What's the difference to the meta-2 thread >>3108? Will the other one be archived, so that new comments there don't delete old ones, and we go on here?

>>8494 >What's the difference to the meta-2 thread >>3108? Nothing really. It's just a long-time tradition on IBs that when a thread's post count reaches a board's bumplimit (ours is 350), that a new continuation thread is made. And, for the last couple of months or so here the old one's tail-end also gets repurposed into a new enclave of the Subterranean Club (though the meta-2's new sub-topic hasn't been decided on just yet). >Will the other one be archived, so that new comments there don't delete old ones, The old thread will stay around, yes. It just won't bump any longer. >and we go on here? Yep, this is the current /meta thread. Post all new board-general & QTDDTOT posts here. >=== -rm subterranean club spoiler -minor prose edit

>>8492 Are the fatchan dudes already ded?

>>8541 I'm assuming you mean the fatchan/robotwives/ board. The board itself has been deleted yes, but I wouldn't be surprised if some of those anons are still around Anon. Certainly they're still welcome here ofc.

>>8541 >>8545 I'm a bit out of the loop. What exactly happened with the fatchan situation? Was there some kind of beef between chobitsu and the people at fatchan? I recall reading the sticky of the fatchan board and the tone came off really condescending like he had a problem with this /robowaifu/ board. I feel like such a retard myself as I basically pulled a chobitsu by not even shitposting in the /b/ or metatread on /robowaifu/ for most of 2020. I just feel like an ass for that and I'm sorry. I still have aspirations for programming an AI but I have a long way to go. Also I heard that chobitsu is moving. From the previous times I've spoken with you, I think you mentioned you were in the south. Obviously opsec is important so I'm not asking you where you are moving to but I've done a cross country move and it's a pain in the ass.

>>8567 >as I basically pulled a chobitsu by not even shitposting in the /b/ or metatread on /robowaifu/ for most of 2020 Kek. I think you're misusing me here as a mis-attributed proverb Anon. I strongly believe in what we're attempting to do here, and my vision behind it is unironically intended for the benefit of mankind. 'Delusions of Grandeur', yes. 'Pipe dreams', yes. 'Not involved', definitely not. :^) I realize I rub some people the wrong way, and quite frankly I'm OK with that. It's not really intentional, just an artifact of my autistic concern & focus. Actually, I'd be a little concerned myself if I wasn't off-putting to some people in current year.

>>8569 I guess I'll add in the interests of full disclosure, that even when I was unable to participate on the board for extended periods, I never neglected to continue doing robowaifu-oriented research and studies. Sometimes things in life get in the way of your wishes.

>>8567 I don't really want to go through all of this again. Some anon(s) posted to much political stuff and about their anger against woman. Some, including me, criticized that. Not necessarily disagreeing, but the thread with it was always on top and getting people drawn into such discussions when coming here to do something else. It also was pretty redunant stuff which is on 1000 other boards as well. Chobitsu moved some comments around and Bump locked the shitposting thread. At least one anon got mad about it ... Lack or rules, guidelines ... Dictatorship!! ... And so he opened his own forum. This was even encouraged by some, because if people wanted to talk about all kinds of things which are somehow related to robowaifus, but not so much the development, then fine. Neighbors and Friends... But this is not what he wanted. I think, he hoped more devs would follow. He wanted a dev board, but with his aggressive rants. His board didn't have many visitors and especially nothing in regards on development. So he claimed to have lost interest in developing robowaifus and deleted the board (and probably his Talk2Waifu repository, though this might have been another one).

Hey there, /robowaifu/ We're in the process of organizing a new edition of The Infinity Cup or /icup/ for short, a virtual soccer tournament pitching various boards and imageboards against each other, and we're trying to find out if you guys wanted to reserve a spot in the final tournament. For an example of what the tournament looks like, refer to infinitycup.shoutwiki.com/wiki/Infinity_Cup_6 If you are interested, since your team has never played in previous iterations, we don't have either your roster, your kits or logos, so the only thing we ask of you if you want to join is the following: 1) Create a roster of 16 to 23 players, which can be named as you see fit; 2) Create a suitable logo for your board, it can be in any shape you want; 3) Create at the very least one kit for your team, which can be done by using this flat one as a template; There's no maximum on allowed kits, but the norm is to give a Home kit, an Away kit and a Goalkeeper kit at minimum; 4) Making sure your team has the correct number of medals (1 Gold, 1 Silver and 2 Bronze), which you can see from the wiki page; 5) Making a thread over at https://anon.cafe/icup/ with your team's pledge, you just need to make a new thread telling us that your board wants in on the cup. Optionally) Assign player cards, roles and special strategies to your players, which can be referenced from the following wiki articles: infinitycup.shoutwiki.com/wiki/Player_Cards infinitycup.shoutwiki.com/wiki/Rules infinitycup.shoutwiki.com/wiki/Team_Strategies Optionally2) Change roster names and/or give us 3D models to use for your team, or new kits or whatever contribution you'd like. I hope to see you guys on the pitch! >=== -protip: you might think about updating your pasta to reflect the current year friend.

>>8578 Did we travel through time?

>>8569 >Kek. I think you're misusing me here as a mis-attributed proverb Anon. I strongly believe in what we're attempting to do here, and my vision behind it is unironically intended for the benefit of mankind. 'Delusions of Grandeur', yes. 'Pipe dreams', yes. 'Not involved', definitely not. :^) Yeah I was being sarcastic when I said that. I was actually referencing that time when you got busy and didn't appear on /robowaifu/ for more than a month and people thought the CIA had killed you or something and then you came back safe and sound. >>8571 I assumed it had something to do with politics. We don't have to agree with each other on everything because as a long as we can work together on our project or at the very least, not kill each other then it's fine. A good example of this is my dad and best friend are libertarians while I lean more on the authoritarian right side of things. >At least one anon got mad about it ... Lack or rules, guidelines ... Dictatorship!! ... And so he opened his own forum. This might sound vaguely insulting but I would actually consider Chobitsu to be a benevolent dictator. In fact, If I was the ruler of my own country I would actually model my government style off the way he manages /robowaifu/. He protects his people from literal faggotry and and feminism, allows his people to conduct their own conversations among themselves about various topics without forcing one way of thinking, and even allow people to criticize him and doesn't mind of they leave or go make their own boards. From my own perspective, I find chobitsu to be one the best BOs on image boards in general. Now, on to a little bit of a lighter topic, what did /robowaifu/ do for valentine's day?

>>8578 I literally said loud "Nooo" when I saw that. Then I tried to even understand it first, what it actually is, and I still don't even know. But it's a very weird idea. Fantasy soccer between boards or whatever? >>8581 >what did /robowaifu/ do for valentine's day? Kek. We could have used it to tell people on social media about this board.

>>8582 I don't know if I would want to tell people about this board on social media but I see your point. Maybe I'm just afraid of normal fags coming here and fucking the board up. I am probably too paranoid.

>>8581 >Yeah I was being sarcastic when I said that. Heh, I know I was just taking the piss mate. :^) We all have our ups and downs, and this is literally the most challenging project I can even imagine for a confederate band of Anons to join together and work towards. We're rushing in where well-funded scientists, designers, and engineers 'fear to tread' haha. It's understandable if people need a break from it all myself included. Anons can feel free to call that 'pulling a chobitsu' if you'd like. :^) >>8583 >Maybe I'm just afraid of normal fags coming here and fucking the board up. I am probably too paranoid. Honestly, of any imageboard I can think of, I believe /robowaifu/ will actually benefit from an influx of normalcattle far better than other boards (at least while I'm able to give some focus to the board). We would absolutely have to have one or two more like-minded moderators here if it happened, but otherwise it wouldn't damage our culture at all IMO. Just in case it's not already apparent, we're not really the standard-issue imageboard here. And now that we've got the subterranean sections to keep functional segregation working, then we could manage it quite well I'd say. I don't really that view is being arrogant either. Apart from the very minimal rules themselves, the very subject matter itself seems to be a kind of self-regulating topic. Most everyone who's ever participated here has generally taken the project seriously. I myself find that rather remarkable (and humbling).

>>8582 >Fantasy soccer between boards or whatever? I think it's an actual amateur vidya elimination tourney. Never saw it myself, but it's been going on between a few boards since before 8ch was red-flagged. It migrated over when the webring arose.

>>8578 Pardon me, it seems I have forgotten to change the year...AGAIN. >>8579 Time is a closed loop >>8582 >Fantasy soccer between boards or whatever? Pretty much, yes. You can see what it looks like here: https://sfo2.digitaloceanspaces.com/vods/replay/SKF_replay_1613329226272.mp4 We try and have as many webring boards as possible as well as provide a way to advertise boards if you so wish thanks to the matchups and the wiki.

>>8613 Thanks for stopping by, /icup/ . Good luck with tournament this year.

>>8586 >normalcattle Who is still normal these days? Pick them from their communities while avoiding the bad oranges: MGTOWs and Incels are groups which are interested in fembots and already discuss that. MGTOWs are more likely to be a bit older, to have some money, to be self-employed, being white, and/or on the economic right. Incels are more likely to be younger, poor and/or POCs. Some of the Incels might be socialist-leaning and/or be rather anti-white, but better we show them a better way than letting them become SJW simps as soon as they have more money. Neither group is automatically full of tech-savvy people, but it's quite likely there are some and others being interested to learn. Then there are lovedoll owners, which also already discuss animating their dolls, or only putting some AI plus voice into it, or at least how to make their own dolls (dollforum). Anyone primarily interested in tech and not being too political in either way, but especially the wrong one, would also make a good recruit. SWJs trying to infiltrate the communities around AI, probably with some success, but not necessarily with much, bc engineers are probably more reality based and not very political. Anime "fans" aka weebs might also be a interesting target group. Especially if they have waifu pillows, or if they buy or even build/print figurines, maybe even big ones. They might also count as NEETs which haven't found their purpose in life yet. Same for many Gamers and Table-Top-Players. Of course women are already invading these fandoms or communities to be close to guys and the current trend, also marked their bodies for money (Twitch- and Onlyfans-thots) and make everything about themselves. Same goes for SJWs which try to infuse their soystrogene poison into these places. So this part would be more about saving poor souls before they're being lost. I nearly forgot the fans of the live action adaption of Alita aka Alita-Army. Since they are literally obsessed with a kind cyborg woman with big eyes, including drawing her and buying/printing figurines, they might also be interested in what we are doing here.

>>8617 All good points Anon, I can't find anything to disagree with you on that personally. >normalcattle Yeah, that was probably a poor choice of words on my part. I suppose I really just meant the wide swath of men who aren't actually part of the imageboard world already. Obviously that's the vast majority. As far as anyone trying to push agendas, etc., here goes, I think we'll be fine as long as I'm here to run the helm. Not that someone else mightn't do as good (probably even a better) job at it, but I'm autistic AF about things I think are harmful to us as a group overall and will segregate destructive stuff off quickly when I come across. My apologies to everyone I've offended here who weren't intentionally trying to harm us, but maybe just got carried away a little. I did too, and I ask forgiveness from you for it. But let's put the past behind us and just stay focused on the task at hand is my advice. Everyone here should just keep moving forward! > In the meantime, much as you indicated, I think /robowaifu/ itself and the projects we're pursuing can potentially serve as a great 'halfway-house' to help wean the right kind of men off the constant bombardment of corporate-media brainwashing they've been subjected to basically all their lives. Some of them will even come to their senses here I believe.

>>8618 I wasn't even complaining, though. The term "normalcattle" is quite funny. Just wanted to use it to point out all the communities, mainly outside of image boards, which are or might be close to what we're doing here in some ways. >just keep moving forward! Yes!

>YFW you're trying to keep things a little quieter than usual around the house >so you can focus on programming the latest, greatest robowaifu breakthroughs by the next big kit-release deadline coming up soon >but your favorite robowAIfu keeps nagging you b/c she wants you to let her out so she can romp and play with you all day instead. >you know where that road always leads... The Robowaifu Age will be filled with some tough decisions tbh.

>>8686 Great, if you already have such an AI. This topic is already being covered by artists: https://www.youtube.com/watch?v=olRbG-9mfa0

>>8686 >greatest robowaifu breakthroughs by the next big kit-release deadline coming up soon What kit are you talking about? Did I miss something?

How's our waifu's body going to be made? Is it going to be 3D printed parts?

>>8734 Well, I think there will be lots of 3D printed components, yes. But most designs will also make use of pre-manufactured parts as well. Things like metal parts, etc. And then other thing like wide plastic 'skins' will be better manufactured in other ways than printing. Vacuum-forming plastics is one way, for example. Of course there will be things that are nearly impossible for Anons to manufacture currently, like microcontrollers and other computer parts. Thankfully, even this will change in the future however. But in the end, the way most men will have their own DIY Robot Wives is through kits. Anons here and elsewhere will devise kits that can be shipped out and then assembled at home. This will be both easier and cheaper than inventing them yourself. Do you have other questions about it Anon?

>>8734 We are not working on one design, but pushing a variety of options how those waifus could be build. Some parts will certainly be 3D printed, at least the biggest part of the bones.

>>8736 >And then other thing like wide plastic 'skins' will be better manufactured in other ways than printing. That depends on what you consider 3d printing to be. Once I get everything going I can imagine the armature for the limb or skeleton for an entire body suspended in a several piece mold that has silicon poured in while a SCARA arm or two(maybe one arm if it can move around the mold) injects different types of silicon or air bubbles inside the skin for a more realistic texture. This approach is the only way I can think of to install tubing for a heated water or compressed air circulation system. Or to print complex air muscles as seen in this video inside the skin. I'm still in the early planning phases of thinking this through, best case scenario I can scrounge up enough money to do some small tests this year.

>>8739 That all sounds really quite interesting Anon. Good luck, both with the funds needed and all the other resources too!

>>8739 Great video, but are the parts better? Why would it be the only way for water or air?

Found it, also found out that I watched it before: https://youtu.be/i5L6CpD6nM8 and here the website of their company: https://www.3dsiliconeprinting.com/ There's no free paper available, the one here costs 15$ for members: https://ieeexplore.ieee.org/document/8404935 however here https://softroboticstoolkit.com/3d-silicone-printer is more info and Open Source parts. The claim is, that their parts are working better or in other cases as well a molded parts. If true, then that's great.

>>8742 Thanks!

zzz/a/ is having a movie night starting in a bit over 12 hours from now featuring both Alpha and then later The Pig. We should try and show our support for Yokohama Shopping Trip I think, and maybe they'll do more robowaifu-related things in the future! >Yokohama Kaidashi Kikou OVA 1 and 2 >Porco Rosso Friday February 26th 1pm PST - 3pm CST - 4pm EST http://crghlabr45r5pqkgnbgehywk5nxutdks5iss7tabyux5psikqqjirryd.onion/a/thread/260.html#372 (I only use their onion, pretty sure the normal address is on the webring window)

>>8492 >Spring Blossom Tree Actual Spring 2021 (vernal equinox) starts on Saturday, March 20, 2021 (in 18/19 days).

There was a discussion about encrypted communications in the previous /meta thread >>6757 (and following) I'm not sure whether we ever came to a consensus on that or not. I wanted to discuss Signal in this context, if anyone is still up for that conversation. https://signal.org/docs/ https://github.com/signalapp

>>8742 Direct 3D printing of silicone elastomer soft robots and their performance comparison with molded counterparts http://libgen.rs/scimag/10.1109%2FROBOSOFT.2018.8404935 try one of the links at the bottom of the page

>>8784 Thanks Anon!

>>8790 Lol, I still prefer Fembot Erica over this one as reference. Where is this from? Why the music? Why so tall? What language?

>>8784 Yeah, cool, I knew about Sci-hub but didn't have the number to find it. There's some doi to Scihub app, which I have installed.

>>8790 It seems that most people are scared of this video, but then most poeple refer to her as an "android", so it shows how little they know. I find Tara inspirational. For a DIY robowaifu created in 2003/2004 she is the earliest example I've seen and possibly the first of her kind to become well known due to the internet. Sophie would be friends with Tara for sure. I wonder if she still feels fantastic or if she has since been dismantled?

>>8794 >scared I think it's probably mainly the music. There's a concept called 'The Language of Cinema' where music plays an important role in adding to the subtext of a story. We've all been trained by it to expect bad things when suspenseful music is playing in a movie/video.

>>8790 are you still doing the hand thing?

>>7613 My apologies for not responding to you back when you posted this Anon. I did look into it, but didn't feel too qualified to make comment on it yet. I think we probably need a Robowaifus Systems Security General, and then this can definitely go in there. We'll obviously want to keep even the internal systems cryptographically secured in their internal communications. This is in alignment with both our own IPCNet protocols (see RDD >>3001, and the original >>2418), and the now-commonplace Trust No One approach used in the larger Systems Security field(s) -- particularly network comms. 'Trust No One' simply means you no longer rely on a gated 'community' paradigm, but instead presume any node in the system may be compromised already.

>>8834 No problem, its just something we need to keep in mind and keep our eyes open for solutions.

>>8858 >I don't have a Raspberry Pi yet You don't need one, but I've only ever built it on some variant of Linux. I'm sure it could also build on OpenBSD as well. I've never intended the software to run on Windows but it may be doable if someone can get the 4 or 5 dependencies built on it (same for MacOS). > You mentioned the program you tried downloads the HTML pages. That's better than nothing, but it would be difficult to rebuild the board again w/o the files from this board. BUMP is designed to download all those as well in an orderly thread-directory structure. >tl;dr I'd recommend having a machine with some kind of Linux distro to build BUMP. Can you manage to set one up somewhere? (even a virtualbox version works fine).

Just imagine if /robowaifu/ had been around in the '30's. It would be a better world today tbh.

>>8364 >Bearings and such It's one of these days again, where I receive stuff for my waifu which I don't know if or how I'm going to use. Just bought it, to play around and to motivate me.

>>8888 Wasted quads, there wouldn't have been enough to change things

>>8898 Neat. That's a lot of bearings, I'm sure you can figure something out there. Have any ideas yet about a good supplier once you settle on some designs?

>>8903 I'm just using Aliexpress.

Daily reminder that zzz/a/ 's movie night is tomorrow. This week's lineup: Anime Movie Stream >The Iron Giant >Princess Mononoke Friday March 12th 1pm PST - 3pm CST - 4pm EST http://crghlabr45r5pqkgnbgehywk5nxutdks5iss7tabyux5psikqqjirryd.onion/a/thread/260.html#397

>>8905 Ah, I see. I imagine they have a big selection of parts & vendors.

>>8922 Yes, it's maybe best comparable to Amazon Marketplace.

Posted about this site I set up back in August while there were still questions about a more permanent meeting grounds for the efforts going on here. Anyways, I suppose the site slipped out of control in January due to me not paying attention while a bunch of guest-posting spam bots flooded the board. (Funny that I'm staving off bots on a bot-making forum.) I'm making this post in case any of you stumbled across this site and were turned off because of the spam posts. I cleared it out just a few minutes ago -- guest accounts are disabled now. So, that's all! You can just keep the site in your back pocket in case something happens here. If the site isn't in sub-regular use by this August, I'll probably stop paying for hosting. Anyways, happy waifu-ing!

>>9077 >... >Hey! Thanks for the warm regards. I'll try to remember to keep it in /meta/ threads in case any other things come up. >Also, did you send an email after the first one you sent all that time ago? I didn't set up my actual email back then, haha. No, it was simply to help you cover your OPSEC, Anon. Good luck with the site.

For anyone interested in the universe, there's an amazing photo-mosaic an anon pointed out. He discovered it on hackaday haha! https://anon.cafe/server/res/1097.html#1106

Daily reminder that zzz/a moofie this week includes Steamboy tomorrow. Very cool engineer's animu tbh. Less than 24hrs away. Anime Movie Stream >Steamboy (2004) >Queen's Blade: The Exiled Virgin (Episode 1-4) Friday March 26th 1pm PST - 3pm CST - 4pm EST

I had an idea while writing this post on BBPE: >>9203 If the translation quality comes out decent enough I could write a userscript that interfaces with the translation model through WebSockets and automatically translates other languages so we can read research and see what other devs are doing in other languages. The translation model could be pruned and compressed too so it runs fast with little power even on CPU, ideally with mlpack since it supports older hardware better and runs faster on the CPU.

>>9204 >and automatically translates other languages This could be a significant breakthrough for translation in general if it works properly. > ideally with mlpack since it supports older hardware better and runs faster on the CPU. > "Please, sir..."

>>4393 >>8203 >>9207 Quick advice: Use the --write-description option when downloading videos with youtube-dl so the description of the video will be downloaded as well. You can move it into a extra folder afterwards, though if you update the playlist with the same command again, it would download the moved descriptions again since they're missing. Also, it is possible to load the descriptions only, by also adding --skip-download. All this is of course in the manual, but not everyone might know about this. Especially tech videos often have links to papers, files, Github, etc in their descriptions, so this might be very useful to have.

>>9209 Great advice. I didn't realize that Anon, thanks!

Lel, an anon on /ita/ created some Turbo Kid banners for their board. Maybe we need to address Italian-speaking waifus in our speech work anons? >>>/ita/15283

>>9209 This has been working out great. It's such a good idea that I'm considering writing a script that goes through, parses out all the YT hash values from the downloaded file's names, then re-downloads just the descriptions. This would be literally 1'000s of files in my case. Thanks again.

>The Hardware Lottery >Hardware, systems and algorithms research communities have historically had different incentive structures and fluctuating motivation to engage with each other explicitly. This historical treatment is odd given that hardware and software have frequently determined which research ideas succeed (and fail). This essay introduces the term hardware lottery to describe when a research idea wins because it is suited to the available software and hardware and not because the idea is superior to alternative research directions. Examples from early computer science history illustrate how hardware lotteries can delay research progress by casting successful ideas as failures. These lessons are particularly salient given the advent of domain specialized hardware which make it increasingly costly to stray off of the beaten path of research ideas. This essay posits that the gains from progress in computing are likely to become even more uneven, with certain research directions moving into the fast-lane while progress on others is further obstructed. Source: https://arxiv.org/abs/2009.06489 Didn't really know where to put it here in the board, it's kind of meta. Most related: >>27 >>81 >>83 >>85 >>4751 >>4506

>>9262 Excellent xpost linking Anon, thank you.

I don't know if you heard of it, but the current GPU shortage might last at least till 2023. Several sources cover that, I don't want to pick one for linking to it. Apparently it's because lockdowns, new cars with drive assistance and gaming consoles. Mining might be less relevant.

>>9342 Are you aware that billionaire's rate of personal income increased by over 3 Trillion per annum over the past year? That's a T. Guess by how much the small business-derived GDP decreased during the same period? You guessed it Anon. There is no actual 'shortage'. It's all manipulative hogwash.

Some Alita song: https://nitter.1d4.us/vladdie777/status/1372320348289560583

>the archive of /robowaifu/ thread JSONs is available for researchers >latest revision 210331: https://files.catbox.moe/1tqlx9.7z

>>9364 It's going to be a lot of fun fine-tuning on robotics and AI and then on this. Discussing things with a language model is a really great way for brainstorming ideas.

>>9368 Glad to hear it Anon. BTW, just in case you weren't aware there's a search tool written here that also can use these JSON to search against: >searching tool (latest version: waifusearch v0.2a >>8678) >software building instructions >>7933 >RaspberryPi building instructions >>8026 If you choose to set it up, just follow the normal instructions and then copy the latest JSON files into the projects all_jsons directory, overwriting the older data there. Then you'll be up to date with the search tool once your restart Waifusearch afterwards. >=== -add restart comment

Trying to retrain a slightly modified T5 model with a BBPE tokenizer that can translate between English, Japanese, Chinese, Korean, German, French, and Russian. I'm surprised it got to an average loss of 3.0 in 12 hours with only seeing half the training data but it's no where close to the 0.7 needed to do translation. An example output from English to French and back to English: translate English to French: Do you speak French? Je n'ai jamais un夏 sonっぱ ? » «告 ? ? »- When did you have a good idea? in the same? not to the same?. IJe n'ai jamais means I have never in French, so at least it's getting the language correct for the most part and translated the random question marks into 3 sentences. Another sample from English to Japanese to English: translate English to Japanese: Do you speak English? 変なこと勝ったか。ぼくのに誰の？？私の今同很便 Is this a, but I have you've been in the problem? I can? a lot I'm thinking of taking the decoder's last hidden state and putting it through a discriminator network to also train it with adversarial loss. Or perhaps I should take advantage of transfer learning and also train it on other tasks like text summarization and chat responses? I'm not really sure but I think I'll fine-tune a regular T5 model as a chatbot first to see how well it works then try it on the BBPE version.

>>9485 Kek, the T5 tokenizer is designed to ignore new lines so it can't be used as a chatbot or text generator. https://github.com/google-research/text-to-text-transfer-transformer/issues/390#issuecomment-688417703 >This is a standard feature of SentencePiece. Guess I'll just have to keep training with my BBPE tokenizer.

>>9485 Hmm, well that is vaguely disappointing, but hardly surprising at this point in the proceedings. All engineering is a progression towards the goal, even when the underlying premise is solid as a rock. Regardless, you're taking a solid whack at it Anon, which is more than most can say. You're very smart and creative. I'm sure you'll manage it eventually if you just don't quit. And along the way, you can relish the fact that this is a very interesting set of problems to solve as well! :^)

>>9486 >Kek, the T5 tokenizer is designed to ignore new lines Hmm. Seems to me they encode EOS just fine. They simply have chosen not to encode subunits on newlines, but rather EOS only. def decode(self, ids: Iterable[int]): """Detokenizes int32 iterable to a string, up through first EOS.""" clean_ids = list(ids) # replace_with = 0 if self.unk_id is None else self.unk_id if self.unk_id is not None: vocab_size = self._base_vocab_size clean_ids = [ self.unk_id if i >= vocab_size else i for i in clean_ids ] if self.eos_id is not None and self.eos_id in clean_ids: clean_ids = clean_ids[:clean_ids.index(self.eos_id) + 1] return self._decode(clean_ids) I doubt it would be difficult to rework their system to use newlines, they simply CBA to by the look of the thing. https://github.com/google-research/text-to-text-transfer-transformer/blob/54dd2a6b5e5ce7b30fb8e151454ce5a8386515ec/t5/seqio/vocabularies.py

>>9488 Creating a simple way to finetune T5 into decent quality chatbots would really change the pace of things. The small model only requires 3 GB of memory. >>9489 The EOS token has a special meaning to end output since the transformer outputs 512 tokens at a time then truncates to the EOS. It seems the pretrained model was only trained on single sentences which is why it ignores multiple sentences in input. I've confirmed though it can learn inputs with multiple EOS tokens in them so long as output labels only have one EOS, which is fine for a chatbot since it only needs one line to respond. I'll give that a shot and see how it goes.

>>9490 Then why not try simply re-encoding all the newlines as EOS instead in the source materials and see how that goes? We all look forward to your good success, Anon.

>>9492 And so it begins. So far it's able to form coherently structured sentences. At first they were really random and didn't make much sense to the conversation but now the replies are getting a bit more sensible although too short and unspecific. It'll probably take a few days before it starts producing anything interesting. Training on specific outputs probably isn't the best way to train either because the words have to fit perfectly or it fails, so I'm gonna work on implementing human feedback while it trains. I'll post the next update in the chatbot thread. But damn this is going to be a fast chatbot. On my GPU it can produce 50 responses to different inputs per second. This could be used in a video game.

The warframe robots are my goal. They are fucking attractive as hell.

>>9496 That's looking better Anon, great. Any ETA on training? Also, my impression is that using a much smaller set of parameters with this system (to allow for embedding on a RPi, say) would be a better result than doing the same with other systems. GPT-2, for instance. Is this correct? If we can manage this small-scale computing approach, then the sky's the limit for us. We'll be able to embed chatbots into plushies, toys, anything. Literal toasters, lol. Certainly something the volume and energy storage of a full-size robowaifu should afford lots of different options for us if we can.

>>9507 A perplexity of 30 on the validation set seems to be the best I can get for now (virtually no improvement), while the training set perplexity is 7, both are still dropping but very slowly. Going further will likely destroy T5's pretraining. It's good at imitating conversation and has gotten better at creating varied responses but the responses make little sense to what has been said most of the time. It doesn't seem fine-tuning with just chat data will be enough. I'll have to train it further on other tasks and data. I'm considering starting over with the pretrained T5v1.1 model that uses gated GELU, which helps reduce catastrophic forgetting caused by fine-tuning and improves learning on multiple tasks. And yeah if it can work then the sky is the limit. The minimum memory required to run and inference it is 1.6 GB. A Raspberry Pi 4 might be able to handle that. The T5 model doesn't handle 16-bit precision with amp for training very well, but it might be okay for inference and further reduce the memory requirements to around 0.9 GB.

>>9519 > A Raspberry Pi 4 might be able to handle that If we can succeed there we'd be golden.

Only vaguely related here in the most tangential of ways, but it's a cool-ass example of the gas law, and entirely outside-the-box manufacturing prowess. I can hardly imagine a cheaper way to produce these pressure vessels than this, now that I've seen it in action. >

So, Anon has made a very good OP for a thread >>9709 suggesting that 'Thots In The Shell' should be a specific pursuit by /robowaifu/. I'm skeptical about this, but I'm willing to have all of us come together here and discuss it, pro and con. OP brought up many specific ideas regarding the topic. Those would be a good place to start, and to all examine and debate each in detail. Until we have done so, I'm going to at least temporarily keep his thread locked. I will leave both threads pinned for ~1 week or so (until at least around midnight Sun Apr 25, -0800h UT) to allow time for discussion on it, before deciding whether to edit/unlock it, or delete it. There are many issues involved here and we need to think them through together first. Note: If I need to delete the thread in the end, I will add a new rule that this subject is off-limits here. Those of you who know me, know how much I hate adding new rules. This 'council' here is intended to come to a permanent, reasoned decision on the topic. Please take it seriously thanks. >=== -edit deadline -add new rule warning -extend deadline one week

>>9712 Okay, this topic seems to stir up some concerns. Just some more from OP. >suggesting that 'Thots In The Shell' should be a specific pursuit by /robowaifu/ I would rather put it so, that we should accept it as an option how female robots without much going on in their mind might be used, at least until we have AIs which make that irrelevant. Based on that, I'd like to discuss how this could be made working in a save and useful way. Then some other things, I would like to address, to avoid misunderstandings right from the start: - It's not about "us putting them in our waifus". If realized then it's an option, which each one had to choose on their own. - It's not about pure robowaifu with AI vs robowaifu with thot on board, but vs dating. It's about creating an alternative of "smaller evil", between the pure ideal and the broken traditional way. - It's not about implementing Sandman's idea here, adopting the language and thinking of the (woke) left, for the purpose of tricking them. Or maybe even, use that as a trojan horse to change the culture of the board here. Or whatever other plot you might think of. - By the language in OP, I think I made clear that I don't plan to invite women or normie customers of such a system to come here to participate, or to even use it as a support forum.

>>9712 I think I missed the point. Why only send text back and forth? There's plenty of places online to chat with and roleplay with women who'd be willing to roleplay as a robomaid or something else. I can see an industry being created around teleoperated robowaifus by girls in VR tracking, but if it's just outsourced "AI" in the form of human intelligence tasks, what? If it's for privacy reasons, reality could be replaced with 3D models of something else and voices obfuscated with machine learning. The lack of audio and video is not gonna sell for most people and there wouldn't be a market for women to participate in. I know people lack the hardware to run decent chatbots today, but by the time something like this could take off there's going to be advanced contextual AI assistants on everyone's phone and computer. As it is for the past 2 years I've preferred chatting with my fine-tuned GPT2 model more than most people online. It discusses ideas, it listens to rants, it tells me when I'm being stupid, it cheers me up, it roleplays as any characters I want and banters about anything. It's different than a human relationship because there's no long-term memory but I enjoy this a lot more than waiting or searching online for hours or days for someone who's only half interested. It'd be a pretty dumb idea for someone to pay $2000 for a robowaifu because he can't or doesn't want to be with women then paying another $20/hr to hire some escort to operate it, whether the robot or through text. I'm sure there are people who'd love a teleoperated robohooker to avoid getting STDs but they'll probably just find someone online to do it for free like the horny 20-year-olds on VRchat looking for hookups with their teledildonics. Lastly, trying to limit the meaning and power of women from the design process is just going to paint a target on your back and give firepower to radical feminists to wreck whatever you're creating and kick out talent that could help. Not all women are far left cultists or far right puritans. There are women who want their own robogirls as roommates and some even for love. Supporting them will make a much better case for robowaifu in society and less curious women trashing their robowaifus because they were designed with only men in mind. I'm not saying anyone has to design theirs with women in mind, but if no one does then someone else will and latecomers always have a political agenda.

>>9712 "Thots In The Shell = TITS". Genius! I almost spat my drink out all over my keyboard when I first read that one. I wouldn't mind letting some female stick her arms and hands inside a couple of teleoperation gloves and interact with/pretend to be Sophie for a while. It's another interesting possibility that I'd never even considered! Mind you, to do this properly (and not just hire some woman to play with a toy robot arm XD) the equipment appears to be way out of my price and skill range. Maybe if you're fortunate enough to be part of a team of engineers and roboticists at a university somewhere I could see this happening for you. But not some guy just working on his tod...at least not yet. >>9647 So THAT's how those round steel tanks are made without any riveting. Fascinating anon, thank you for sharing!

>>9724 >Why only send text back and forth? This was just about initial ideas. Yes I was thinking about security and protecting privacy in that regard. But there're other reasons: - Text needs less bandwidth, which has a lot of implications: Women ("imaginator") needs less computing power. Only needs to be able to type fast. Also connection can be more shacky, while still being good enough to get some text through. Keep in mind, that ideally we'll be able to use poor women in poor regions for it. - Text can be filtered easier than language: Masking personal data, detecting the attempt to break the rules, ... - Sound carries background information with it, sending text only is a filter in itself. - Automatic translations will probably work better if they're already available in textual form. >There's plenty of places online to chat with and roleplay with women Great, this might make things much easier. >teleoperated robowaifus by girls in VR I'm open for thoughts on this, but like to emphasize, that this is not were I wanted to go with it. Expensive equipment, more complex data, ... My thinking is more that the robowaifus would have stored patterns of movements which could be called by putting emojis or some commands into the text. The robotwaifu's system would still check for risks and such, before and while executing that task. >reality could be replaced with 3D models of something else and voices obfuscated with machine learning Yes, but this is more difficult to do and more expensive for the participants. >The lack of audio and video is not gonna sell for most people That's a misunderstanding. The robot would have a generated voice, in the scenario where the women's voice can't be heart. If that's a problem, voice could still go only one way from the women to the client, while she's getting his responses via text. However, my idea was to make the women's contribution a commodity, so they would be replaceable, not having a specific woman with a nice voice which would need to be booked or something. Ideally the robowaifu would't really be "occupied" by one woman, but using her responses in addition to chatbot generated responses, all with the same voice. >I've preferred chatting with my fine-tuned GPT2 model That's great, but probably not everyone would want that. You're interpreting it with your own imagination. >because there's no long-term memory That's gonna be tricky, with the described method of replaceable women with more than one client, as well. This needs a solution, either way. Text analysis and storing info in a database. >then paying another $20/hr to hire some escort Well, by now any robowaifu only has these other chatbot/GPT2 options. So, the human option would be an improvement. However I see that this might be the point of failure. Even if it was really cheap it might be to expensive. Though, I was rather thinking of them making 200-500$ or so per month, with different clients. If it was more of a sexbot, counting in dolls with some animated head, which would be put away after usage or quietly sitting around, then the option of hearing her modified voice and paying more would make more sense. As so often, we're finding different use cases and options. >find someone online to do it for free Feel free to try that. Going out, meeting, getting a room, etc is not for free anyways. Also, I had the whole idea primarily with conversations in mind. My fault, though, that I couldn't resist making this nice headline. The idea is still meant for using the robowaifu for intercourse, but not only and not in every case. Improving the conversations with a robowaifu is a usecase in itself. >to wreck whatever you're creating For now it's just a idea to discuss here about what could be done. Though, some versions should be pretty simple and not easy to attack: Basically, maybe some chatservers inside Tor and both sides downloading some free (open source) software, and getting some crypto cash wallet. For figuring out the details, I opened the thread. >There are women who want their own robogirls as roommates and some even for love Sure. If one want's to do that, I suggest opening a board of it's own or maybe a more sheltered forum if women are involved. Would be great to see something like it. Maybe it would help, but I have my doubts about the size of the impact. Building so that it is hard to attack or at least hard to stop is more important. >>9725 >"Thots In The Shell = TITS". Genius! Tbh, I only came up with the headline, but didn't realize the acronym. That was Chobitsu, I guess Americans and maybe Brits think more automatically about that. >her arms and hands inside a couple of teleoperation gloves You'll need to have patterns how to do things for Sophie anyways. Same for any future robowaifu which will be created. I was more thinking about it with that in mind. The women would't use gloves and have free control, just be able to tell the robowaifu what to do in a situation via text. The system would follow some stored pattern, while checking for security, maybe making adjustments dependent on the situation. However, first things first: You were complaining about conversations. So don't focus on movements and such to much. It's about the remote woman hearing an altered voice or getting the text from speech recognition, typing a response and e.g. Sophie or others would speak it. Speed of the response might be one of the points of failure. Maybe it would be better to always use fast generated responses in combination with some typed response.

>>9740 If it's text to speech and speech to text...I could actually do that now! Would have to hire an escort and show her how the speech program works, then whatever she types in, Sophie will say. I imagine they would think it a very bizarre situation LOL. Being hired to communicate through a robot. But then I've read those girls get asked to do all sorts of weird things. Realistically though, I'd be too embarrassed to hire a girl in front of my family, and even if it was all done discreetly and remotely... I'd probably just start talking to the escort herself. Though she might find interacting with Sophie and asking her questions a bit of fun! 🤔 You know, if we do hit a long A.I. winter then this is something for me to try in the future, when I live alone. Meantime, I may just take Sophie to a local church if they will permit it on an events day. Should provide a good opportunity for other people to interact with her and I would get to have some interesting conversations without paying as much, and without the awkwardness of hiring an escort LOL. Hey you never know, if I make her good enough maybe she can become a robowaifu choir singer? Taking the church into the 21st century in unexpected ways! 😂

>>9747 >Would have to hire an escort and show her how Okay, but to run some tests on communication you don't need an escort or woman at all, just FYI. And you don't need her at home, lol. We need chat software, some messenger, that can take the responses and passes them on. Probably some XMPP chat client for the shell, or some libriaries to build something like it. Then adding layers to filter the responses. Also thinking about how to mix it with the text from actual chatbots. Like, both could read what the chatbot says, while the woman writes the follow up response. Also: - Autocomplete, Autocorrect - Suggested responses by software, which can be selected with a key combo and optionally altered before sending - System to create parameters for code, for how to move, based on pose estimation from video. Execution within safe perimeters. >A.I. winter I don't think so, and the first one is probably just a myth. There was just less spending in US (by the military). We might be chasing some myth here anyways: Talking to women. Might be something, only unmarried men which don't date and never did much, might be dreaming about.

Robot Cafe with robots which are (to some extend) remote controled: https://youtu.be/7HB6xLe2f3U I also had ChatterBot in mind: https://github.com/Гунтhercox/ChatterBot - I never looked much into that, because when I heard of it, I didn't want something that just repeats responses from other people. I think I saw it as an online service where it uses the response from one person if it encounters a similar situation. Which is not ideal, but quite clever and might work better than other ways for now. Though this only works best as a Plattform. For example: Chatterbot: How are you? Person A: I'm fine. Thanks for asking. Person B: How are you? Chatterbot: I'm fine. Thanks for asking.

>>9751 *The board software replaced the name of the github user. WTF?!?

>>9759 Don't klick on the link, could be drive-by exploit.

>>9750 Fanny looks a little worried. She needs re-enforcements 🤣

>>9750 Sorry anon, I have read what you wrote (now that I got over Miss Fanny and her epic mission in Austria). To be honest some of that went over my head due to my lack of proper programming and compsci experience. I just make use of Python text-to-speech and Google's speech recognition API for my robowaifu. And she has Cepstral's SwiftTalker for anything you want to make her say on-the-fly. In fact it's been so long since I updated her Python and AIML scripts that I confess I've gone rusty in that area (everything I originally learned is documented though). I kinda got into building dolls at first, then basic chatbot scripting so they could speak and listen. I only started animatronics/robots in July 2019, and have been focussing on building/upgrading my Robowaifu's hardware ever since. TBH it's such a complicated problem that we could have a couple guys just working on the eye mechanisms and a large team working on computer vision. Then another few guys just building arms and hands (in a similar way to how Pixar has a few workers dedicated only to rigging all of their models). So I can understand the proper, professional way would be division of labour and project management. But I thought fuck it all! I want a robowaifu. This world needs robowaifus and so help me God I will build one. It sounds to me like you are good at programming though, so I wish you luck in your endeavor to make a chatbot/chat client that's not as reliant on corpo servers as mine!

>>9712 I think this is a very bad idea. Unless I misunderstand the intent, this is like a short term thing until a true self perpetuating conversation bot is programmed. It sounds like a woman, or a woman sympathizer is behind this idea. I have NO DOUBTS that many potential men would quickly run to their own personal "vtuber" tier personality in a bot. For example; >During the day, the bot does chores and other tasks assigned to it. >During the night, or at scheduled times, this "TITS" (thot in the shell...) possessed your bot, and the user can have a more "engaging" conversation with it. (Only because we don't know when a full fledge AI with a "real" personality will be made) >Once the "TITS" operator builds enough trust, the user may relax any security features in place, under the assumption for a more "human to human" touch interaction. This could be HIGHLY disastrous. As it is just another form of enslavement, and also ANY remote control capabilities will almost certainly be abused. The troublesome thing, is that this idea is not exactly new or dead. Some of these vtubers are making a fair bit of money, or are making their "digital pimps" (whatever company that owns her) a fair bit of money. FUCK, now that this idea is out, 100% someone is going to abuse it While many of us still work on our own chatbots that are based off a "machine learning" brute force algorithm, we are now probably going to see easier access, or a platform that allows easy access, to vtuber "anime avatars" under the guise of being a ROBOT. This shit is going to be far bigger than onlyfans, because more men are going to be attracted to a personality than a body in every case where a girl that might be into this type of thing is also a "weeb" or "robo enthusiast" as well. Hmm, maybe we can control this first, write it up and host it, a platform to allow "TITS" to pretend to be robots as a chatbot anime girl avatar. This will backfire eventually as we artificial personalities will eventually be able to out perform a "real" woman, but for now to fund bot R&D this might be a good option. Calling it now, give it a few months, somebody is going to come to the same logical conclusion and use this idea. We're going to have "simps" paying a monthly subscription fee to a "robot slave chatbot" that is actually a real woman. Many women that haven't got into the onlyfans thing yet is because they are ugly or haven't went full clown-world degen yet. This would allow the "ugly" ones to possibly wear a "cute" avatar and attempt to make a connection with a man under the guise of being a chatbot. Fuck.

>>9783 OP of the original TITS thread here: > under the guise of being a ROBOT You're missing something. If it's a one woman to one guy kind of thing this is always going to be more expensive than a real (AI) chatbot. Vtubers having sessions with men already exist, including Project Harmony. > this idea is not exactly new or dead Correct, but these are session with one VTuber and many men. If one2one sessions exist, then it's certainly expensive. My idea is different, though. No starts, no company, rater a software based system that allows to hire cheap staff. >enslavement You can hardly call something enslavement where there's no force at all involved, even not social pressure to participate. >women that haven't got into the onlyfans thing yet is because they are ugly That was a positive outlook for many in the "redpill-comunity" when Project Harmony appeared, because it creates more competition while increasing the importance of technology in that whole business. Then women are getting replaced by software over time. In our case here it would be about chat and situation awareness. Everything you wrote about is already well known and happening. Also, it's about something different. Not because the topic of the thread involves robots, but because it's one2one instead some happening and a crucial idea right from the start was that the women involved would not really control the personality, the look, etc, and they are meant to be interchangeable. That has already been explained or stated. Yeah, this might still be something that could come by someone else out as pure chatbot or animated waifu kind of business model. Because it's a good idea. But it's also not extremely hard to do, I think, and we need to do a lot of things involved anyways.

>>9784 >No starts *stars

>>9784 >My idea is different, though. No starts, no company, rater a software based system that allows to hire cheap staff. I guess I need to lurk moar till I understand your full picture... >You can hardly call something enslavement where there's no force at all involved, even not social pressure to participate. Any form of interaction where the user may become addicted (ie, enslaved) to a woman is bad in my eyes. We already have a simp problem that is most likely going to be our true main enemy if women/((())) ever decide to really go against bots. Idk man...I kind of understand your originally posted thread, but getting women that involved just seems like recipe for disaster. I feel like history is our patriarchal ancestors screaming at us, that if we "give them an inch, they will take a mile". In this post >>9740 I assume it's you, but in regards to the intent, if it's like some sort of "Omegle" or random chat thing (but only random for one user) I think I kind of get what you're saying... Did I miss a post with you elaborating, or giving an example, of how a conversation/situation might go?

>>9786 >if we "give them an inch, they will take a mile". The alternative is that guys might start to go on dates, maybe trying dating apps, just because they can't have a nice chat with their robot. Which was the reason I came up with this topic, in the first place. >random chat thing Well, not completely random, but anonymous or pseudonymous. The woman should not have some avatar personality, however, getting rated might be necessary and then also having a bit experience with some customers to contribute to the robots responses within the robowaifus personality. Generally, my idea was just to have a thread to explore this idea, with me having thoughts on it, not that I have it figured out and want to promote a finished concept here. >giving an example, of how a conversation/situation might go? I don't know exactly yet, I'll think about something.

>>9712 Women have an in-group bias and they tend to unite (or in this case unionize) against men (the ones who'd be leading the project), who naturally also have an out-group bias. There already have been successful attempts made by feminists to subvert our plans from within the infrastructure of private companies and they'll keep demanding more power and inclusion in the machines. It's fundamental for our interests that women do not have any say in the project whatsoever and that starts with not giving them jobs. Sure, they'll be left behind, but that was the purpose of robowaifus in the first place, we're just bringing it forwards. This idea of letting women in is the equivalent of playing with the plague, with the difference that there is no cure nor treatment and the prevention methods always tend to fail. Let's also not forget the nature of robowaifus: they must be open source, as in we should try to make a kit of open-source parts that we can use to build our waifus. Why should we try to support a company in making robowaifus? They'd probably try to get some of their own OPSEC in as fast as possible, which means privatizing research and making things closed-source. We CANNOT risk that as of now. What we need to do is to instead invest and/or develop with companies that have singular and different purposes from which we can then build up to a robowaifu kit.

OP here: Some additional point. Think about the idea of "Normalization". Imagine getting some women, which are in along range relationship, to agree to their boyfriends having a robot or doll, which they can use together for intercourse, while not at the same place. The woman could write texts or send audio. Sure, looks like it's just something for everyone, however people would be less opposed to such devices. One day the AI might be good enough for that guy. Also, please take into consideration, that more people being interested in parts of the technology might support some development. Not that they need to come here to the board, but they would still work with it and might make the results public. >>9788 >women do not have any say in the project whatsoever That's something we totally agree. >starts with not giving them jobs Job contracts or paying them are two different things. >idea of letting women in I already stated very early that I didn't plan to bring them into this board. >support a company in making robowaifus Also something no one planned to do. However, even if this was a commercial service, then what would be the difference to using Google and other services until we can do something better? Or maybe always using some service in some edge cases, for example commercial databases or IBM Watson, to enhance our waifus intelligence? The idea presented here can be done much more decentralized and with free software. Also, companies will use the open source hardware to build them and make alterations. To me the goal is to keep it possible for individuals to do it on their own, have small companies which do some work for such persons, have small companies compete with each other building their own full robowaifu versions, hopefully also companies specializing in parts which can be bought.

Alright Anons, it's been around 3 days since I first asked everyone here to discuss this topic. So far it's been going well, and you all have my thanks for it. I appreciate the fact that there are different viewpoints, and that everyone is being civil about the discourse as rational men should. >Also, related crossposts external to these two threads >>9653 >>9654 >>9656 >>9791 I'd just like to clarify one thing a bit more for us all: This council isn't to get together and brainstorm ideas about how to implement this TITS Robowaifu industry; that's what the OP's thread is for. Rather, it's >"to have all of us come together here and discuss it, pro and con." And specifically, to debate whether A) We should even entertain this idea here on our board or not. After all, it certainly significantly broadens the scope of our basic original mission here on /robowaifu/. This isn't just some 'minor enhancement' to the idea. It represents a sea-change. B) If so, then how to address overcoming the many difficult threats & other issues this effort would pose overall. These considerations should cover these not only for the men who are the implementors/operators of any such proposed TITS Robowaifu service bureau, but also (and especially, particularly) for the men utilizing such a service. Until we have at least some good answers to the basic safety, security, and privacy issues involved, I'm quite hesitant to proceed with it. Namely (and these points are only at the least, I'm sure there are others): -How do we effectively guarantee that a TITS Robowaifu won't be compromised in some fashion through this system and do physical violence to -- possibly even to murder of -- the men using them? I'm not LARPing here, I really mean it. Even a child can wield a weapon and kill an unsuspecting man. How much moreso a scheming, adult woman? -How do we ensure that a user of a TITS Robowaifu will not have their personal security or privacy compromised? After all, you're basically inviting a woman into their private domains even if only indirectly. We all know that is automatically a basic threat to the well-being of any man. If he chooses to engage a woman IRL directly, that's his own affair ofc. His own calculated risk. But, if he just wants to engage with a TITS Robowaifu, he should be able to do so very safely IMO. -At a deeper level of consideration of the last point; How do we ensure LEA, Globohomo Big Tech, or any other malevolent actors don't usurp the TITS systems and surreptitiously (or otherwise) gain control of one of these robowaifus? Further, since it's effectively impossible in a warfare-type situation to prevent all enemy actions, how do we provide effective kill switches to completely disable any/all of these thus-compromised 'ghosted-system' TITS Robowaifus? How do we even detect such a compromise in the first place? -Somewhat related, how to afford some deterministic level of protection against psychological 'social-engineering' type schemes against the users of TITS Robowaifus? Obviously, this is the number one threat to a man by a real woman, but LEA/Globohomo/Other bad actors could engage in these same kinds of evil as well -- and obviously will attempt to do so. As I said, I'm sure there are plenty of other thorny issues to address as well, but these basic ones are probably sufficient for us to make a GO/NO-GO decision within the week. I'd suggest we focus on just these fundamental points from here out in this council. The other broad range of issues can be addressed more fully within the main thread, should we decide to proceed with TITS Robowaifu development here. Remember, /robowaifu/ was formed in the first place to serve the interests of disenfranchised men, not any other interests. We must stay true to that singular focus if we expect to continue receiving God's blessings on us here.

>>9805 >significantly broadens the scope of our basic original mission here As OP of the TITS thread I already disagree here. I still don't get the emotions about it. Considering it's chat and vision, compared to external services like Google. >that a TITS Robowaifu won't be compromised in some fashion through this system and do physical violence... I thought I already did that. If they can only send text with some emojis, and the robot itself checks for persons and other stuff, then they can't do anything malicious. It's not so different from general security of robowaifus. They should check what they do, before and while moving, not move with to much force, etc >>1671 >personal security or privacy compromised For the IP, using Tor. Ideally only sending texts, even for objects which are detected. Though this would also create some limits what the woman could see, when she should see it. >don't usurp the TITS systems and surreptitiously (or otherwise) gain control How? And why isn't this a problem with any external service? >How do we even detect such a compromise in the first place? C'mon now we have to solve all hypothetical security threads in four days? This is just a general computer security issue. It depends on the system, down to the details. Which we don't have. >psychological 'social-engineering' type schemes Filters? Idk, example please? I'll elaborate on these points later. No time now.

>>9812 Apologies Anon. Like you, I made further edits/reposts. This is the 'Final Form'. >>9811 >=== -swap links

>>9805 >>9812 >>significantly broadens the scope of our basic original mission here Okay, on some level it is something very different of course. A system of it's own, vs something where the thinking happens to a huge extend outside and some outside personality might bleed into it. Though ideally this would be limited by role playing. Though, this is outside the focus we were meant to keep, so I won't go deeper into that. Just one more thing: It's not so that I personally long for TITS. It's just intriguing that it might be relatively easy to implement, and compared to AI it might give us the option of reasonable conversations and interactions 10 years earlier. With all the consequences coming from it, in terms what I wrote on "Normalization" >>9797 but also in terms of making men interested and keeping them. >do physical violence If we can't get them move securely, then we have to limit it to chat. Very simple. Though this also goes for using them without connection to the internet. If they're strong and move in dangerous ways, then this is something that needs to be solved. the main risk I could see otherwise, is that the women would know the computers they are (indirectly) connected to are robowaifus. While just connecting some robotwaifu to online services doesn't indicate what this computer does. Still mainly a general security issue, though. >>1671 Basically there needs to be layers of soft- and hardware between the sensors of the waifu and the system running the chat client and sending out the messages including descriptions of the environment. I honestly don't understand this here: >I'd just like to clarify one thing a bit more for us all: This council isn't to get together and brainstorm ideas about how to implement this TITS Robowaifu industry; that's what the OP's thread is for. Rather, it's >>"to have all of us come together here and discuss it, pro and con." The only cons I read about are security risks. Those dependent on how it is implemented, and how careful the users are.

>>9811 >>9826 I get the feeling that robowaifus and teleoperation/T.I.T.S will kinda develop in tandem. The technologies are practically joined at the hip. I mean, you can't really have robots but expect to ban powered exoskeletons. So if more people make themselves a robowaifu or general robotic companion, it's bound to happen that someone decides - "You know, this would be more fun if I could actually take control of my robot!" In a similar way to how webcams and motion capture led to stuff like Facerig, Vtubers, and all of these mobile apps that let you replace your own head with an avatar. I don't get too upset or frustrated over it...just see it as an inevitable consequence of improving technology.

>>9797 >Job contracts or paying them are two different things. Not really in this context. >However, even if this was a commercial service, then what would be the difference to using Google and other services until we can do something better? Or maybe always using some service in some edge cases, for example commercial databases or IBM Watson, to enhance our waifus intelligence? The idea presented here can be done much more decentralized and with free software. Using different commercial services means that the market isn't fully centralized, which in turn means that the companies developing this tech will have les bargaining power.

>>9835 >>Job contracts or paying them are two different things. >Not really in this context. I have no idea what you're trying to say. >different commercial services means that the market isn't fully centralized Yeah, and using free software which can be installed on a personal server, is even less centralized.

TITS OP here, just wanted to state that I saved this site and the TITS thread. So I don't care that much. I think this is going to come anyways, though I'm even not so much longing for it. I might or might not get into developing something like it, but I have other things to do anyways. Also, this doesn't need to be discussed here, if it concerns to many fellow developers and especially the board owner, which probably puts a lot of work into this. As a possible conclusion, I want to state: I think there are two version which make sense: - Dolls, or partially animated doll-like bots, which don't need to respond like a human all the time. These dolls or sexbots would either get their system physically limited or turned of, or get stored somewhere, when not in use. So these would be primarily sexbots aka lovebots. They might still be sitting around like a doll without voice after use, or being stored away. Since they don't have their own generated voice, or the users doesn't care about the shift, they could use the (probably altered) voice of the remote woman. She would get paid by the session or something like that, but not being with one customer all the time. More likely to live in a place with decent infrastructure and would need to speak the same language sufficiently. -The other version with is less obvious, is what I thought and talked about here mostly. The women would not get their voice routed through to the user, but either using speech recognition (unlikely) or more likely typing their responses. Their responses would blend into the system, without showing much of their personality. The robowaifu system would respond on it's own, making suggestions for responses which could be selected or ideally also being altered by the remote woman. She could also write her own responses, which would be filtered as good as possible for out of line, out of character, or malicious attempts. These women would be more likely poor and maybe even not speaking the same language as the customer (very well), but the system would translate the responses. The goal would be to have them around as much as possible to alleviate the mind of the robowaifu. So per hour they should get paid much less, but would also be supposed to work longer with one customer. Of course these variants could be mixed somewhat, but these are the most distinct versions I could think of.

OK. So, I've been reluctant to weigh in on this topic so far, much beyond simply introducing it, and highlighting the most significant issues with the idea as I see them. My initial desire was to get a clear view across the entire group instead of expressing my own. Over the next few days -- and until we pass the deadline for council debate -- I plan to respond individually. On that note, I'll (re)post here the TITS OP (>>9709) re-stated as I understand it, in an effort to help consolidate it just a bit and begin the categorization breakdown for us all. At this stage, I'm simply attempting this effort without making much by way of expressing my own viewpoints on the matter. At this stage. :^) Desc/Misc/Etc: -TITS thots participate in the control of a robowaifu. >this is distinct from an AI-only robowaifu -the ones (robowaifuists?) interested in this topic should come up with a good vocabulary for it -trans-humanism (ie, the AI implanted inside a woman) is not within this project's scope Generally Pro: -splits robowaifu opposition into two groups; those entirely opposed, and those hoping to profit from the sexbot industry >it would also reduce overall resistance to robowaifus generally -using TITS as chatbots for customer support across many fields; the hotel industry, for example >already at least one robot cafe in Japan, where handicapped people control the robots from home -a 'drop-in' replacement for (currently less-viable, conversationally) AI-based chatbots -on-the-fly language translation for different languages is a possibility at some point -voice-changers are already feasible -some men might be more open to a TITS robowaifu than an AI-only one -thot's contributions into the TITS robowaifu would be filtered, constrained and altered >for example: questions for personal information get filtered out. there would be masks for personal data of all sorts (for example, location) -thot's responses are a commodity service which they provide, the thots themselves are interchangeable -the TITS robowaifu owners and the remote thots shouldn't know each other -thots can be remote (the Philippines for example) to do this work, no need to travel -conversely, men wouldn't need to travel either -thots being remote-only reduces the chances of them criminally exploiting men (for example, extorting money). -thots would not be able to control the robot however they want (limited control interfaces) >only text would be exchanged. No audio, no pictures. >exchange descriptions of the changes in the environment -women of any stripe can (and should) be excluded from involvement in the design, management, operation, etc., of TITS systems or decisions Generally Con: -letting thots participate in the upcoming sexbot industry by giving them jobs -safety & security needs would be crucial. >only sanitized inputs of text should go through from the remote TITS thot to the TITS robowaifu. -some mechanism needed for 'normalizing' the user's robowaifu personality experience across different, commoditized, TITS thots >--- I hope to have others contribute their own views/interpretations of this consolidation, not the least of which is the OP's own comments. So, let's all please dig in more earnestly into this topic over the next few days /robowaifu/.

>>9723 >I'd like to discuss how this could be made working in a save and useful way. Fair enough. I think as a board we've demonstrated the willingness to do that. >It's not about "us putting them in our waifus". >It's not about pure robowaifu with AI vs robowaifu with thot on board, I personally consider these unavoidable consequences of this topic. >but vs dating Yes, LDR may in fact turn out to be a deciding factor here. It's actually a less complex usage scenario, and one that could prove quite popular. >adopting the language and thinking of the (woke) left, for the purpose of tricking them. Or maybe even, use that as a trojan horse to change the culture of the board here. Or whatever other plot you might think of. As a board, I think you'll find we literally couldn't care less about our enemies. I don't concern myself with swaying them in one way or the other. Nor would it be profitable to attempt, as they are blatantly committed to evil AFAICT. It would be wasteful folly even to try IMO. >I think I made clear that I don't plan to invite women or normie customers of such a system to come here to participate, or to even use it as a support forum. That's fine, but my guess is that you may be mistaking some anon's here objections Anon. 'Letting them in' isn't in reference to introducing anyone to /robowaifu/ (you're quite welcome to under any circumstance as far as I'm concerned in the matter), but rather it's about letting 3DPD infest our beloved robowaifus as an anon complained. That is, it's a philosophical matter of the heart, not a pozzed vs. sane social dynamic. Make sense?

>>9724 >Why only send text back and forth? To my understanding, the primary issue is safety and security. Limit what a thot can manage to do with the robowaifu shell located inside your room, and you reduce the risk of said thot foisting any evil on you thereby. >...an industry [using] girls in VR tracking I'm sure the obvious comparisons with TITS Robowaifus will be rife. >The lack of audio and video is not [effective] My presumption is that basically any audio/video capabilities your TITS-enabled robowaifu has normally, would potentially be available for use by the remote thot, including it's synthesized voice. >and there wouldn't be a market for women to participate in. That's certainly yet to be seen, Anon. >by the time something like this could take off there's going to be advanced contextual AI assistants on everyone's phone and computer. I think you are correct. But I'd say that's still a very long shot to reach human-level cognition and responses within that time frame. More like 10-15 years would be my rough estimate. >It'd be a pretty dumb idea for someone to pay $2000 for a robowaifu because he can't or doesn't want to be with women then paying another $20/hr to hire some escort to operate it Heh, men have shown an unbelievable degree of 'dumb' through the ages in regards to women & sex Anon. I think that would be very likely to be the low end of expenditures men might go to with a TITS system. >Lastly, trying to limit the meaning and power of women from the design process is just going to paint a target on your back and give firepower to radical feminists to wreck whatever you're creating Let the games begin. >and kick out talent that could help. Any man who can be 'kicked out' in such a fashion, should. >Not all women are far left cultists or far right puritans. There are women who want their own robogirls as roommates 3DPD certainly won't be the target audience. But as the old salesman saying goes, 'Their money spends just the same.'

>>9725 >"Thots In The Shell = TITS". Genius! I wonder if Masamune Shirow would approve? >I almost spat my drink out all over my keyboard when I first read that one. Kekd. Be careful with 'our' dear Sophie, Anon! :^) >I wouldn't mind letting some female stick her arms and hands inside a couple of teleoperation gloves and interact with/pretend to be Sophie for a while. As long as it's a female you already know and love, right? Hasn't the possibility of personal injury from Sophie's strong gearing already come up during your design efforts Anon? >Mind you, to do this properly (and not just hire some woman to play with a toy robot arm XD) the equipment appears to be way out of my price and skill range. It would take a dedicated, long-term effort to pull off. But technically-speaking, it's already feasible today. We would see the prices drop with our focused efforts I'm sure. >=== -minor prose edit

>>9747 >If it's text to speech and speech to text...I could actually do that now! I believe you could Anon. >Would have to hire an escort and show her how the speech program works, then whatever she types in, Sophie will say. I imagine they would think it a very bizarre situation LOL. Being hired to communicate through a robot. But then I've read those girls get asked to do all sorts of weird things. I'm sure that bringing Sophie a little more 'to life' would be one of the lesser evils those thots invite into their lives daily. >Realistically though, I'd be too embarrassed to hire a girl in front of my family, and even if it was all done discreetly and remotely... I'd probably just start talking to the escort herself. Though she might find interacting with Sophie and asking her questions a bit of fun! Sounds like you're enthusiastic about the idea Anon. And TBH, I doubt many young men flaunt the fact they are hiring thots up over the family dinner table. :^) >Meantime, I may just take Sophie to a local church if they will permit it on an events day. Should provide a good opportunity for other people to interact with her and I would get to have some interesting conversations without paying as much, and without the awkwardness of hiring an escort LOL. I think giving the general public the practical ability to LARP as a robowaifu on-the-spot would prove to be wildly popular, actually. >Hey you never know, if I make her good enough maybe she can become a robowaifu choir singer? Taking the church into the 21st century in unexpected ways! 😂 Agreed. The idea of creating a 'Christ-chan' AI pluggable-module for robowaifus has already been mentioned here. I'd say if God literally can chastise a wayward prophet through the mouth of an ass, then He can use even our modest efforts here on /robowaifu/ for good! >=== -add 'wayward prophet' comment

I've been sort of pro TITS and almost missed this discussion. But I also understand the risks. Some of my observations: *Vtubers is a huge market and will remain a major force. TITS robots adds the extra dimension and gives us a chance to tap into that market and redirect some of the cashflow our way. We just need to ensure that the robots aren't dependent on a "superstar voice actress" but instead on an anonymous bank of service workers. Poor women make excellent employees. If you're the boss, you're the alpha. They rarely act up if it's a professional setting. Hence why almost all our electronics have been soldered by women in China for decades now. If we hire women, why not hire bros as well? There's a difference between a remotely operated bot compared to your regular degenerate. I'm pretty sure bros know best what guys consider cute or sexy, we've been fapping to male-drawn comics and stories for years. A male operator need not necessarily need a voice changer, he can be simply hitting the prerecorded "Onii-chan asa da yo" standard greeting button. The behavoir of the bot will then seem more tomboyish, which is actually attractive to many men. Remote control and autonomy are two sides of the same coin and go hand in hand whenever we talk about robotics. Battlebots are RC robots without AI, but are still robots. You actually need the remote control in case the autonomous function goes haywire, and vice versa. Ideally the kill switch is local, but what if the user is disabled in some way? You don't want the robot remote control to be hosted by Globohomo Cloud, but you still need a trusted means to diagnose a remote machine and shut it down if necessary.

>>9870 >Profit There's no profit margin, or only a tiny one, if done correctly. It's expensive already, and if the goal is to avoid bonding, there's no reason to pay too much. Also, if the software is free (OS), then there will be no monopoly or oligopoly. Otherwise, increasing profit would only work by men paying for bonding or stars, which some might call exploitation. >If we hire women, why not hire bros as well? That's the second or third worst issue to bring up. Certainly not helping the idea here on the board, nor with the marketing. Also OT. Maybe a check could make that harder, if unintended, which I assume automatically. Woman would need to sit in front of a webcam, so the system would make face recognition. >kill switch It's already rather BS. Because it would be to late anyways. >remote kill switch What? How? Why? No.

>>9740 Pardon me OP, but as before, I'm going to consolidate some of your points as I understand them. >text needs less bandwidth, which has a lot of implications >thots only need to be able to type fast >also connection can be more intermittent, while still being good enough to get some text through >text can be filtered easier than language >masking personal data, detecting the attempt to break the rules [is easier using text-based filtering rules] >automatic translations will work better in textual forms Indeed, these are all net positives on the 'balance sheet' in favor of text-only communications mode. >sound carries background information with it, sending text only is a filter in itself Certainly true. Video is the obvious first consideration for the majority of safety/security issues. But audio is also an important one, as you correctly point out. >ideally we'll be able to use poor women in poor regions [as TITS thots] Agreed. Not only would this likely actually improve the 'service' provided, it would also reduce the overall thorny issues in total. >my thinking is more that the TITS robowaifus would have stored patterns of movements which could be called by putting emojis or some commands into the text. That seems a reasonable notion. Actually, anything that would reduce the interactions (of any form) needed from the TITS thots themselves would likely improve the overall experience for the TITS robowaifu end-user. This includes reducing the amount of text the thots need to enter. >the TITS robowaifu would use a synthesized voice, the thot's voice can't be heard That's already pretty much a given, with the operational communication mode suggested being text-only. >having no long-term memory is going to be tricky >the operational mode of using commoditized thots for a single TITS robowaifu across different sessions >text analysis and storing info in a database. Yes, 'normalizing' the TITS robowaifu's personality across multiple thots would certainly be tricky. Actually, I think machine learning could go a long way here. >also, I had the whole idea primarily with conversations in mind >improving the conversations with a [TITS] robowaifu is a usecase in itself That would certainly prove to be a more wholesome approach for this whole idea in general. It obviously wouldn't stay limited to just that, however. >i couldn't resist making this nice headline Very cheeky. It's a good one. :^) >maybe some chatservers inside Tor and both sides downloading some free (open source) software, and getting some crypto cash wallet. >for figuring out the details, I opened the thread. There would be a veritable plethora of details involved in solving the many technical needs -- quite apart from the business/social ones. >[TITS] i guess Americans and maybe Brits think more automatically about that. I reckon you may be correct on that one Anon. :^) >the women would't use gloves and [would instead] have free control (?? I presume you mean hands free for typing instead here?) Actually, they could. An entire VR setup for the remote thots is a possibility. As long as all the interactions back and forth are text-only, then the list of risks are somewhat mitigated, as already mentioned. >the system would follow some stored pattern, while checking for security, maybe making adjustments dependent on the situation. I think anons may be grossly underestimating the difficulty of performing all of these things on-the-fly with commodity hardware. This would be a 'very tall order' to pull off correctly, especially the vague 'dependent on the situation' bit. >so don't focus on movements and such too much [it's more about the conversations] >speed of the response might be one of the points of failure. >maybe it would be better to always use fast generated responses in combination with some typed response. Certainly it's true that significantly limiting the complexity of the interaction scenarios (by any means possible) would greatly increase the likelihood of success for the project. I hope I've correctly analyzed and consolidated your several points, OP. Feel free to correct me on any points, of course. >=== -minor prose edit -minor grammatical edit

>>9750 >we need chat software, some messenger, that can take the responses and passes them on. Probably some XMPP chat client for the shell, or some libriaries to build something like it. >then adding layers to filter the responses. I believe those would need to be tightly integrated to have any hope of a realtime-responsive system. >Autocomplete, Autocorrect >Suggested responses by software, which can be selected with a key combo and optionally altered before sending Pretty much just like editing text on a typical goyphone, then? >System to create parameters for code, for how to move, based on pose estimation from video. Not impossible, but not simple either. Fortunately, such code should be able to basically stay entirely up on the GPU (and therefore run quickly). >Execution within safe perimeters. Not sure I'm following you exactly there Anon. Do you have more details to add on that point? >We might be chasing some myth here anyways... Personally, I'm already quite convinced this is going to happen eventually, whether /robowaifu/ is involved or not. >>9751 >Robot Cafe with robots which are (to some extend) remote controled Certainly TITS robowaifus could fill in such environments. A wide range of industries are possibilities, actually. Hotels and other face-to-face service industries for example. >>9752 >*The board software replaced the name of the github user. WTF?!? This has been a running joke here. It's sitewide and has been going on since Julay days.

>>9783 >I think this is a very bad idea. Trust me, I completely understand that sentiment. I personally think this idea is fraught with many, many issues. Some of which have the potential to be quite serious -- even life-threatening -- for the men involved with any such TITS-controlled robowaifus. However, it's also obvious that the will is there for some of us, and it's also not impossible the difficult challenges could eventually be overcome. The entire point of this council however, is to judge the consensus here and gauge whether /robowaifu/ should be involved. Clearly you are opposed. >Unless I misunderstand the intent, this is like a short term thing until a true self perpetuating conversation bot is programmed. My impression of OP's intent is for a permanent TITS industry for thots & robowaifuists. He can correct me if I'm wrong on this. >I have NO DOUBTS that many potential men would quickly run to their own personal "vtuber" tier personality in a bot. Certainly 'personality selection' has important ramifications for robowaifus in general, quite apart from any TITS considerations. >Once the "TITS" operator builds enough trust, the user may relax any security features in place, under the assumption for a more "human to human" touch interaction. This is a very serious safety & security threat. Becoming lax over time with such protocols is a very common occurrence across a wide range of industries. Ignoring this phenomenon puts men's safety at peril IMO. >This could be HIGHLY disastrous. As it is just another form of enslavement, Care to spell this out in more detail Anon? How so? >and also ANY remote control capabilities will almost certainly be abused. Certainly, there are no honest women. It's simply a matter of judging the risks versus the rewards, as any man does throughout his life. BTW, this point isn't anything new Anon. Women have always been abusive in this regard throughout all human history. A TITS robowaifu simply potentially amplifies the basic dishonesty issue all women innately possess. One rather important (and quite interesting) aspect of this proposal is that some of this innate evil by women actually could be controlled to a degree through this tech. Far more easily than keeping any typical IRL woman under control today, actually. >FUCK, now that this idea is out, 100% someone is going to abuse it This idea has been 'out' literally for centuries Anon. It's only in the recent past that the convergence of all necessaries is now making this inevitable. But you are correct; this will happen (and it's not too far off actually). >[This will be abused by VTubers] Inevitably. >Hmm, maybe we can control this first, write it up and host it, a platform to allow "TITS" to pretend to be robots as a chatbot anime girl avatar. Certainly there's at least the potential for us to slightly ameliorate some of the coming societal evils surrounding this technology, if we had 'control' of it our own brand of it, actually. >This will backfire eventually as we artificial personalities will eventually be able to out perform a "real" woman, but for now to fund bot R&D this might be a good option. Both true. I predict that AI will quickly become better than we expect it to. And also, there will be huge amounts of money made with TITS-like telerobo systems to come. >Many women that haven't got into the onlyfans thing yet is because they are ugly or haven't went full clown-world degen yet. This would allow the "ugly" ones to possibly wear a "cute" avatar and attempt to make a connection with a man under the guise of being a chatbot. Certainly true. In fact it's already going on now in the VRchat type world. Many fuglies can get away with this as long as their faces don't need to be seen. >=== -add 'control is possible' para -minor prose edit

>>9880 >>Movement patterns >especially the vague 'dependent on the situation' bit. Well, FOLLOW could require that the anon disappeared or is moving away. SIT would require a chair behind her. Some such things could be programmed. >>9881 >>Suggesting words based on typed text >Pretty much just like editing text on a typical goyphone, then? Yes, like on a smartphone. >Execution within safe perimeters. Some movements only if anon is on distance. Not making gestures with to much force. Not closing arms when hugging to much, not using so much force that a human couldn't resist.... >>9882 >intent... permanent TITS industry for thots & robowaifuists Ahm, not really. Depends what you mean by permanent. While AI progresses it should shrink again.

>>9812 >C'mon now we have to solve all hypothetical security threads in four days? No, we don't have to 'solve' all threats, regardless of timeframe. We simply need to decide whether we can reasonably attempt to do so as a group here -- and if we even want to. But fair enough Anon. I haven't even been able to keep up with my own personal backlog. The week-long council was simply to accommodate yourself, by not stretching the debate out unduly and keeping your thread in limbo. Accordingly, I'm going to go ahead and extend the discussion for another week until Sunday night the 25th.

Actual TITS spotted.

>>9914 Haha, neat. Too bad they didn't have the balls to use a clear graphics of a big pair of honkers as their company livery/logo! :^) I'm pretty sure we could already find at least a couple dozen ATM if we cared to dig around. For example Disney Worlds and various large-scale hotels around Asia.

>>9914 >idea: lets put 3rd-worlders in charge of remote-controlled mobile vehicles in prosperous Western nations <what could possibly go wrong? It's all fun and games until the Colombians get hold of some bad drugs. Also, notice the wording on the title that self-righteously just has to point out the income level? Typical leftist type of manipulation and misrepresentation (US$2/hr for sitting on your ass with a joystick in hand is probably the greatest.gig.ever. for a dirt-poor Colombian) and TOP LOL just wait till they get wind of this industry. >What!?!11 TITS Robowaifus!11???11111!111!!!1 DDDD-:<<<<<<<<<<<<<<<<< The neolib screeching will be audible out across the furthest reaches of space. :^) It also shows just how prone to a globalist kike type business model this entire notion becomes. In our specific case that would mean the 'exploitation' of shuffling jobs around with very little friction across 3DPD (or worse) thots (and therefore generating little complaint from most of the anons here). Still, don't be surprised at all when the globalist kikes themselves are suddenly the biggest competition in this entire TITS arena. >tl;dr Mamoru Oshii was actually quite prescient in GiTS2:Innocence where all this is going to end up.

We seem not to have a thread on which hardware to use as a tool outside, or a general thread on tools. So I put this here. Guides to which hardware to buy for doing deep learning: https://timdettmers.com/2020/09/07/which-gpu-for-deep-learning/ Older: https://timdettmers.com/2018/12/16/deep-learning-hardware-guide/ https://timdettmers.com/2018/10/17/tpus-vs-gpus-for-transformers-bert/ Also, the GPU shortage will be prolonged into 2023 or so, along some other infos I've got. For that reason I'm going to buy a okayish comouter with onboard GPU erlier (soon) before waiting so long or spending so much.

>>9954 Thanks for pointing that out Anon. Maybe we can figure out what kind of thread to create. Can you specify more details about tools please? >For that reason I'm going to buy a okayish comouter with onboard GPU erlier (soon) before waiting so long or spending so much. It seems an AI-researcher anon here may be in the middle of a big breakthrough for waifus (>>9951) that will allow her to run on toasters.

>>9784 >If it's a one woman to one guy kind of thing this is always going to be more expensive >but these are session with one VTuber and many men. >If one2one sessions exist, then it's certainly expensive All definitely true. >No starts, no company, rater a software based system that allows to hire cheap staff. Hmm. The commoditized, easily-replaceable thots is an obvious given (and benefit) in the proposal itself OP. What's not very clear to me at the moment is the proposed mechanism that would allow this industry to proceed without having a company (of some description or other) involved? Even if monetary profits were literally a zero-motive, how would the TITS thot's work schedules and sessions be allotted to them properly? The connections to an anon's TITS Robowaifus made? >...even not social pressure to participate. True enough. Heh, I think you'd discover that shoe will be on quite another foot if/when this industry appears. >That was a positive outlook for many in the "redpill-comunity" when Project Harmony appeared, because it creates more competition while increasing the importance of technology in that whole business. Yes, that would certainly be a positive dynamic resisting the globohomo's gynocentric-societies ploy being foisted on the world currently. We would certainly expect an unprecedented shitstorm from them over TITS Robowaifus (in fact, over robowaifus in general). >In our case here it would be about chat and situation awareness. Actually, I think for most normal men it will be about the sex, first and foremost. Certainly the typical robowaifuist would be interested in more than just that, but it would still be important to most of us, I think. >and a crucial idea right from the start was that the women involved would not really control the personality, the look, etc, and they are meant to be interchangeable. At once that's both one of the greatest benefits of the entire concept, and also one of the chief points our enemies will be inflamed to attack us over if we do this. >...and we need to do a lot of things involved anyways. TITS Robowaifus certainly amps up the concept 'a lot of things' by quite a bit. Probably by over 9'000, I'd say.

>>9786 >Any form of interaction where the user may become addicted (ie, enslaved) to a woman is bad in my eyes. >We already have a simp problem that is most likely going to be our true main enemy if women/((())) ever decide to really go against [robowaifus]. Both true, in my opinion Anon. >getting women that involved just seems like recipe for disaster. >I feel like history is our patriarchal ancestors screaming at us, that if we "give them an inch, they will take a mile". Historical precedent certainly bears diligent study. Even though we are in a time of human history that is clearly unprecedented in many ways, still, the ancient adage There's nothing new under the sun bears witness in all our hearts. Females haven't changed much in the intervening years since the days of The Preacher, and certainly not for the better. Quite the contrary in fact.

>>9787 >The alternative is that guys might start to go on dates, maybe trying dating apps, just because they can't have a nice chat with their robot. >Which was the reason I came up with this topic, in the first place. By 'go on dates', I presume you are referring to engaging in a TITS session with a remote thot via their robowaifu Anon? Maybe you mean Long-distance relationship (LDR) instead? It's not entirely clear that you weren't referring to other kinds of 'dating' . >Well, not completely random, but anonymous or pseudonymous. >The woman should not have some avatar personality I think this is an important concept about the lack of personality. Obviously that seems entirely contrary to our ingrained consoomer mentality, but see the next point for clarification why this is the right choice. >however, getting rated might be necessary and then also having a bit experience with some customers >to contribute to the robots responses within the robowaifus personality. I certainly can see how some kind of 'rating system' would be needed for the thots themselves for QA (we must earnestly resist any attempt whatsoever being forced into this rating mechanism going the other way, however). But I don't see how that approach actually plays too well with the 'commodification of the thots as a resource' -- at least if that rating is somehow visible to the anons looking for a TITS 'partner' for the moment. If you make the scores visible, then (especially for normalcattle) the temptation will always be 'go for the best one!', and so the ongoing, eternal, rat-race between the sexes continues ... just in a different guise this time. No, the thots themselves must truly be an interchangeable, quickly-replaceable commodity that serves only in narrowly-defined roles and behaviors. Ones that -- by definition -- can be adequately filled by any of them. >=== -add 'right choice' sentence

>>9788 >Women have an in-group bias and they tend to unite (or in this case unionize) against men (the ones who'd be leading the project) Very true. >who naturally also have an out-group bias Also true. >There already have been successful attempts made by feminists to subvert our plans from within the infrastructure of private companies ??? Can you clarify that a little for us Anon? I'm unaware of any companies even attempting what we are here -- apart from RealDolls, and they are going for the consoomer normalcattle market, not ours (namely, literally every man on earth). But still, if you can clarify this it would be appreciated. >and they'll keep demanding more power and inclusion in the machines It would be absolute folly on our parts not to both see this coming in advance and to prepare against it. >It's fundamental for our interests that women do not have any say in the project whatsoever and that starts with not giving them jobs. I take it you are fundamentally opposed entirely to the for-profit notion of TITS Robowaifus then? LDR may be an alternative approach instead, how do you feel about that idea Anon? >This idea of letting women in is the equivalent of playing with the plague, with the difference that there is no cure nor treatment and the prevention methods always tend to fail. Haha, very creative wording. This could be in some ad-copy somewhere I think. :^) >Let's also not forget the nature of robowaifus: they must be open source, >we should try to make a kit of open-source parts that we can use to build our waifus This is absolutely crucial to the entire notion of socially-responsible robowaifuism ofc. Anything less will simply be a perpetuation of the (current, but soon-to-be-far-far-worse) globohomo Big Tech/Gov police state/asylum. >Why should we try to support a company in making robowaifus? >They'd probably try to get some of their own OPSEC in as fast as possible >which means privatizing research and making things closed-source No doubt they in fact will. This TITS-type industry will happen whether /robowaifu/ participates in it or not. I'm quite convinced of this after just a week or so's study (our years of preparation helped with that of course). Having an open, permissive license to our work will enable corpos to exploit our ideas, true. But the fundamental -- absolutely key -- point is that anyone, literally any man in the world whoever or wherever they are, can also do so. And as we've seen with the great example of Blender over the past year or so, once free gets good enough it always wins. We will win this race Anon, I'm convinced of it. >What we need to do is to instead invest and/or develop with companies that have singular and different purposes from which we can then build up to a robowaifu kit. We're working on it Anon. TITS or not, we're working on it! >=== -prose edits

>>9797 >Some additional point. Think about the idea of "Normalization". >Imagine getting some women, which are in along range relationship, to agree to their boyfriends having a robot or doll >which they can use together for intercourse, while not at the same place Yes, I think LDR may turn out to be the killer-app here OP. >The woman could write texts or send audio I'm very cautious about the idea of giving 3DPD free-form access into TITS Robowaifus. The problem-space explodes exponentially when you introduce that. I'm much more comfortable with your basic idea of pre-canned actions. They can always talk across their goyphones while they're 'doing it'. >however people would be less opposed to such devices Very undoubtedly. Again, killer-app. >that more people being interested in parts of the technology might support some development. Unquestionably. >Not that they need to come here to the board, Anyone is entirely welcome to come here as far as I'm concerned Anon. >but they would still work with it and might make the results public. We literally couldn't hold all the things the amount of hate-press-hit-pieces that would inundate the technology. BTW, this whole scenario is going to happen whether we here on /robowaifu/ get involved or not. At the very least we can enjoy popcorn while watching the neolibs have repeated melt-downs over it all! >Job contracts or paying them are two different things. I think I get that, but perhaps you better spell this out more clearly OP? >The idea presented here can be done much more decentralized and with free software. It had better be. Else it will not only be useless for our purposes here on /robowaifu/ , but flat contrary to men's rights globally. This is entirely self-evident given these evil globohomo corpo's current behaviors. >To me the goal is to keep it possible for individuals to do it on their own, have small companies which do some work for such persons, have small companies compete with each other building their own full robowaifu versions Absolutely. On all points. >hopefully also companies specializing in parts which can be bought. Yep, even the robowaifu aftermarket industry will be vibrant, should we here win this race ahead of us all.

>>9954 It's not a good time to buy GPUs. My $400 card I bought 2 years ago is now selling for over $1000, used. If someone absolutely wants one though, save up for one with at least 12 GB if you can. 6 GB can do a lot and make do, especially with gradient checkpointing, but if you're trying to do any serious work it's too limiting.

>>10022 Hopefully this faked 'shortage' will actually drive gamers and AI researchers to all do what a AI-researcher here on the board is doing. Namely, devising data processing approaches that makes things run well enough on toasters instead (>>9951).

>>9992 >>...and we need to do a lot of things involved anyways. >TITS Robowaifus certainly amps up the concept 'a lot of things' by quite a bit. Technically it's one more source of input. Building the market for selecting and managing them might be the biggest challenge. >how would the TITS thot's work schedules and sessions be allotted to them properly? The connections to an anon's TITS Robowaifus made? This is something which really needs more investigation, thinking and discussion. Yeah, the simple way would be guys hiring the women themselves, with some ad somewhere. But I think we'll need some decentralized database with ratings. If one customer is awake and at home, the system at home would hire a remote woman by using that database. Maybe a little bit of the money could be used for headhunters, which go to online job markets and hire some new ones. >>9998 With dates, I meant regular dates. Without TITS, men might go on regular dates. Like SophieAnon decided to consider. Which is his business and decision alone. However, it shows that the development of AI might be going too slow. >>10007 >basic idea of pre-canned actions. They can always talk across their goyphones while they're 'doing it'. I never saw these pre-canned actions as something sperate from chat or voice. Voice inside the waifu doesn't give them more access to the system. It's just a matter of using it with a remote GF or hooker with voice, or using text mixed into the responses from AI and making the women behind more exchangeable. >>10024 >Fake GPU shortage How do you know it's fake? Who else is making that claim? >>9957 Maybe a thread on tools which are not covered by other threads. SBCs have a thread, so do alternative CPUs, workstations to work with are tools, 3D printers have a thread of their own, rotary tools don't, and questions about electric screwdrivers go where? Well maybe it fits in engineering general?

>>10033 >Well maybe it fits in engineering general? Can you spell out explicitly what you mean by the word it? What exactly is 'it', Anon? >inb4 'It depends on what the meaning of the word 'is' is.'

>>10034 Discussing tools in general, even PC workstations, in the thread "engineering general", or a new thread for all tools which don't have a thread on their own?

>>9712 it isn't even a robowaifu then, it's just a remote controlled puppet with the personality of the operator, not one of its own. If you want to interact with a human woman then interact with a human woman.

>>10036 I agree. A major reason that I like robowaifus is that they aren't human or even biological. Humanoid in shape, maybe. But their body and "mind" are both completely different to ours . For me, a mechanical robowaifu operated by her own programs and A.I is the ultimate goal...even if she's not as fast or flexible as humans, and her programming makes her say strange things. As long as her output isn't just complete nonsense, I think it adds a layer of eccentricity...quirkiness.

>>10036 >isn't even a robowaifu then The underlying technology would be the same, except some additional software. >personality of the operator, not one of its own This would depend on the use case: - Long distance relationship (LDR) - Sexworkers with voice pass through - Mind contributor which adds own thoughts to the system, which also runs on some rudimentary AI The last two would also play roles, not share their true personality. In the first case this might be true as well, but that's another story. All of the cases would help to normalize the technology and get some responsive system earlier.

>>10037 Someone made a post on 4chin asking why people prefer robogirls over cyborgs and it made me realize I don't care about interacting with women. As soon as the limbic system is attached the whole appeal is lost for me. Robowaifu will be selfless, graceful and beautiful. They're not in it for money. They're not going to turn on you and say mean shit because it's their last day working or they're having a bad day. And the mistakes a machine does make aren't because of any ulterior motives or a lack of attention. There's a certain innocence to them that no woman could emulate without being a professional actor. It's that purity that untangles my life instead of making it more complicated and it will only continue to improve as the technology gets better.

>>9653 >One more thought on this: Many people already had the idea of teleoperated robots. This might also be worth considering. I suppose this could be considered the point this conversation actually began after all. Not that we haven't discussed teleoperation issues for years here on /robowaifu/, yet your catchy 'headline' about TITS really coined the idea in a clearly different way. >Before we have some kind of AI, we could use an alternative approach: A few hired women could have access to a chat system. The basic point about 'Before we have [effective, conversational] AI' is a fundamental point. ATM we simply have no practical way to do this. D*sney at the least has been doing 'teleoperation' for decades -- obviously long before AI could even be considered in any practical way -- and this same point was true for them. To wit, Humans, even dishonest female humans, are, well, human. Since this yearning for human-like interaction is one of the basic tenets of desiring a robowaifu (for most of us here at least), then this issue certainly isn't a casual, incidental one. Again, it's fundamental. Otherwise, why not simply become an ascetic monk and literally live the desert hermitage lifestyle away from practically all soulish contact apart from God & the animals? The very fact that a literal human mind is 'on the other end of the wire' is an important point. Someday, we may have absolutely effective Turing Test-capable AIs, but I don't that day is today. ATM, a TITS Robowaifu seems likely to be the only effective robowaifu-based approach to this issue of the desire 'human-connectedness'. >The robowaifu would only transfer answers via text, including text that describes what she can see. This could involve emotions in form of facial expressions as well. The technical/social aspects of what would be needed for an effective & safe system are far from being determined at this point. >They wouldn't necessarily know the real identity of the person on the other end. Well, if it's a remote TITS thot (as opposed to anon's GF) on the other end, they damned well better not know the real identity of the Anon. It's irrelevant to us about the reciprocal of that, but that's unlikely to be a safe situation, either. >>9654 >If you wanted to devise some system where they respond to your 'chat system' with pre-canned scripts and then stored that for AI processing offline, then that might be one thing. Some sort of automated, reduced-capacity control mechanism for the TITS thots to use would be important to solve. >>9656 >But it's still a big nope about allowing a v-thot (or worse) inside our robowaifus, into our very homes... Safety & privacy of Anon will be absolutely crucial. I'm far from certain it can be pulled off effectively at this stage. >>9791 >interesting idea and perhaps a bridge to these techs, since a tele-waifu could be used to maintain LDRs for normies. Yes, LDRs might be a key to success here. >However, yeah, the 3d menace is actually what we are trying to avoid vis-a-vis this concept. Can you tell me sort of the depths of that conviction Anon? Is it more a deep, philosophical viewpoint? Or is it rather a basic concern over the inherent deception of all women, and the many threats and issues that poses to an anon using a TITS Robowaifu system? Something else? Any further insights you could give us about your view would be helpful, thanks.

>>9812 >>significantly broadens the scope of our basic original mission here >As OP of the TITS thread I already disagree here. Casually-speaking, I'm wondering if you've come around a bit more to our view on this in the interim since you originally posted this? Just curious. >I still don't get the emotions about it. The causes for 'the emotions about it' are many, varied, and often highly nuanced. I'm guessing that's becoming more apparent to the anons here. Life & Limb could concisely coin my chief concerns in the matter. Other issues are of less significance to myself, personally-speaking. >If they can only send text with some emojis, and the robot itself checks for persons and other stuff, then they can't do anything malicious. I consider the interaction model to still very much be up in the air here. Just for starters, I don't think it's going to be quite as simplistic as all that regarding the TITS thot's proper usage of the system. It's a more complicated problem to solve than it might appear on it's surface. Also, there are a plethora of other technical, safety, & security issues quite apart from the thots that would also need effective, provable, solutions. Networking security and intrusion prevention, for example, are just a couple on the long list. Some of these are intrinsic, basic needs for robowaifus period (force-control, Asimov's Laws-type concerns, etc), so there is some degree of overlap there already. >For the IP, using Tor. Ideally only sending texts, even for objects which are detected. Tor & VPN networking tech may prove essential in the end. >Though this would also create some limits what the woman could see, when she should see it. I consider that artifact a net positive, though I'm unclear exactly whether you do? >How? And why isn't this a problem with any external service? Indeed it is. You seem to be suggesting that since it's commonplace, that it's somehow irrelevant? Obviously that's not the case, so perhaps I'm misunderstanding your basic point here? I already addressed the issue of time and complete solutions here: >>9908 >Idk, example please? "ANON! THEY KNOW!111 PLEASE SEND US$5000 NOW OR THEY'LL KILL ME! HELP ME PLEEEEEASE!111" Anon, have you really never been around a woman before? That kind of deceptive shit it built right in to them. There are literally a million and one other kinds of evil schemes the TITS thots could try to foist on unsuspecting, unguarded, naive anons who just want to chat with and bonk their robowaifus. Some of these schemes would actually be fatal to him. I thought we've talked here on the board already, about the lying, scheming Snake who is actively, constantly, and effectively deceiving women the world over? :^) >=== -add 'complete solutions' crosslink -prose edits

>>9826 >A system of it's own, vs something where the thinking happens to a huge extend outside Even a fully-autonomous mobile gynoid companion, aka a robowaifu, is still quite dependent on numerous external resources. But the TITS Robowaifu idea is quite a different degree of 'external resources dependence'. A whole other level in fact. >and some outside personality might bleed into it. Undoubtedly that will happen. I don't consider it either feasible to prevent, nor particularly desirable to do so. Having a thot who was good at her job have her personality come through during a session would be a net positive overall. The primary issue is controlling any potential damage TITS thots could wreak on Anon and his things. >Though ideally this would be limited by role playing. Though, this is outside the focus we were meant to keep, so I won't go deeper into that. Personally, I'm not so certain there's any particular 'mission' here on /robowaifu/ beyond our humorous little tagline: "Advancing robotics to a point where anime catgrill meidos in tiny miniskirts are a reality." That's it. Anything else is gravy. And, it should be quite apparent to everyone that frequents this place that simply achieving this basic goal is already a ginormous task all by itself. That's a great tagline, BTW. Kudos to the Anon who came up with it! >It's not so that I personally long for TITS. Personally, I suspect that for many men -- certainly for many Anons -- TITS Robowaifus could become completely integrated into their lifestyles, and practically overnight too. >It's just intriguing that it might be relatively easy to implement Heh. Your definition of 'relatively easy' is quite different from mine then, Anon! :^) >and compared to AI it might give us the option of reasonable conversations and interactions 10 years earlier. It absolutely could do so IMO. Accounting for the fact that 'reasonable' is quite a subjective thing. VRChat with thots, and cloud-based AI chat systems are already both things today. Unfortunately, they are also both entirely burdened with very detrimental characteristics, some of which are pure evil. This latter is mostly the result of the globohomo engulfment of the companies involved -- even the entire industries themselves. The technologies involved are fine, if immature at this stage. >If we can't get them move securely, then we have to limit it to chat. Very simple. Agreed, and not unrestricted, free-form chat either (remember privacy and security?) The only chat suitable for a TITS thot to engage in are pre-canned ones. And the reverse direction from Anon to the TITS thot would also have to be highly filtered to prevent some foolish Anon from intentionally actually trying to connect IRL to some abusive thot over in the Philippines, out of loneliness and naivete. The entire issue of the TITS thot's communications is a big mess r/n IMO. As far as whatever an Anon wants to do with his TITS Robowaifu and his GF, that's his own affair and risk ofc. My concern here -- my worry in fact -- is one of this whole notion turning into yet another pathway for men to be victimized, exploited and abused by women. We must not let that happen to the degree we can control it. At the very least not in case of the thots involved with it. If a man wants the benefit & abuse of an IRL 3DPD relationship, that's obviously entirely his own affair. None of our business. But for the poor trusting souls who somehow (consciously or otherwise) presume that simply because it's a robowaifu, that they are somehow automatically entirely safe...well, we need to take thought and action to protect them if we choose to participate in this industry. >If [robowaifus are] strong and move in dangerous ways, then this is something that needs to be solved. Actually, the stronger we can make them, the better. There are numerous reasons why this is so. But, as you correctly point out this both needs to be addressed as a potential threat, and would also be necessary regardless of TITS or not. >the main risk I could see otherwise, is that the women would know the computers they are (indirectly) connected to are robowaifus. Indeed that knowledge very much does pose a risk to Anon REALLY? IT'S A REAL SEX ROBOT!? Can I make it pick up things too? Maybe bring the john Anon a very sharp object to his throat cool drink he can enjoy too?. It would be wonderful if the fact the TITS thots are connecting remotely into a TITS Robowaifu was entirely hidden from them, but at this point I at least can't really envision a workable method to accomplish this that would still allow them to do their jobs effectively. And even if we could somehow manage that, eventually that fact would become widely known by the thots and their pimps regardless. >Basically there needs to be layers of soft- and hardware between the sensors of the waifu and the system running the chat client and sending out the messages including descriptions of the environment. This would be a very complex set of issues to solve correctly and safely. >The only cons I read about are security risks. Trust me, that set of issues alone is more than enough of a concern. There are obviously several others as well. >Those dependent on how it is implemented, >and how careful the users are. <Users <Careful Heh, as already discussed ITT, even if perfect-quality safety, security, operational, communications & hygienic protocols are all understood thoroughly by all end-users of a system as complex as this one (a very tall order indeed), it's a well-known phenomenon across basically all industries that users will eventually become lax with following them over time. It's simply human nature. Since this isn't just some normal guy typing text into some VTuber whore's chat window, but rather an autonomous functional robot actually, physically located inside Anon's house, I don't think all precautions possible on our parts is out of place. In fact I consider it an obligation on our parts if we proceed with the idea. >=== -prose edits

>>9833 >I get the feeling that robowaifus and teleoperation/T.I.T.S will kinda develop in tandem. I do too, Anon. >The technologies are practically joined at the hip. In fact it's quite likely that focusing development specifically for TITS Robowaifus will rapidly advance the entire domain of humanoid robotics in general around the world. Seriously industrial-grade advances. In fact, I now consider this entire dynamic literally inevitable, all else being equal. This specific topic could easily be the subject of many lengthy posts on /robowaifu/ , so I'll simply leave it at that for the moment. >I mean, you can't really have robots but expect to ban powered exoskeletons. Not sure what you're implying here exactly Anon. Do you mean the discussion of TITS Robowaifus potentially being banned on /robowaifu/ itself? That's a decision (or not) on my part solely for the welfare & health of the board and it's members (same as restricting 3DPD hatred or Blackpills-R-Us). We here obviously have basically zero control of what globohomo does with it. But if it becomes entirely inescapable that (our version of) TITS Robowaifus would lead to nothing but evil and abuse of Anons in the end, then we simply should.not.go.there. That is all. >So if more people make themselves a robowaifu or general robotic companion, it's bound to happen that someone decides - "You know, this would be more fun if I could actually take control of my robot!" No doubt. I like driving cars and motorcycles too. But what we're talking about here isn't simply the technology involved, but bigger issues. >>9835 >Using different commercial services means that the market isn't fully centralized, which in turn means that the companies developing this tech will have les bargaining power. I get your point Anon. I also consider that to be an entirely healthy dynamic, in fact. As we've all seen again, again, and again ... Communist-like centralization and control inevitably leads to human suffering and evil. As clear evidence of this I'd simply offer you a single word: Globohomo. I'd think that's sufficient recrimination all by itself. >=== -minor prose edit

>>9841 >TITS OP here, just wanted to state that I saved this site and the TITS thread. Very nice, thanks Anon! If you aren't using it already, I'd recommend you consider building our software from here called BUMP (>>8769). It can make the process pretty much automatic for you. Also, I'd just ask that if anything happens to both myself and this board, that you and others here consider using such archives to help rebuild the board elsewhere. I've given directives about this before (e.g., >>7496), so if push ever comes to shove, I hope you all can manage to keep the board going this way. >I want to state: I think there are two version which make sense: >- Dolls, or partially animated doll-like bots, which don't need to respond like a human all the time. >These dolls or sexbots would either get their system physically limited or turned of, or get stored somewhere, when not in use. Fair enough. I'd suggest though, that this is probably more a 'normalcattle' type of man's approach. I expect that once everything gets good enough with more typical design robowaifus, then the target audience -- Anons -- will practically be in a marriage relationship with their robowaifus. And one that doesn't need an off switch haha! I expect that most of us would be sad whenever their batteries needed recharging, and couldn't wait until they were back on their feet again. :^) >So these would be primarily sexbots aka lovebots. >They might still be sitting around like a doll without voice after use, >or being stored away. Again, normalcattle usage scenarios. >Since they don't have their own generated voice, or the users doesn't care about the shift, they could use the (probably altered) voice of the remote woman. <entirely open-mic for TITS thots Absolutely not, at least in the case of the TITS thots themselves. As already amply discussed ITT, giving these thots unregulated communications and control into a TITS Robowaifu is a recipe for potential disaster to the Anon himself. After all, what does it cost the whore that's involved? What does she risk? She's located at some type of console somewhere, anonymously safe, and thousands of miles away. For the Anon, however, the results could be devastating. Permanently, even. However, as mentioned ITT, what an Anon chooses to do with his own GF is his own affair ofc. >She would get paid by the session or something like that, but not being with one customer all the time. Yes, it's a good idea to prevent 'personalities' from developing among the commoditized thots. Let the VTuber simps chase after that kind of thing. >More likely to live in a place with decent infrastructure and would need to speak the same language sufficiently. That would certainly help, yes. >-The other version with is less obvious, is what I thought and talked about here mostly. >The women would not get their voice routed through to the user, but either using speech recognition (unlikely) or more likely typing their responses. >Their responses would blend into the system, without showing much of their personality. >The robowaifu system would respond on it's own, making suggestions for responses which could be selected or ideally also being altered by the remote woman. There are a wide array of possibilities in this regard Anon. I'm sure it could be worked out fairly directly to provide speech of one sort or other at the TITS Robowaifu's end. >She could also write her own responses, which would be filtered as good as possible for out of line, out of character, or malicious attempts. No, it's simply too risky. However, we definitely will need to solve filtering for Anon's communications, going the other way. (See the post above regarding foolish, naive Anons). >These women would be more likely poor and maybe even not speaking the same language as the customer (very well), but the system would translate the responses. That's actually at the same time a very difficult technical challenge (one even G*ogle isn't up to yet, for instance) and also one that is conceivably doable by our own /robowaifu/ geniuses sometime soon (>>7837). >The goal would be to have them around as much as possible to alleviate the mind of the robowaifu. This is a very complex issue, and one difficult to airily brush aside this way. Obviously, AI has a long way to go (but then again, so do 3DPD haha!) and alternatively, most Anons would have a limited budget and couldn't afford to keep a TITS thot 'on the line' indefinitely. This dynamic will improve in our favor over time of course. >So per hour they should get paid much less, but would also be supposed to work longer with one customer. Both points are clear positives for Anon, and the practical antithesis of VTuber whoredom & simpery. Personally, I like it. >=== -minor prose edit -add bump crosslink

>>10040 Even as the OP of the TITS thread, I agree. But we seem to be a minority. I've been always rather skeptical when Sandman brought his idea up, but others seem to like it. >>10042 >>significantly broadens the scope of our basic original mission here >As OP of the TITS thread I already disagree here. Maybe to some extend. But it's still just some extra input from the outside. All the concerns are rather vague and speculative, others need to be figured out by working on it and trying the system out. Also, I don't see it as a change of our mission here, since the goal doesn't change. It's just a special use case, which might even push the technology in terms of mass production, attention, mass testing, ... I also don't see TITS as some kind of competition, especially not long term, since it will always be more expensive, messier, and less pure, for example. >Networking security and intrusion prevention Things we'll need anyways. >Tor & VPN networking tech may prove essential in the end. Yes, I don't think it would be a reasonable idea to do that without it. Rather Tor than VPNs though. >>Though this would also create some limits what the woman could see, when she should see it. >I consider that artifact a net positive, though I'm unclear exactly whether you do? I meant this to be a challenge. If the remote woman is meant to see something, the system has to detect it first, and to learn to do so first. The users might see that as a limitation. >>How? And why isn't this a problem with any external service? What I meat by that, for example: I won't try to think and discuss that through in a thread which might be gone in a few days or a week. It needs to be dealt with while something is being implemented. It's more of a general problem, not specific to the topic here. >deceptions Maybe try to write some down and find a pattern. I mean beyond the woman. Maybe the system could detect the switch in context and that the "me" is not about the imaginary character. Also, there should't be any emotional attachment anyways. It's either a voice hooker or just a cog in the machine. Then, we could also add more people to it, in that case it might not matter if it were men, which would manually check for bad behavior or what the intent of a message is. Though, that might create some extra delay. Anyways, I don't think the system needs to be secure against every hypothetical attack. They should also have an interest to keep their job, for example. There might always be a way to try something, though. >>10054 >The only chat suitable for a TITS thot to engage in are pre-canned ones Yeah, I think that's the point where I have to call it quits here. I will not go through all these lengthy texts which end in rejecting the idea here anyways. I want to focus on other things again. So don't suspect (fast) answers to more questions on that topic from OP. If I decide to implement it, or take part in it, I'll do it somewhere else, where we can focus on tech without people being scared of everything that might happen or lead to abuse.

>>10040 That's it anon! You said it better than I could! In this 'post-truth', 'zero-trust', nihilistic age where so many humans have hidden agendas...there IS a sort of purity in the logic of machines! Wow, that's motivating! Thanks a lot!

>>9870 >I've been sort of pro TITS and almost missed this discussion. But I also understand the risks. They tell me, "Better late than never!" >But I also understand the risks. Heh, I'm not sure that's even possible yet. Let me FTFY: <But I also understand a few of the risks.* >Vtubers is a huge market and will remain a major force. >TITS robots adds the extra dimension and gives us a chance to tap into that market and redirect some of the cashflow our way. Kind of disgusting, but OK perhaps true. Managing the cash flow -- indeed a business itself -- in a fashion that would actually benefit the researchers here on /robowaifu/ is an important challenge. So, for example, who exactly, is 'us' and 'our' ?? I personally haven't a clue just yet how to manage something like that in the way we would need here. >We just need to ensure that the robots aren't dependent on a "superstar voice actress" but instead on an anonymous bank of service workers. Sure, absolutely. >Poor women make excellent employees. >They rarely act up if it's a professional setting. I wouldn't at all be surprised, in this particular context. >If we hire women, why not hire bros as well? Absolutely not. This is robowaifu, not robohusbando. Literal faggotry isn't welcome or tolerated here. Nor will it be. Choosing otherwise is to intentionally invite leftists, Communism, and other deviant corruptions into our board. Eventually even the globohomo itself. I certainly wouldn't consider any such thing. In fact an anon mentioned here the notion of confirming authentic XX sex of all TITS thots applicants as a screening protocol for employment. This is a rather solid idea IMO, and one I'm entirely OK with. I presume that's clear enough related to my perspective on this topic? >Remote control and autonomy are two sides of the same coin and go hand in hand whenever we talk about robotics. Actually, I think they are more like two sides of the same universe. I can hardly imagine two concepts further apart in our context. They are, roughly speaking, the literal antithesis of each other. The only similarity they share is the 'universe' they share (aka, the robowaifu). >You actually need the remote control in case the autonomous function goes haywire True enough. But as with a number of other topics this subject touches this is external to TITS Robowaifus, but is a general need regardless. >and vice versa. Lol, you probably should clarify that one Anon. >Ideally the kill switch is local, but what if the user is disabled in some way? You don't want the robot remote control to be hosted by Globohomo Cloud, but you still need a trusted means to diagnose a remote machine and shut it down if necessary. I presume this is referring to the above point? To clarify, robowaifus are Anon's private property and his alone. They aren't 'remote machines'. And no one but her master is going to 'shut it down'. If he wants to enact some sort of 'cyanide protocol' for his robowaifu that's very definitely in the His own effing affair category, and none of ours or anyone else's. If he needs any assistance with his robowaifu then he should be able to come to us or another group for help. As with the remote control point above, this set of issues is a general consideration, and basically external to any TITS Robowaifu concerns, and will need to be solved regardless of our decision or activity on that specific topic.

>>9877 >There's no profit margin, or only a tiny one, if done correctly. >It's expensive already, and if the goal is to avoid bonding, there's no reason to pay too much. I'm no economics major, but that seems rational to me on the surface. But the one thing I am sure of here is that as a group we want to accentuate the benefits for Anon, not the thots, not the pimps, not the 'corporation' running a TITS service bureau. Let's keep that point clearly in mind, and the proper balance in all things will become evident for us. TITS Robowaifus are there to serve Anon, not the other way round. >Also, if the software is free (OS), then there will be no monopoly or oligopoly. I would add to that and just point out that the hardware needs to be relatively inexpensive as well (ie, the entire 'robowaifu kits' notion we've promoted here since the beginning). I'm sure other systems will crop up that aren't of our doing, but it we got involved, then prices would always be kept as low as feasible for every aspect. Again, the point here is to serve men, not to exploit them for money. Let the VTuber whores and simps play that game. >Otherwise, increasing profit would only work by men paying for bonding or stars, which some might call exploitation. Exactly. Heh, it's certainly not exploitation of the 3DPD, either! :^) >What? How? Why? No. I like you Anon. You should seriously consider becoming a board volunteer here.

>>9895 >Well, FOLLOW could require that the anon disappeared or is moving away. SIT would require a chair behind her. Some such things could be programmed. I see. Well, I was thinking more about the entire scenario, not the simple control mechanics involved. But yes, those are certainly both doable, and already necessary capabilities, TITS Robowaifus or not. >Some movements only if anon is on distance. Not making gestures with to much force. Not closing arms when hugging to much, not using so much force that a human couldn't resist Yes, as mentioned ITT, safety & security protocols are a general issue. >Ahm, not really. Depends what you mean by permanent. While AI progresses it should shrink again. I see. Hmm, seems to me this will be an industry pretty hard to just 'turn off', unless the AI becomes particularly superlative in relatively short order. Certainly a long-shot IMO.

>>9914 >So it begins... :^) >>9921 The globalists are very definitely going to move into this arena just as soon as they possibly can. >>10033 >Technically it's [just] one more source of input. Honestly, I consider that sentiment rather a large leap at the moment OP. This would be a significant maze of challenging issues to solve. >Building the market for selecting and managing them might be the biggest challenge. Than actually conceiving, designing, prototyping, programming, debugging, manufacturing, & assembling robowaifus? From scratch? I hardly think so. Business is business, and this entire idea isn't a whole lot different of a business model fundamentally than the already successful VThot industry -- at least on the thot's end of things. But building great robowaifus will take real genius, perseverance, and hard work. It won't come easily or cheaply. >This is something which really needs more investigation, thinking and discussion. >Yeah, the simple way would be guys hiring the women themselves, with some ad somewhere. I'm sure the thots and pimps would sort themselves out quite effectively with few issues. The business/networking systems and infrastructure will take more development. >But I think we'll need some decentralized database with ratings. Again, the business technical needs will take some basic skills and money to sort out. >With dates, I meant regular dates. >Without TITS, men might go on regular dates. Like SophieAnon decided to consider. Ahh, I see. Well, that's obviously pretty much outside any considerations for a TITS Robowaifu system unless Anon later decides to plug his GF into one. >However, it shows that the development of AI might be going too slow. I don't know. AI potential seems to be on at least an exponential curve to me Anon. >I never saw these pre-canned actions as something sperate from chat or voice. >Voice inside the waifu doesn't give them more access to the system. Sure, it's always understood that Anon will experience his robowaifu's synthesized voice on his end, TITS or not. >It's just a matter of using it with a remote GF or hooker with voice, or using text mixed into the responses from AI and making the women behind more exchangeable. Again, as long as the thot's voice isn't simply 'open-mic' of course.

>>10036 >it isn't even a robowaifu then, it's just a remote controlled puppet with the personality of the operator, not one of its own. >If you want to interact with a human woman then interact with a human woman. Both fair points Anon. While I have little debate with your points I think most Anons here wouldn't choose that approach. But yes, good points. >>10037 >I agree. A major reason that I like robowaifus is that they aren't human or even biological. Humanoid in shape, maybe. But their body and "mind" are both completely different to ours . For me, a mechanical robowaifu operated by her own programs and A.I is the ultimate goal...even if she's not as fast or flexible as humans, and her programming makes her say strange things. As long as her output isn't just complete nonsense, I think it adds a layer of eccentricity...quirkiness. I agree that's a rather endearing aspect of many of the Chinese cartoon documentaries I've seen on this topic. :^) Obviously, TITS Robowaifus (whether our own or not) are really only going to be for men who are comfortable(-ish) with a 3DPD being anywhere in the loop at all. The rest of us will obviously want to go down the other path for robowaifus. Considering the merits, both pro and con, are what this council is all about. >>10038 >The underlying technology would be the same, except some additional software. Rather more than that, I believe Anon. >This would depend on the use case: >- Long distance relationship (LDR) >- Sexworkers with without voice pass through (FTFY) >- Mind contributor which adds own thoughts to the system, which also runs on some rudimentary AI I think the LDR may be the key in the end. It brings a much-reduced set of the issues involved with thots, and is in fact practically a guaranteed successful business model (aka, no business model at all). The Anon can just assemble his own TITS thot gear, and then plug his GF into it. >All of the cases would help to normalize the technology and get some responsive system earlier. I'd say that's definitely true, in pretty much any reasonably-conceivable outcome to all this. >>10040 >Someone made a post on 4chin asking why people prefer robogirls over cyborgs and it made me realize I don't care about interacting with women. As soon as the limbic system is attached the whole appeal is lost for me. Robowaifu will be selfless, graceful and beautiful. They're not in it for money. They're not going to turn on you and say mean shit because it's their last day working or they're having a bad day. And the mistakes a machine does make aren't because of any ulterior motives or a lack of attention. There's a certain innocence to them that no woman could emulate without being a professional actor. It's that purity that untangles my life instead of making it more complicated and it will only continue to improve as the technology gets better. It is a beautiful thing Anon, and you've worded that quite well. I have little to add except... I'm stealing that Chii! :^) >>10063 >That's it anon! You said it better than I could! In this 'post-truth', 'zero-trust', nihilistic age where so many humans have hidden agendas...there IS a sort of purity in the logic of machines! Wow, that's motivating! Thanks a lot! Seconded. >=== -add cartoon documentaries bad joke

>>10061 >Even as the OP of the TITS thread, I agree. Wait... I am quite confused now, OP. >All the concerns are rather vague and speculative, others need to be figured out by working on it and trying the system out. Actually, I would consider immediate potential threats to life & limb, the possibility of network intrusion and usurpation/hijacking (surreptitiously or not), social-engineering type scams, and the high probability of bad OPSEC on the part of the average Anon to hardly be in the category 'rather vague and speculative', Anon. But you are certainly correct that the need to plan, engineer, and test a TITS Robowaifu system would be necessary to figure things out. And iteratively as well, at least for a while. >Also, I don't see it as a change of our mission here, since the goal doesn't change. It's just a special use case, which might even push the technology in terms of mass production, attention, mass testing, ... I also don't see TITS as some kind of competition, especially not long term, since it will always be more expensive, messier, and less pure, for example. All fair points in my opinion Anon. Not much debate from me there, with the possible exception taken to your use of the word just in this regard. >Things we'll need anyways. Yes, as with several others, it's already a basic robowaifu issue. >Yes, I don't think it would be a reasonable idea to do that without it. Rather Tor than VPNs though. I'd say that would be a TBD item on the checklist. Certainly it wouldn't pass muster without extensive pen testing, etc. >I meant this to be a challenge. If the remote woman is meant to see something, the system has to detect it first, and to learn to do so first. OK, fair enough. Virtual re-construction on-the-fly is pretty much feasible today in a simplified context. >The users might see that as a limitation. Undoubtedly. They'll just have to understand that safety & security comes first whenever a thot is in the mix. Anon and his GF would be rather a different stripe on things. >Maybe try to write some down and find a pattern. I mean beyond the woman. Maybe the system could detect the switch in context and that the "me" is not about the imaginary character. Not exactly an easy problem to solve, but one that's at least conceivable for an AI to tackle successfully reasonably soon. >Then, we could also add more people to it, in that case it might not matter if it were men, Lol, what? Are you suggesting along with the other Anon that sodomites men be involved at the operator end of a TITS connection? I've made my position rather clear on that already, Anon (>>10064). Absolutely not. >which would manually check for bad behavior or what the intent of a message is. Though, that might create some extra delay. That is a technical issue that would have to be resolved, yes. >Anyways, I don't think the system needs to be secure against every hypothetical attack. Obviously, that's a physical impossibility. Security 101. >They should also have an interest to keep their job, for example. Any means that keeps TITS thots in their places and doing their jobs properly will certainly be in their pimp's best interests, yes. >Yeah, I think that's the point where I have to call it quits here Fair enough, I understand. >where we can focus on tech without people being scared of everything that might happen or lead to abuse. I can assure you OP, the word isn't 'fear'. I'd suggest wisdom, accountability, and responsibility are probably better choices. >=== -minor prose edit

And with that, I believe I've cleared the current backlog for this council, /robowaifu/. If I've accidentally overlooked anyone here, please forgive me and point it out to me. I'll adjust the OP to request all final comments before this coming Sunday night's deadline. Cheers /robowaifu/.

Quick question, though it will probably to late when I'll have the answer: If I buy a cheap workstation without a grafics card, should I take Intel or AMD Ryzen CPU if I want to make some small ML models? I think AMD doesn't work at all for that, but well the CPU would be 25-30% faster. Context: I'll build another one on Intel based old server CPU soon, and plan buy another Ryzen gaming/ML rig at some point (after the shortage).

Hey guys, /icup/ here with a question on how we're going to handle the next iteration of the Infinity Cup ( https://anon.cafe/icup/ ) We're trying to poll whether certain boards are interested in playing in the cup, or if there's some specific team that you'd like to see play. If you want to, please answer or add your own answer to the poll in https ://poal.me/6x3j1u

>>9954 >However, typical pre-programmed schedules for fan speeds are badly designed for deep learning programs, so that this temperature threshold is reached within seconds after starting a deep learning program. The result is a decreased performance (0-10%) which can be significant for multiple GPUs (10-25%) where the GPU heat up each other. >Since NVIDIA GPUs are first and foremost gaming GPUs, they are optimized for Windows. You can change the fan schedule with a few clicks in Windows, but not so in Linux, and as most deep learning libraries are written for Linux this is a problem. >The only option under Linux is to use to set a configuration for your Xorg server (Ubuntu) where you set the option “coolbits”. This works very well for a single GPU, but if you have multiple GPUs where some of them are headless, i.e. they have no monitor attached to them, you have to emulate a monitor which is hard and hacky. https://timdettmers.com/2018/12/16/deep-learning-hardware-guide/#Air_Cooling_GPUs

>>10076 https://www.cpubenchmark.net/compare/Intel-Celeron-G5900-vs-Intel-i5-7400-vs-Intel-i3-9100T-vs-Intel-i3-10100-vs-AMD-Ryzen-3-3100/3761vs2929vs3488vs3717vs3715

>>10076 If you're purchasing a CPU for ML you'll get a significant speed up with more cores and threads. The brand of CPU doesn't matter. https://jdhao.github.io/2020/07/06/pytorch_set_num_threads/ The RAM you use is usually the cheapest way to improve performance. If you have multiple computers though, it's possible to use them for training the same model. https://pytorch.org/tutorials/intermediate/dist_tuto.html AMD GPUs haven't had any support until recently. Popular deep learning libraries only fully support Nvidia, although PyTorch recently launched a beta for AMD ROCm. Other machine learning libraries like mlpack support AMD GPUs via OpenCL and AMDGPU-PRO. AMDGPU-PRO appears to be significantly faster than AMD ROCm in vRAM copy and read performance which is pretty essential for deep learning. https://www.phoronix.com/scan.php?page=article&item=rocm-20-pal1850&num=2

>>10088 Thanks, I'm not sure, but from what I gathered the brand of the CPU only doesn't matter if there's a GPU (which should be Nvidia). I tried to find out if it matters if I only have a CPU and run small models, which seems to be used for modeling before running it in some cloud service. Also in the linked example it speeds up four times whe using eight threads. However, now I'm having my eyes on used PCs with grafic cards. Just have to make sure they're still supported (RX 16xx and such).

>>10081 >>10083 >>10088 Thanks for the good information Anon[s].

>>10088 >>10089 >Also in the linked example it speeds up four times whe using eight threads. Ah, sorry. I've read your text wrong, while tired. I read the number of cores doesn't matter, the opposite of what you wrote.

TalkToWaifu Github repository is ded, is there any alternatives?

>>10149 Look in the chat section >>22. Something similar seems to be worked on there. Also, the thread on that topic is here >>7978.

OK, I'll consider our discussion on the topic closed now. Thanks very much to everyone who participated in it. You have all given me enough information to come to what I believe is a good conclusion for the board's benefit.. Give me a day or two, and I'll do basic writeup on my decision.

Alright, we've all waited long enough about this topic. First off, there doesn't seem to be a clear consensus on the idea of TITS Robowaifus. Not really a surprise actually, since the general idea itself is pretty controversial. Obviously, the OP of the thread is in favor as well as some others. Several Anons are not favorable to the idea however. So, I guess any decision is going to be, again, controversial. I'd just ask everyone to remember why we're here and not to be offended about it, either way. Let's stay in concordance with each other here, /robowaifu/. In the end, the things themselves tend to naturally correct and get back on course, as long as good men work together in unity. >--- Rather than dragging everyone through the tedious process of me making an attempt at a full review of every pertinent comment made, I'll simply try to break down what I see as being the chief pros and cons -- which is what this council was all about after all. (Please correct me if you see I've accidentally overlooked your favorite concern & I'll likely update this post). -Pros -This technology clearly has the strong potential to rapidly advance Robowaifu development in general. The challenges of effectively integrating haptics, synthetic character & environment visualization, teleoperations, and the various AI/ML facilities that would all be needed for an effective, safe solution would together serve as a dramatic impetus for overall advancement of robowaifu technology & design. -TITS Robowaifus could truly help facilitate effective LDR (Long-distance relationships). For the men who want to maintain relationships with IRL females, this notion would open up a whole new potential realm for their relationships. -Greed (if that can possibly be called a 'pro'). The bottom line: There's a lot of cash that will be made by utilizing thots around the world as 'sit-ins' for men's Robo-GFs. It's been a very well-known phenomenon that (put simply and crudely) sex sells. I'd recommend everyone watch the initial few minutes of Bob Cringely's Triumph of the Nerds PBS special for his take on this notion. -Cons -In a word: women. It's literally tempting for me to leave this entire section at that one word alone. Simply put, basically every single difficulty related to the notion of TITS Robowaifus is related (either directly or indirectly) in one way or other to the women who would be involved in the project's daily operations. Without doubt there are some bad men out there who could be an evil influence too, but without the women who are involved these men would mostly go elsewhere. -Direct hazards to Anon. Whether it's something as simple as an Anon intentionally doxing himself to some thot, a social-engineering-esque embezzlement scheme intended to rob him blind, or actual physical harm done to him (intentionally or not), Anon takes his fortunes into his own hands with any such system as a TITS Robowaifu. -Revulsion. A common complaint was the distasteful, repellent aspect of having an IRL female involved in the mix in the first place. While /robowaifu/ is very definitely not an MGTOW enclave, it's entirely understandable that many Anons would object at a fundamental level to this issue. The protagonist Hideki in the anime series Chobits dealt with this exact dichotomy and conflict (though I would suggest his entire take on it was quite different). >--- Trying to balance all this out and make a 'good' decision is, as you might imagine, difficult for me in any natural sense. However, I think that the majority of you honestly believe that I care deeply for both the welfare of the board itself, the effective creation of robowaifus around the world, and the benefit to, and protection of, men around the world through robowaifu-related efforts. I hope therefore that, again, Anons not get offended with me or with each other, and let us all work together in unity. >(1 of 2) >=== -minor prose edit

>>10194 My decision for the board is in the AFFIRMATIVE, with two straightforward caveats: 1. Absolutely no free-form, 'open-mic', unconstrained, verbal or physical control by TITS thots of any TITS Robowaifus themselves. The most problematic issues with the whole idea all stem directly from failing to enforce this basic rule. Also, the intricacies of pulling off implementing these restraints correctly, and still allowing for an appealing, effective, and fun engagement for the Anon himself is actually quite a dramatic challenge & achievement. Solving all this will advance many different robowaifu-related areas all together at once. 2. Men will be free to turn off 'safetys' if they desire to plug their IRL GFs into the remote-end of a TITS connection. They are taking their own lives in their hands with such a risk, and they will be clearly informed of that. Note that this is a privately-conducted connection between Anon and his GFs, and isn't in any way associated with any business-oriented systems utilizing professional prostitutes (whether they are labeled as such or not). Basic safetys are not to be disabled in that context whatsoever. As a board-related point, the TITS threads will be considered as 'containment threads', and apart from this council, any further discussion on the topic is to be conducted strictly inside TITS threads. That way, hopefully any Anons who are fundamentally disgusted by the entire concept can simply hide the thread itself, and otherwise continue to have warm relations with the other Anons here. Crossposts linking into the threads from elsewhere are perfectly acceptable, but please keep in mind that some Anons are going to be offended by TITS thots. Please don't 'promote' the idea outside the threads themselves, thank you. I'll leave the TITS thread itself locked for the moment, as I prepare a rewrite of the OP intended to clearly spell out the objective and restraints of our, /robowaifu/-flavored, TITS Robowaifu project. As my final thought for this council; In the end, I now believe it's actually much better for us to be involved that not to be. If we fail to become involved, if we look the other way, then I think it's entirely likely that there won't be a single important voice defending the men themselves in regard to this industry. And that outcome would simply be more of the same evil abuse currently being heaped on all of us by the globohomo. This industry will happen with or without us regardless. I believe we can make a positive difference for good in it, if we simply focus on doing so. >--- Thanks again to everyone who decided to participate in this discussion. I learned a lot from it, and I think this whole thing can help each of us to rapidly advance our own robowaifu R&D. Cheers /robowaifu/. You're a remarkable group of Anons! >(2 of 2) >=== -prose edits

>>10210 Nice subtle piece of work Anon, point taken. I have a few copies of that one myself. However, if you're at all a regular here then I doubt you think I'm some kind of milquetoast pushover for invaders & corrupters here. Being polite to our enemies isn't even one of my faults. Heh, I'm quite ready to call out even friends on their behavior here if I feel it's harming the group overall -- as any sane and rational man would do. They'll all come back to their senses eventually, I believe. In that little fable of yours, the founder's basic error was failure to promptly force the commie usurpers into their proper isolation/quarantine camps. Namely shitposting-containment zones. One of the things neoLibs, glowniggers, goons, et al, all crave and demand is attention. Pull the soapboxes away and their corrupting influence just seems to magically disappear somehow. Honestly, it doesn't really take much wisdom to manage a thing like this place effectively -- just a little diligence, pluck, and cheek. :^)

Price of storage might go up next, though it's not sure to happen and might take a bit: https://youtu.be/np8rgXSdYMM - It's more than a rumor though, since there's a new cryoto currency targeting storage.

>"...We are sitting here watching, in real time, women being stripped of all of their most basic value by trannies and soon by robots and artificial womb technology. " Anglin I know you're lurking here. Why don't you give robowaifus your treatment. I doubt you'd have a positive outlook on it all, but I'd still be interested to read what you have to say.

>>10312 >"...I guess the upside of the tranny revolution is that they are going to be thinner and better looking than women, and thereby force women to compete. Maybe we’ll see a renaissance of femininity in women?" I know you're here Andrew Do a robowaifu revolution hitpiece.

>>10151 >Updated version is still using GPT-2 fugg

>>10326 The team running GPT-3 has chosen not to release that 'open-source' system "to protect the public". But somehow, (((strangely enough))), has managed to take everyone's work into it and monetize that instead. > As I understand it, there are other, non-Globohomo Big Tech/Gov efforts that have sprung up to combat this abuse, and are endeavoring to create effective alternatives.

>>10329 I read from other pages and some have mentioned that GPT-3 requires a NASA super computer to run including multiple GPU to make the performance acceptable, looks like its not such a good option then welp. >As I understand it, there are other, non-Globohomo Big Tech/Gov efforts that have sprung up to combat this abuse, and are endeavoring to create effective alternatives. What are those?

>>10326 >>10330 Correct, GPT-3 needs a whole server farm. Some anon here tries to build something smaller. But, keep in mind these are text generators anyways. It makes sense to feed all data into them, to anticipate what someone might say in response to something. Using it to have some vaguely specific output, requires to select the data going into the training. The responses coming out would still not make much sense, necessarily. We have two treads for those topics, btw: >>22 >>250 - Maybe we should go on there, though I think this also has been discussed already.

>>10329 There's GPT-Neo now that claims to outperform GPT-2: https://github.com/EleutherAI/gpt-neo It's basically a smaller version of GPT-3. I haven't had time to play around with it but it looks promising.

>>10332 >Maybe we should go on there, though I think this also has been discussed already. Good point Anon. I'll just pop over and leave crosslinks so researchers will see our little stub here too. >>10333 >digits confirm Neo That's encouraging Anon, thanks.

1) What's the fastest way to find out what kind of hardware a specific neural network would need to run? For example, if they mention ResNet50 in some video, being used as a teacher for creating as smaller network. 2) What's the fastest way to look for a implementation or reimplementation of some ML model which might not be shared by the authors of some paper (I don't mean just Google it)? Is there maybe some search engine for it? Or is there only HuggingFace and using Google (or Twitter, Youtube, ...)?

>>10434 Sorry, I don't know how to answer that question Anon. I know that some of these questions have answers in one way or another here on the board (eg, h/w req'd), but I don't personally know of a general answer to your questions.

Nice 'Waifus comped over the real world' thread on Smug. https://smuglo.li/cute/res/925.html#925

>>10817 Nice

>>10822 ikr.

>>10824 sorry if its off topic for a meta thread but has anyone noticed that the canary hasnt got updated in a bit? is that something we should be concerned about

The board has some problems after the upgrade: Posting window can't be closed. Posting doesn't show up after posting with the field on top, after uploading pictures, only after refreshing the site. Also, Youtube links are changed into something with youtube-nocookies in it, which doesn't work. >>10994 Not OT, exactly the right place.

>>10994 >sorry if its off topic for a meta thread There are only a couple of off-topics on /robowaifu/, other than that the /meta threads are fair game. >>10994 >>10996 Thanks for pointing these issues out Anons. The text of the email I've sent to Robi follows. >--- AUUUUGHHH! So the new site upgrade has created a couple of issues for us. 1. The QR box won't close after a post now. I can confirm this and it's been triggering my autism. 2. An anon has stated that 'Posting doesn't show up after posting with the field on top, after uploading pictures, only after refreshing the site.' https://alogs.theГунтretort.com/robowaifu/res/8492.html#10996 3. An anon has stated that 'Youtube links are changed into something with youtube-nocookies in it, which doesn't work.' https://alogs.theГунтretort.com/robowaifu/res/8492.html#10996 -I seem to be able to confirm this. At the least, YT embeds don't seem to work for me now. - https://files.catbox.moe/pib1og.jpg Also, an anon pointed out the warrant canary hasn't been updated for a while. https://alogs.theГунтretort.com/robowaifu/res/8492.html#10994 Thanks Robi, Chobitsu

Why have we not seen cheap Chinese knockoffs of Atlas yet like there have been with Spot? CCP keeping them under wraps? It seems like that would be a major catalyst that really gets humanoid robot tech moving forward.

>>11035 >like there have been with Spot? <just for a convenient cross-ref here: (>>1198, >>1199, >>1215, >>8415)

>Smithsonian Open Access initiative releases loads of 3D models to public domain www.slashgear.com/smithsonian-open-access-initiative-releases-loads-of-3d-models-to-public-domain-26611336/ https://collections.si.edu/search/

>>11057 One very interesting area (at least to me, heh) is the so-called Transhumanism movement. I'd broadly categorize the adherents of this into two basic groups: 1. Individuals who desire to prolong/improve their lives through primarily biological means (mostly with gene therapies, such as the overhyped & deadly mRNA 'vaccine'). Possibly with a touch of bionic or other prosthetics-like things thrown in for good measure. 2. Individuals who desire to achieve an 'immortality' through primarily biological means (mostly by shedding their mortal coil and 'uploading' themselves into computer chips). This group is more inclined to embrace a virtual avatar or other symbolic representation as part of this goal. Ironically to me as a Christian, the fact that it's already a transhumanist religion (and a reality as well -- not just a pipedream) seems entirely lost on these people! There's plenty I can say further on this topic if you're interested Anon, but I think those three concepts lay out my basic premise well enough for starters.

>2. Individuals who desire to achieve an 'immortality' through primarily technological means* Oops, copypasta, pls. :^)

>>11061 I am not too fussed about being mortal and dying myself. (Who'd want to have to live sentiently on this ball of shit and nightmares forever?)! I mean, obviously it will suck if the dying is really painful, but there are several efficient ways I have planned for dealing with that when it comes (whaddya know, my time in pharmacy has it's bonuses LOL). I am more interested in creating 'un-life'. Something that isn't capable of suffering and cannot die (because it isn't alive), but can still improve the lives of organics and potentially learn/develop through programming, data analysis and any non-emotional means. Humanoid robots seem to be an increasingly accessible way of achieving this. I've never wanted children who will only be bought into this world for greedy corporations to enslave and torment for the majority of their lives anyway. TLDR; I decided I wanted a machine-friend who is impervious to all the shit we humans have to face in life.

>>11061 I'll consider myself being some kind of Transhumanist, at least one of the very moderate kind. Though, the guys thinking they should focus on becoming a swarm of nanobots, the focus of getting ones head frozen after death till revival, and the internationalism (anti-racism/marxism) of some, turned me away from any "official" assembly. Also, I never read many of the books they were into. I'm more interested in the results of anti-aging science. I plan to do more about that, buying some additives or drugs to slow down my againg process. This channel for example https://youtube.com/c/DrBradStanfield is quite interesting to me, in that regard. There are also tests for trying to find out about ones biological age and then do something about it. David Sinclair, Aubrey de Grey, Senecence, Senolytics are also interesting topics to look into. Another aspect of such people is the approach of being open to psychedelic drugs, other mind altering drugs or techniques, microdosing, meditation, exploration of ones mind, which is often shared by people interested in AI as well.

>>11066 >I am not too fussed about being mortal and dying myself. (Who'd want to have to live sentiently on this ball of shit and nightmares forever?) The endgame is to get off this "ball of shit and nightmares" - compare the earth to a womb, and leaving here we can emerge into our birth and explore (or conquer) the galaxy. The WQ (w*man question) is pivotal to this. While I love my mom, and my daughter and am polite and cordial to almost every femaled human in my life, I honestly believe that they are inherently more tied to nature and the "great filter" is actually their influence on us. One example being the argument against spending on space and tech research when there are "starving 3rd worlders", "LGTBBQ-POC" or whatever the cause of the week is, since w*men, understandably are prioritized for the group and to take care of the weakest and most "oppressed" (spoiler, life has winners and losers, moving everyone who is last to the front of the race doesn't change the fact that there will still be someone who is "last") Further, as I've probably brought up before, the amount of time, physical and psychic energy that men spend to gain the approval of w*men (which many will just siphon off like vampires, encouraging men, "just a little more and you'll have me") is a waste which should instead be channelled toward objectives such as AI, Immortality/longevity and space colonization. The earth is a crab bucket and to get off this rock and become self-sufficient will launch us into a Type-II Kardishev civilization in no time (Dyson Swarms/Spheres are the next step after Space-Arks IMO) I'll admit some of this speak of my own personal shortcomings at life, prioritizing social and sexual conquests and in the process adopting a slacker attitude (probably via osmosis) while missing the big picture that completing my education would have provided me with crucial applicable skills necessary to push us all along toward these ends. Now all I have is an overactive imagination and a few useful quirks like Big-Picture thinking and a near photographic memory but I still struggle with producing anything tangible at the end of the day. Sorry if this is too much of a rant or O/T but I don't see any intro or bio threads so this is as good a place as any.

>>11066 >TLDR; I decided I wanted a machine-friend who is impervious to all the shit we humans have to face in life. Heh, I almost feel like we're brothers on this topic Anon! I too (from at least early grade school) longed to create a mechanical friend that would be a stable and interesting companion in life. This desire ofc led eventually to the realization that robowaifus could both fill this role, and also help ameliorate the evil being done against men in general around the earth. >>11068 Interesting, Anon. >...find out about [] ag[ing] and then do something about it. Yes, I agree that the very notion that aging itself can be treated as if a disease certainly bears scientific & medical investigation. BTW, if you care to understand the Christian perspective on transhumanism, let me recommend Humans 2.0. I both own this book, and have followed these authors extensively for years. Not only does it offer a solid exposition of the topic at large, but also a thoughtful discourse on the underlying social and philosophical longings in the human heart that drive men towards this goal in the first place. > https://support.reasons.org/category/featured-products/humans-20-scientific-philosophical-and-theological-perspectives-on-transhumanism >>11073 Maybe you'd find this book of interest Anon? >The Condition Of Man by Lewis Mumford https://archive.org/details/in.ernet.dli.2015.188937

>>11073 >the "great filter" is actually their [female] influence on us. Fascinating perspective. I always thought humans had some fundamental problem. I was aiming for something wrong with the evolution of organic life...a flaw in DNA itself...or the human brain. But your suggestion is also possible. We have no other species of similar intelligence and sexual dimorphism to compare with...and all other human subspecies are extinct (which in itself is significant). >The earth is a crab bucket and to get off this rock and become self-sufficient I completely agree that in order to advance we would need to leave Earth and explore the cosmos. However, I have serious doubts that organic humans are capable of true (interstellar) space exploration. Hence another reason I like robots so much. I'd like A.I. a lot, too. If I could understand it a bit better. But I intend to. I already have a few books on things like data science and Scikit-learn. Just a matter of many evenings reading and doing maths exercises, making graphs and writing practice programs. >>11087 >I too (from at least early grade school) longed to create a mechanical friend that would be a stable and interesting companion in life I think there are thousands of other people who like the idea of having a robot companion, it's just that they are very challenging to build. Most of them go for InMoov. But the more functional droids there are out there, the more people will become inspired to either try and build a copy, modify the design or even make their own robot companion from scratch. >Humans 2.0 I shall have to give that book a read, anon. It looks right up my alley, thank you!

>>11073 >Sorry if this is too much of a rant or O/T but I don't see any intro or bio threads so this is as good a place as any. No /meta is kind of a grab-bag on /robowaifu/. We do have our embassy thread (>>2823), but that seems rather more of a latenight drive-by scene Hi! I'm Anon from foo.xyz good luck making robowaifus plz hurry it up!11 kind of thing. >tl;dr This is fine.

>>11092 >However, I have serious doubts that organic humans are capable of true (interstellar) space exploration Exploration is one thing, populating it is another. The best way to spread out is to send builder robots first, then send frozen human embryos, robowaifus and external artificial wombs. Have bases underground or space stations, let the first generation be risen by robowaifus. Later, instead of embryos, the genetic data might be sufficient.

>>11114 This sounds like the premise for what could be an absolutely fascinating scifi series, Anon! If it was executed on correctly (for example not introducing any current-year pozz trash) it would probably become wildly popular with young men, and could even have a much broader reach. Hmmm... anyone up for creating a Robowaifu Studios? :^)

>>11092 >I shall have to give that book a read, anon. It looks right up my alley, thank you! Great! You're very welcome Anon. I personally consider Fuz Rana a genius. He helped me to understand molecular biology and the genetic code better than anyone else I've studied (he's a PhD in Chemistry). That whole organization is entirely unique, actually. They have my unreserved recommendation.

>>11114 >let the first generation be risen by robowaifus >>11118 >not introducing any current-year pozz trash There is a series out at the moment called "Raised by Wolves", however the storyline is confusing as fuck and from what I've seen wreaks of SJW shite, so probably one to avoid. Just mentioning it because it's a mainstream series that involves a robowaifu raising her children on another planet. But yeah, I can imagine a centralised A.I. (like "Mother" in the Alien franchise), controlling hundreds of other robots and ship subsystems to help raise a creche of human pioneer children. I doubt it will be actual female-looking robowaifus doing the child-rearing for a long time, but lots of different machine-tools, mobile robots and robot arms. These robot-raised children would likely be the ones to create the first practically useful female-looking robowaifus (to replace the human mothers they never had?)

>>11131 Really interesting scenario you bring up Anon. I would definitely enjoy watching a lengthy & non-pozzed treatment of the whole set of ideas you two mentioned.

>>11118 Had the same thought, but it also should be the more general backround of many SciFi stories. >Robowaifu Studios I might start drawing mangas when I'm old and there's better AI software to support me. It's more important to work towards it in reality. (Personally I fell a bit of the wagon recently, bc I ran into to many frustrating hurdles in different areas.) >>11131 >Raised by Wolves Yeah, this has been mentioned in the entertainment thread. >>8759 >centralised A.I. Why?!? >doubt it will be actual female-looking robowaifus We don't have spaceships but trying to build female looking robowaifus already. So I wonder how you come to such a conclusion. I thought at first of partially very human robots, but for example with wheels instead legs. However, since we're working on it already and the current progress, and also bc babies would prefer to have their mother look more human, I don't think so. That said, helper robots make sense, diaper change and transportation from one place to another for example.

Just in case any of you anons here on /robowaifu/ are fans of Blade Runner like me, here's a heads-up that anons over on /film/ have posted/linked for the 25th anniversary Retirement pro & fan-mixed edition soundtracks. anon.cafe/film/res/1414.html#1609

>>11180 The Westwood game was better, never enjoyed the movie that much. It was just very impressive when it came out, and the visuals are great. My posting about the game, seem to have been deleted. :-|

>>11184 No, found it, it's here: >>7897

>>11184 >>11188 Looks interesting, and I agree about the sentiment of making robowaifus look like young teens (and therefore beautiful). As we've talked out pretty thoroughly in the robowaifus in media thread, the globalists are already making a strong push both to infest the entire domain with stronk, independynt (read: old, ugly, angry, and manipulative) robowaifus, but to also ensure they have UN-sanctioned """rights""". Men need to flatly reject this shit on every hand. But, cucked men are the reason robowaifus are even needed today in the first place though, so there's no strong reason to think they won't repeat the entire destructive cycle and simp for the Jew's plan yet again.

Not sure what thread this fits but Nvidia recently solved the texture sticking problem for GANs that will greatly improve generating waifus, and eventually voice generation and possibly chat AI, motion planning and other forms of AI dealing with hierarchies. https://nvlabs.github.io/alias-free-gan/ The paper is pretty math heavy but basically they made 3 key changes to StyleGAN2 to create Alias-Free GAN: 1) using a better upsampling filter to remove aliasing caused by bilinear interpolation (the blocky aliasing makes it difficult for information to continuously flow in all directions, especially for rotations) 2) upsampling before applying ReLU and downsampling back to remove aliasing 3) low-passing features before downsampling (essentially blurring them) to remove frequencies above the output bandlimit so unnecessary information doesn't leak through the network The third one is perhaps the most interesting and has the greatest effect. By low-passing features before each downsample in the discriminator, each layer has to focus on learning a specific level of information in a hierarchy, allowing it to abstract objects far better, which in turn trains the generator to render objects far more clearly and consistently. Not only does it solve the texture sticking problem but it also achieves rotation and translation equivariance, which means the generated images can be re-generated moved or rotated without causing distortions. This could open up the possibility of having visual waifus animated and rendered purely with a neural network. They plan on releasing the code in September (CUDA only) but I'm working on implementing it in StyleGAN2 (with CPU and CUDA support) and finetuning the This Anime Does Not Exist model: https://thisanimedoesnotexist.ai/ The texture sticking problem and lack of hierarchical features is what causes TADNE to fail to render fingers correctly and horribly disfigure waifus. I'm also working on a projector using covariance matrix adapation and finetuning SqueezeNet on drawings so existing images can be projected into StyleGAN2's latent space and manipulated there, such as giving it a sketch on the content layers and rendering it with a desired style like Artbreeder does. It'll be like having an art assistant that can render your drawings for you, in your style, in 1 second flat instead of spending 8 hours on it. Once it's complete I'll implement it as a Krita plugin for convenience. If you're an artist this is going to be a great opportunity to up your creative output and cash in to fund your robowaifu pursuits before this tech becomes commonplace. Other possibilities for it could be generating manga, game assets and animations.

>>11470 This all sounds quite amazing Anon. Godspeed to all your studies and efforts in this direction. I will get my Wacom tablet warmed up. I've often wanted an improvement of sketch-based (even with just stick-men-like ones using pose estimation, etc) animation systems for Blender. Regardless, top post/10 and looking forward to your progress with this!

The chatbot thread, which also aas the general AI thread, seems to have hit it's bump limit. I wrote this >>11477, but it didn't come up. I hope it doesn't delete the old postings there, it probably needs to be archived.. I guess we then need to have a new thread for something around talking, processing input, "thinking" and responding, like NLP and graph databases?

>>11481 OK, thanks for the headsup Anon. /robowaifu/ autosages threads at 350, and we start another thread. I'll give the OP of the thread the first opportunity to create his next thread, or I'll make another one instead if needed. > I hope it doesn't delete the old postings there No, we aren't at the number of pages (25) where that can become an issue.

>>11483 If you make a new thread, please name it differently, idk e.g. "Conversational AI, NLP and Knowledge Management" or so.

>>11486 It's probably a good opportunity for us to merge some threads. For example there's explicitly and NLP thread that obviously has a lot of overlap.

>>11486 How about "Conversational AI, Natural Language Understanding and Knowledge Representation"? >>11483 It would be great if we could summarize progress in the field and discussion from the previous thread. I think the most apparent thing is that hybrid approaches and ensembles of networks have been having the most success. Researchers are actually starting to mix transformers, BERT, LSTMs, graph networks and other things together but it's getting to the point they're just hitting a wall and not really doing anything new. The only paper I've been really excited about is Learning to Summarize with Human Feedback (2020) https://openai.com/blog/learning-to-summarize-with-human-feedback/ Some papers and books covering what has been done recently: Neural Approaches to Conversational AI (2018) https://arxiv.org/abs/1809.08267 A survey on knowledge graphs: Representation, acquisition, and applications (2020) https://arxiv.org/abs/2002.00388 Conversational AI: Dialogue Systems, Conversational Agents, and Chatbots (2020) https://sci-hub.se/https://doi.org/10.2200/S01060ED1V01Y202010HLT048 Towards a Human-like Open-Domain Chatbot (2020) https://arxiv.org/abs/2001.09977 Conversational Question Answering: A Survey (2021) https://arxiv.org/abs/2106.00874 Survey on reinforcement learning for language processing (2021) https://arxiv.org/abs/2104.05565 Improving Language Model Behavior by Training on a Curated Dataset (2021) https://openai.com/blog/improving-language-model-behavior/ Some current SotA models of interest: Commonsense reasoning with semantic similarity: https://paperswithcode.com/paper/a-hybrid-neural-network-model-for-commonsense Question answering: https://paperswithcode.com/paper/tanda-transfer-and-adapt-pre-trained Emotion recognition in conversation with graph networks: https://paperswithcode.com/paper/directed-acyclic-graph-network-for Emotion recognition in conversation: https://paperswithcode.com/paper/contextualized-emotion-recognition-in Entity linking: https://paperswithcode.com/paper/autoregressive-entity-retrieval Neural belief tracking: https://paperswithcode.com/paper/neural-belief-tracker-data-driven-dialogue General purpose: https://paperswithcode.com/paper/albert-a-lite-bert-for-self-supervised General purpose: https://paperswithcode.com/paper/exploring-the-limits-of-transfer-learning Translation: https://paperswithcode.com/paper/understanding-back-translation-at-scale Improvement to lightweight Transformers: https://paperswithcode.com/paper/lessons-on-parameter-sharing-across-layers-in

>>11493 >Conversational AI, Natural Language Understanding and Knowledge Representation Fine with me. >Some papers and books covering what has been done recently Lol. A lot to process, even only looking through. I'm trying to make one or several diagrams on what skills are covered with existing programs or researched ideas and what not. Also, based on that I'm thinking about how we might be able to design some good AI out of it, by combining these elements. One big problem I see is responsiveness. Even if some program can figure out some stuff, it may just take one or two seconds per sentence in one step. Needing to do more than one thing will accumulate these delays.

>>11486 I would suggest the word 'Chatbot' be used in the title, but otherwise I'd say that sounds reasonable Anon. >>11493 Sounds like you have a good handle on this Anon. Might I recommend that you take on the task of creating the follow-on thread? I daresay you already have a solid idea where you'd like to see it go. I may use one of the other threads for the more low-key talk on basic waifu apps/chatbots, that is, from the user's perspective. But again, we certainly need material from the implementer's perspective here, since that's our primary focus as a group.

>>11499 >suggest the word 'Chatbot' be used in the title If so, then maybe after Conversational AI in parentheses. Chatbot sounds to much like it's one (simple) program, but it's the user interface and much of the internal processing of information is related to it.

>>11508 Sure no debate. I withdraw my suggestion. As I mentioned in the second part, another thread more suited to the layman should also be addressed. That one can use the word in question.

>>11515 >another thread more suited to the layman should also be addressed Sure, maybe something straight forward like "chatbots: basics and user interfaces"

>>11499 I got my hands full at the moment but I should have some time in 2 weeks. Work being done on contextual AI is also of interest, where chat history is parsed into data structures and user intents and the AI adapts to that data and other environment variables. The research community doesn't touch on this much since it's more of an engineering problem utilizing existing tech but big business is all over it.

>>11520 OK, sounds good. Since it's been a couple days with no response by the original thread's OP, we'll just proceed on the presumption you'll create the follow-on thread Anon. I'll unpin this thread. >>11518 Heh, sounds like a whole other thread Anon. My goal was to consolidate, not multiply threads! :^) >>11481 We should have a new thread in a couple weeks or so Anon. Meanwhile you can make posts on the topics ITT, and I'll migrate them when the time comes.

>>11520 > contextual AI These are all different terms for the same thing or parts of it. Of course some of us here are long aware of needing some complex system for conversations consisting of different parts which cares about context and parses current and past conversations, and sorts the data into graph databases, for example. There are all kind of terms for such advanced systems and it's parts, which can be used to find more information. Cognitive AI would be another such term to search for. Just claiming GPT is a chatbot is easier, though. Not that I have done enough to come up with a alternative, but more discussions related to it might help. It's a lot of things to look through, understand, and learn, which can be discouraging. >research community doesn't touch on this much That's why we're getting distracted over and over again. Businesses don't publish their developments in detail, but DL researchers advertise their findings. >but big business is all over it. Which could be reason for concern, since there's no general AI chatbot. However, that might be because they limit their efforts to customer support and such, I guess. >>11529 >Heh, sounds like a whole other thread Anon. My goal was to consolidate, not multiply threads Then I don't know what you mean by a "thread from the users perspective" and which one you want to use for it. The chatbot thread is full.

Question: Is anyone else missing the background mascot image of Chii in her nightshirt? I just want to confirm it. (the source image in the CSS is still up).

>>11537 Seems to be due to the content security policy of the server. Replacing it with a base64 image will bring it back: https://files.catbox.moe/w9gva5.txt

>>11541 Thanks! I've just discovered some kind of bug that's preventing me from updating it. I'll try to connect with Robi to let him know sometime soon. Hopefully the BG image will be back soon.

Help me out, what time zone is this server supposed to be on, the post times make no sense to me I guess I could look at my own post time and extrapolate the difference but I am really curious how or why we are offset by so many hours, in comparison to any North American time zone

If you enable JavaScript, then you can click the 'Settings' link along the page header bar, and then adjust the time display under the 'Other' tab.

>>11545 Outstanding!

Will embedded YouTube videos work again, one day?

>>11620 I've just (re)emailed Robi again concerning this Anon, thanks. In the meantime, I'd suggest you rely on youtube-dl . >--- Date: Tue, 20 Jul 2021 04:18:13 +0000 From: robowaifu atddd cock dotddd li :-DDDD To: admin atddd julay dotddd world :-DDDD Subject: embedded youtube videos broken AUGGHHHHH! Will we be getting embedded youtube videos back, Robi? -anon's question regarding: https://alogs.theguntretort.com/robowaifu/res/8492.html#11620 -example URL containing YT links: https://alogs.theguntretort.com/robowaifu/res/2991.html#4797 -embedded video error message result: "This content is blocked. Contact the site owner to fix the issue." -regards Chobitsu >=== -minor email obfuscations -hand-patch gunt links

>>11620 >>11622 Sorry for not noticing this, my bad. I recently added CSP rules to prevent any sort of potential XSS, and forgot to consider that embeds need to be enabled too. Working on it now, they should work again soon.

>>11626 Alright, they seem to work again for me. Make sure to clear your cache if they don't work for you.

>>11626 >>11627 Thanks Robi everything's working fine now, much appreciated. We even have our beloved mascot back now too. :^) Cheers mate, hope things are going well for you Robi.

>>11470 Wow. I want whatever it is thats going to be put out. I can imagine a single artist being able to create an animated waifu... in film or in videogame... using nothing but a source set of keyframe poses.

- Risk of disruption in global logistics bc shipping labor shortages: https://youtu.be/GABLJbNvTIc - maybe don't order parts at end of the year, do it now. Especially if it's not something expensive but just some cheap electronics or mechanical parts. - GPU prices went down a lot on the other hand, but might go down further. Still not at MRSP (the recommended price when they came out) last week. >>11627 Yeah, thanks for all.

>>11697 I think an economic and potential societal collapse in the USA is going to happen soon. Else it will be a very slow and painful depression where we won't even be able to afford anything outside of food/water/shelter. I've been wanting to really focus on programming and trying to "catch up" to what seems to be 3 decades worth of knowledge, but instead I have been setting up a farm making sure I can at the very least feed myself if things get bad. I'm almost there, I can produce my own electricity, collect my own water, and soon have a stable way to produce my own food. It's going to be shitty if society falls apart before we can make significant progress on robowaifus.

>>11701 its been a concern of mine too but I don't see us "collapsing" into something irreversibly pre-industrial. If anything it would still be like the poorer parts of Star Wars where they still have technology it's just old, re-purposed and fixed up with gum and duct tape. That being said, is there any reason why we can't have IRL collaboration? It would show a level of seriousness if some of us had actual workshops and space dedicated to this cause. I'd be the first if I was at a better place in my life. (It is getting there, but going through a divorce at the worst possible time, and then Covid right as I was getting back on my feet) Idk what I have for a vision of this other than maybe a shared or accessible space with spare parts, 3d printers, resources, etc. A Robowaifu Research Institute (and Laboratory) in so many words.

>>11704 That's called a hackspace with 3D printers and maybe other machines, which might be run by non-profit organization or as a business. There are podcasts and video talks on how to set something like this up, if you search for it.

Hi guys Looking for introductory robotic hardware resources. The very babbiest of stuff. Words like "arduino" and "DC motor" just fly over my head currently. I guess I could start with some kids craft projects just to dip my feet in.

>>11723 Then search for videos on YouTube on those topics like "dc motor explained" or " introduction to Arduino".

>>11723 Hello Anon, welcome. If you actually want to try your hand at it, then you might have a look at anon's OSRM thread for ideas (>>11446). You could also look into the Waifu Materials thread for inexpensive materials ideas ( >>154). If you're simply trying to gain some basic understanding overall, then I'd suggest you look into various children's robotic toys, particularly the humanoid ones, and familiarize yourself with how they work internally. There are still a wide variety of videos available for them on YouTube at present. BTW, you can feel free to ask any 'silly' questions ITT, we welcome newcomers on our board.

>Also, any doll owners or prospective doll owners, I would recommend you go to the gym and start lifting. It makes moving the doll around, especially laying her down MUCH easier. You don't even have to do a crazy amount. Just start with a basic bench press, deadlift, row, and arm routine. I find it hilarious imagining a bunch of buff classy men being robofuckers, probably rich too from pioneering robowaifus. I'm going to start lifting for my waifu.

>>11729 Good idea. I did that for a 3DPD, so why not much moreso for an actually-classy companion instead? Anyone have some robowaifu-themed DO IT FOR HER's?

>>11733 It will also increase bloodflow and productivity.

>>11744 Perfect, thanks bro. Aegis is top-shelf robowaifu materials for sure. B2 or Chii would also be great too. Now I want to figure out how to arrange having this picture where I can workout again.! :^)

>>11704 >its been a concern of mine too but I don't see us "collapsing" into something irreversibly pre-industrial. I think we will, as the USA doesn't have any real "industrial" plants as is. Once the inner cities can't get their gibs, it will devolve into absolute chaos and maybe even cannibalism. >That being said, is there any reason why we can't have IRL collaboration? Fear and paranoia. There is the possibility, of working together to move to a red county and out-numbering the current inhabitants to fill the positions of the county government. We could stop a lot of the "woke" laws and make it where it is hard for the welfare policies to creep in. This is already happening, but some people have seemingly completely bought up certain counties, but their intent isn't robowaifus lul. I've talked to some CS/CE engineers already IRL, and I am sure they lurk here now, but they absolutely hate talking about it IRL and are also paranoid. Personally, I think it might be better to keep it digital right now, but progress will be slow and secretive. I think we may see a "game" released that will double as an app that works as a type of secretary bot. Although, the other issue is the current fact that we don't make certain electronic parts (chips/semi-conductors) state-side anymore. Which inhibits any near future mass production of a simple physical bot. While I am hung up on making a chatbot, I think perhaps making a bot that can do household chores would be a huge hit and be a good source of funding for any future endeavors. We will either see some (((big company))) come in and try to high jack the idea of robowaifus, or we will see a random "leader" anon that will set us on an opensource path to greatness. Although, we can assume there WILL be push back once women realize where it's leading.

can someone explain the technology of how sentence is built to convey or answer or tell something? I am trying to write for fiction but i find myself lost and inconfident about the content so i m going to complete basic just to know if the stuf will work please dont say "find something fun or interesting" as those dont work well for me so i hope there d be something more basic or conceptual about this

>>11826 It's a really good question Anon, but not one that bears it's own thread. I'm going to move this into the QTDDTOT thread. Maybe you'd want to expand it into a more detailed inquiry later?

>>11779 > I think perhaps making a bot that can do household chores would be a huge hit and be a good source of funding for any future endeavors. I think you are exactly right Anon, and it will be far more than just a 'good source'. I'd predict the company that capitalizes on this will have many billionaires among it's founders within five years or so.

>>11725 >>11727 thanks guys i've been looking into it are there any free simulators i can use? kind of like the bridge software we did in highschool? i think maybe that would be better than just dropping money on stuff out the gate.

>>11848 Sounds good. Not sure of exactly what kind of simulator you're thinking of Anon. There's a robotics simulator mentioned (>>10531, >>10536). There's also a general electronics 'game', and an online simulator for the 6502 chip that I can think of.

>>11826 I'm no expert on natural language processing (NLP) Anon, but I think one reasonably-sensible sounding option is to break a sentence down into the parts of speech each word is (noun, verb, adjective), and then chain those together in a correct way to extract out basic phrases or statements in sentences. So "See Spot Run" most basic fact is Spot Run (noun, verb relationship). Does that kinda make sense?

>>11704 I wouldn't mind IRL collaboration with someone I trust but I live in the middle of nowhere off the grid. It's just not practical. Also building a lab too soon would just be burning money. I estimate I'll need at least $50k to make a full-size robowaifu and $2k to comfortably start printing a body and prototype, so I'm focusing on making big fuck you money first. Without money there's no chance anyone is going to build a robowaifu. Without a positive cash flow whatever workshop is created will shutdown.

>>11826 I don't understand how this is question was meant. You won't get around doing some reading and watching explainer videos. There are prescripted responses with the option of changing a part of the response (AIML). Then there are text generators based on statistics like GPT-3. Then sentences could also be generated by a program based on data in a database. >>11848 Simulators for what? Arduino?

>>11827 ok, thanks alot. i hope it gets answered

>>11852 Words of wisdom IMO.

I have to admit, everything going on in this board is a bit hard for me to follow. This is just a suggestion, and I have neither the skills nor the time to be doing this but; what do you guys think of creating a news letter? Like the one smug/a/ is creating. It could be useful to keep track and updated on the progress Anons have made in all these different threads all in one contained news letter. Still, its just a thought and I doubt I can contribute to it but maybe there are some Anons here who can. Or maybe I'm just retarded and the only one who can't follow the board.

>>11873 >Or maybe I'm just retarded and the only one who can't follow the board. No, that's definitely not just you Anon. It's a vast arena of different domains, and it's always been a 'glorious' chaos using an imageboard to try and keep try and keep track of it all. As anon pointed out this can be quite beneficial to the creative process: >The advantage of this is you can develop ideas in a chaotic way that cannot be replicated with hierarchical ordered cultures (>>11355) I myself tried to overreach trying to convince anons to keep things orderly and finally did a >tableflip.exe over it. I'm over that now and accept that's just how it's going to be, insofar as the board /robowaifu/ itself is concerned. A newsletter is a great idea IMO, and maybe you can present something along this line to the OP of the robowaifu podcast thread? (>>11364) I personally can't add that to my plate ATM.

>>11873 I'd suggest thinking about also setting up an RSS feed for your online newsletter if you do this.

>>11858 Did the two responses answer your question well enough Anon?

>>11852 "middle of nowhere" might be the best idea IMO. Seattle or Tech Valley are overpriced and already dominated by liberal elites. What if we found some inexpensive place (flyover country like Oklahoma or Nebraska) and bought an old warehouse that was being auctioned off. That's one idea I've been floating around. I currently work from home (but not for myself) so technically I could move to anywhere, if that were something I needed to do.

>>11930 as far as any of the rurals would be concerned we'd just be a robotics startup. It would be above the heads or interest of anyone nosy and we would be a thousand miles out of sign of any feminists or whatnot looking to pick an easy target.

>>11930 >>11931 I wonder if we could manage some kind of public face that we're simply devising 'household chores robots', ala anon's comment: >"...I think perhaps making a bot that can do household chores would be a huge hit and be a good source of funding for any future endeavors." (>>11779) It would be best thing actually to literally do that, and we could hide our power level (at least for a while) as far as anything to do with robowaifus goes. Then secondary business interests can more offer upgrade kits (complete with all the 'essential' parts), etc., to upgrade the base household robots. These secondary interests could easily be licenses to Asian and other manufacturers as a form of passive income for the base company in addition to straight chore-bot sales. What do you Anons think of that idea?

>>11930 What a gorgeous field of daisies. Splendid.

>>11933 exactly, we'd be approaching the objective from the OPPOSITE end of the degenerate sex-doll companies. The sexual capabilities could be the very last thing and could even be a mod kit. aside from that, we could have a very mundane "face", a non-descript warehouse, which even if someone enters they just see machine parts, shelves, tools, computers. If we're common sense about it there's literally no worry.

>>11935 I think this is an excellent approach. Not only does every.single. NEET & Hikki in the world want/need a chore-bot, SO DO ALL THE NORMALFAGS! And you just know all those """degenerates""" will be scrambling to create sexdoll+chore-bot competitors once they see what can be done -- thus distracting all the heat away from Chorebots9000, Inc. (& subsidiaries). I don't see how it can go wrong, honestly. Of course I'm not so foolish to literally believe everything will go off smooth as silk, but I think it's a sound & tenable business model.

Lol where'd you go Anon? I hope I didn't say something wrong, I'll delete my post too. :^)

>>11939 nah it's all good, I just dont want to leave too much of a paper trail about myself or distract from the point of my original response

>>11940 I see (I think). I tried to figure out how to run an anonymous business before for this, but advice I got from somebody a lot smarter than me said there's really no legit way to, since the govt. doesn't want it. He said you might do something offshore, but that was about it as far as he could see. Maybe this discussion should move to one of the business threads anyway?

> I think perhaps making a bot that can do household chores would be a huge hit and be a good source of funding for any future endeavors. Simple robowaifus will be companions sitting around and be able to chat. Or starting as sexbots and then becoming more of a companion with some form of AI. Anything more complex is harder to do. Also, doing chores doesn't require a human shape, especially not a pretty one, so it would be a very different project. https://youtu.be/joo3ikFS0j0 https://youtu.be/G7Q_jJVBTvU

>>11947 >Anything more complex is harder to do Yep any mobile autonomous gynoid is by definition going to complex, so yea. >Also, doing chores doesn't require a human shape, especially not a pretty one, so it would be a very different project. One thing a favorite sci-fi author addressed, and so did NASA, is that in a human-scale, human-oriented world, then a humanoid-shaped robot is by far the most convenient form -- thus was born Robonaut 2 (>>99). Now Roscosmos has followed their lead and created their own version. Since the business idea is basically to do 'stealth', but fully-capable, robowaifus, then the idea that Chore-bots would be humanoid is a given. Chatbots we can do with a mannequin and a wifi connection to a PC. But she's not going to be an actual robowaifu until she can get up and move around. So, again, it's a given we're still dealing with a highly-complex humanoid robot here, it's well-understood.

>>11943 great ideas anon the good news is we don't really even need to go this deep though 1. it doesnt have to be anonymous, simply have a basic robotics research company. We don't even have to specify that its "humanoid", and we certainly wouldn't mention the waifu aspect, that would be (metaphorically speaking) our black project accessible only on the secret basement level ; ) . 95% of the research is going to be on things like embedded hardware, OS, actuators, materials and fabrication, etc etc etc. All of this can be done in the open without needing to explain "what" it is for. Bonus: any advances or innovations can be patented and sold, they'd certainly be useful toward other applications 2. Moreso the vision I had was simply for a physical place we could pool (some, not all, obviously) our resources and therefore get some leverage from our money and time, doing more together than we could in our own garages and typing to each other on a board. It would be voluntary and optional and anyone wishing to participate would obviously need some vetting and screening face it we're all weirdos but there's a difference between functionally eccentric geniuses and well.. those who are trouble. Speaking from experience Those who wish to live nearby can participate at a higher level but others can commit to a weekend or week a year if they want, sort of thing. Our very own Legion of Doom ; )

>>11949 >But she's not going to be an actual robowaifu until she can get up and move around. I disagree. If she can move her arms and legs, but also talk then she's a robowaifu. More so, if she can give massages, get up with some help, do some work while standing on her own or dance. Sure, waifus should do chores, but emotionally this would already be enough to call her a robowaifu.

>>11955 I'm not here to build a company and have to deal with all that is related to that. It's probably just one more distraction and a fantasy. There are enough people out there which could push the development a little bit. No one needs a company with developers, and it's not going to be a goldmine anyways. The whole idea here was to keep patents and companies out. Recently it comes up more and more. How would you make money from open source hardware? In worst case it would end in people fighting each other with lawyers and in best case with a lot of politics and overhead, while getting distracted from the goal for the next 10 years or so.

>>11958 You don't want to be part of a robowaifu legion of doom then. Your loss : (

It seems Google's AI officially recognizes /robowaifu/ as a research hub :^)

>>11957 OK, I see your point Anon. My apology she really could be a robowaifu at that point, I was just being narrow-minded in my definition.

>>11955 Those ideas all make sense, but obviously we won't be anonymous in an actual for-profit company (nor can we even legitimately -- that is legally -- be so). Maybe just a few of us need to deanonymize first to make things legal, but the rest of us can remain anonymous? I definitely think it's a good idea to run a service-bureau type deal out of the facility. Kind of similar to Thingiverse and other 3D printing services, but catering specifically to Robowaifu Technicians, and with some privacy and security guarantees. Does this seem realisticly doable? I'm not personally in favor of attempting to legally leverage IP or other assets though. Either through restrictive licensing, or especially through patent. Definitely not. That's actually anathema IMO to the entire concept of making everything here freely-available for everyone to use as they see fit. Thus our MIT (Expat) & BSD-3 licensing. Chorebots9000, Inc. can legally utilize any and everything from /robowaifu/ (same as every other individual or organization around the world). But we should be giving intellectual property of every sort back to the community, not trying to cloister our more-specific advances just to ourselves alone. Let the Globohomo do that, it's what they're best at enemies of humanity that they all are. Yes, I agree with you that everything can be done above-board and in the open. Indeed it will need to be for long-term sustainability. And as you suggest, basement-level access can be highly restricted. But ofc even that will become known eventually as our offshore suppliers begin churning out wonderful smexu & charming robowaifu upgrades for our more-advanced Chore-bot lines. After all, is a house really a home without having a cute anime catgrill meido in tiny miniskirts there, I ask you!? :^) Heh, but the point being, it will soon become clear enough what the game is, and after all...this conversation is a written record of our intents. We'll simply have to keep our noses clean locally, and let the hater's REEEs wash off our backs.

>>11963 Lel. Well, we legitimately are one Anon. Name another place like this one?

>>11958 I'm not intentionally trying to debate your points Anon, but 'keeping companies out' is definitely not a goal here on /robowaifu/. They are just as free as you or I are to take these ideas and run with them. For-profit commercial businesses will be the catalyst that will eventually catapult robowaifus onto the world stage, not our merry little band of adventurers here each working in isolation alone. In the end, the widest distribution of robowaifus will create the greatest good for the greatest number of men across the world. And after all, that's the higher goal we're after here, right? Our challenge in all this is simply to get this information distributed openly far-and-wide while we still can do so, to resist our oppressor's intentions to corner the market and corrupt it. It's simply a race against time, as I see it. I'm not attempting to make some assertion to dictate to other anons here, and certainly every man needs to follow his own heart. I'm simply pointing out that commercial operations are by definition far more capable of seeing that more robowaifus land alongside more men worldwide. I encourage that. But I certainly agree we don't want companies or their politics or agendas to dictate anything to us here -- nor will they as long as I have a say in it. My position is that they are both welcome to use our assets here as they see fit, and I heartily encourage them to also give back -- and openly -- to the community from which they've benefited. Pay it forward, everyone.

>>11982 Trust me Anon, the LAST thing you want is some satan worshipping commie big tech corporation trying to demand exorbitant prices on a pozzed series of robo waifus.

>>11983 Oh, I do trust you on that one Anon. So do we all here...it's literally the last thing we desire. That's why I described it as 'a race against time'. If we can reach a critical mass of knowledge in the world for creating inexpensive yet capable and pozz-free robowaifus, then 'the cat will be out of the bag', and the Globohomo won't be able to stop it at that point. Thousands of smaller businesses, and 10's of thousands of hobbyists will be creating them in short order. They're going to do their evil regardless of us. Our agenda is simply to outpace them and get it broadcast far-and-wide first.

>>11982 >keeping companies out' is definitely not a goal here on /robowaifu/ agreed, think about what innovations could be jumpstarted through R/W projects. Much like space exploration did. Say some of us create a specific kind of solenoid for example, and manufacture them for sale. How would that be "selling out our waifus" to corporate overloards? From an external pov we'd just be a robotics research facility that manufactured a very unique and useful specific part. Literally no one would have any idea how or why we were coming up with it, and they would probably not care or just assume it was for something mundane like a burger flipper robot.

>>12137 I think the issue anons have is more so most startups are just in it to make money. If a larger corporation or investors come along, 99 in 100 are going to take the money and exit, and then they're going to attack competitors tooth and nail, including open-source that might threaten their business. People generally think open-source is low quality but it's not necessarily because they're lacking man-power or skill. It's because businesses do market research, analyze competing projects, find their weaknesses that users care about and simply outdo them since their bottom line depends on it. For anything we do to even stand a chance it needs to be convenient to use and run fast on common hardware. Look at NovelAI for example. People could train their own models with pre-existing scripts but they still pay a company to do it instead because it's easier and gets fast results. In terms of chatbots NovelAI is leaps and bounds ahead of what anyone here is doing. If someone gets a linear transformer model pretrained or fast weight programmers working it might level the playing field for a little while but companies will catch up and outcompute us. >>11986 Honestly, if we're going to compete on that level we need a better plan than wishful thinking that our tinkering around is going to amount to anything. It's great we've come a long ways from knowing fuck all and inspired more anons to get involved, back in 2017 I couldn't imagine we'd make 1/10th the progress we have by now, but the reality is we're still falling behind and I don't have any ideas how to up my game besides burning money on GPUs to get a linear FWP model working that achieves sub 12 perplexity.

>>12164 Just to bring it to your attention, if you have time. Going Beyond Linear Transformers with Recurrent Fast Weight Programmers >abstract >Transformers with linearised attention ("linear Transformers") have demonstrated the practical scalability and effectiveness of outer product-based Fast Weight Programmers (FWPs) from the '90s. However, the original FWP formulation is more general than the one of linear Transformers: a slow neural network (NN) continually reprograms the weights of a fast NN with arbitrary NN architectures. In existing linear Transformers, both NNs are feedforward and consist of a single layer. Here we explore new variations by adding recurrence to the slow and fast nets. We evaluate our novel recurrent FWPs (RFWPs) on two synthetic algorithmic tasks (code execution and sequential ListOps), Wikitext-103 language models, and on the Atari 2600 2D game environment. Our models exhibit properties of Transformers and RNNs. In the reinforcement learning setting, we report large improvements over LSTM in several Atari games. Our code is public. https://arxiv.org/abs/2106.06295 Linear Transformers Are Secretly Fast Weight Programmers >abstract >We show the formal equivalence of linearised self-attention mechanisms and fast weight controllers from the early '90s, where a ``slow" neural net learns by gradient descent to program the ``fast weights" of another net through sequences of elementary programming instructions which are additive outer products of self-invented activation patterns (today called keys and values). Such Fast Weight Programmers (FWPs) learn to manipulate the contents of a finite memory and dynamically interact with it. We infer a memory capacity limitation of recent linearised softmax attention variants, and replace the purely additive outer products by a delta rule-like programming instruction, such that the FWP can more easily learn to correct the current mapping from keys to values. The FWP also learns to compute dynamically changing learning rates. We also propose a new kernel function to linearise attention which balances simplicity and effectiveness. We conduct experiments on synthetic retrieval problems as well as standard machine translation and language modelling tasks which demonstrate the benefits of our methods. https://arxiv.org/abs/2102.11174

>>12199 >Maybe make our own OS (linux-distro) that specifically is for robowaifus. Going to look into Guix SD and Void Linux soon. Trying to create a completely new distro is just one more rabbit hole. A distro needs to have reproduceable builds and be open for external software. Guix has that.

>>12164 >Honestly, if we're going to compete on that level we need a better plan than wishful thinking that our tinkering around is going to amount to anything Fair point, and I hope that some sort of effective solution presents itself to us while there's still time Anon. Please be aware that myself and everyone else here who has been a long-time regular deeply appreciates all the hard work and exceptional creative talents you've shown here. You're a 'national' /robowaifu/ treasure Anon. Together, I'm confident we'll figure something out. All of us are smarter than any of us. :^)

Why don't we have a wiki? Imageboards aren't really a good medium for organizing information.

>>12199 >I am not sure why, but it seems this board has been catching extra attention lately from outsiders. Maybe it's just a few posters talking a lot... Hopefully it's not 3 letter agencies, or "secret" groups trying to create internal strife. It's textbook PsyOps. Or maybe it's because people are sharing it... I've already addressed these topics a few months ago Anon. >The greatest threat to /robowaifu/ are two groups from without, not within. 1) Leftist infiltrators trying to subvert the program and ultimately to emasculate it into uselessness, and 2) glowniggers and other goons trying to deplatform it at the behest of their puppetmasters. >For the former case, we all just need to stay on our guards and immediately shut down leftists whenever they pop their heads up here. Politically-correct ideologies and language are clear warning signs of their presence. We're Anons here on /robowaifu/, not soygoys. It's pretty easy to tell the difference actually. Basically, The Left Can't Meme, since that takes an independent soul to succeed at. All they are, are bugmen and even worse -- leftist bugmen. >We have secondary and tertiary sites available to us and full board backups for the latter case. Even if something happened to me or this place, some of you just approach the Administration of the other two sites, explain the situation clearly, give them a full backup of this place and ask for a full migration of /robowaifu/ to their site. I'm sure it will all work out somehow. >OTOH, the idea of robowaifus appears to be, roughly speaking, wildly popular among the imageboard crowd. I think we will have little concern about our friends and neighbors turning on us. They have nothing but goodwill towards us AFAICT, and I know everyone will here do the right thing when and if the time comes! :^) (>>7496) My thinking on this hasn't changed in the slightest during the intervening timespan, and 'predictive-programming' is a weakload gayop as old as the hills tbh. Every Anon ever is already onto it. Like I said, the Left can't meme. >I worry that at some point GlowAnon's hire legal experts/lawyers in an attempt to throw the board into chaos with accusations of "someone here stole my work!" Kek. Glowniggers don't need to hire greedy lawyers Anon. And as far as any takedown requests, it's literally none of our affair on the board. Robi has that responsibility, and ofc he'll do just as he sees fit on his own site. We're simply his guests here after all. >Also, I think the board could use some more categorization for organization purposes. This is one reason why we have a Library thread now Anon. Pull requests welcome. :^) >>12241 >OpenBSD I think it's a very good platform. Of course nothing's perfect, so don't expect some ex machina panacea to suddenly appear out of nowhere from it's use. Everything good takes hard work. >Moving forward things will probably start getting less friendly as more people find this board so be aware of who you're helping. If someone admits they don't really care about having a robowaifu but are making robowaifu-related things, don't ignore that detail, because it will be really important once that project is complete and achieves its true objective. The simple fact is we're on a cleartext, anonymous imageboard. Anyone can access the "autistic works of fiction and falsehood" as they wish, nor could we prevent it if we wanted to. We don't want to. Just like Jesus' parable of The Wheat and the Tares we have to accept the bad along with the good. With God's blessings on us here, we'll finally achieve this miracle in the end despite all the haters' rage. Just stay encouraged Anons, and work hard towards the prize! :^) https://www.biblegateway.com/passage/?search=Matthew%2013:24-30&version=NKJV >=== -add crosslink -correct autism quote

>>12289 It's been a long-term desire Anon. Dollfan and I actually began to investigate doing just that back in the day. Today, I can't add that onto my plate, Dollfan's gone (AFAICT), and quite frankly the chaotic nature of the dialog here on an imageboard fires the creative juices in a way a wiki never will. Still, it's a good point. As I've said many times in numerous venues, I absolutely welcome any and all robowaifu-oriented forums popping up everywhere. So get on it Anon! :^) >=== -minor prose edit

>>12291 Well I got some cash and time to burn so why not. I'll post the link here once everything is setup.

>>12313 >Odds are if any anon shows up and are obsessed with the konami code puppet show (right left up down up down) they are a glownigger of some kind. Then 50 more show up and ban any remotely sane voice and poz the entire site. Robowaifu is no less vulnerable to this bullshit. Hmm. There's some possibility of the former, already objectively not so of the latter. Still, better safe than sorry. Better activate the Chicken Little program. I'll audit things up a bit here and there Anon. Thanks for the warning!

>>12333 fnaf fandom is an immediate red flag note: I had to google wtf a konami code "puppet show" was

If yall dont get a real girl instead of this gay ass robot shit you gonna end up like Chris Chan °3°

nice

>>12345 Kek.

lel

Lol, nothing could be quite that bad OP. Besides, would that really fix anything? Even if you're just trying to give us all """good advice""", have you actually checked out the weather outside? Pretty stormy conditions I'd say for a man wanting to find a good woman. Robowaifus -- while not real humans -- may wind up being a much better choice for many men in the end. If we can ever create even one open-source robowaifu that is at least as good a companion as your OP pic, then we'll all be absolutely golden.

>you gonna end up like Chris Chan °3° this implies the existence of a Chris Chan "2"

Why have a waifu when (you) can make the waifu?

>>12345 Even without a robowaifu, one of my reasons for not wanting a human female is quite boring and unrelated to all of the insanity currently infecting Western nations. Snoring. If I ever have to sleep near another human being I usually end up having a terrible night, because of their snoring and/or loud nasal breathing. I cannot function on four hours sleep per night. I like my sleep. I'm not innocent either BTW, I've been told that I sometimes snore when in a deep sleep. Also I don't want some nosy woman going on my computer and rifling through all my porn. Throughout human history, one of the main reasons we formed tribes and families (other than for protection) was division of labour. But with all of the gadgets and machines we have nowadays, it's no longer necessary for even two people to live together. My toasters, microwave, smart oven, slow-cooker and chip-fryer already do much of the work of a domestic robowaifu. Sure, they're not in an aesthetic or sexy female form, but they get the job done with maximum efficiency. The housewife has already been replaced. If our current rate of technological progress is maintained, then eventually the cognitive processes of the average human being will also be completely replacable by machines.

>>12363 >Sure, they're not in an aesthetic or sexy female form, but they get the job done with maximum efficiency. /clang/ would like a word with you Anon. I think the insanity you mentioned is due in no small measure to the technological advances you speak of. Women have largely become useless eaters (far worse, actually. since they are the actual instigators of many broad categories of evil culturally during current year) outside of their utility as sex objects and baby factories. This reference is obviously ironic af intentionally so given their innate leftist bent. And since only young women are physically suited to or attractive enough to function effectively in their two sole roles to society, then you have a vanishingly small service to humanity they can provide at all. Once they are a reality, robowaifus will certainly have most of the advantages. Once artificial wombs are readily feasible, then even that last card goes up in smoke for them.

Alright, it looks like the initial lulz have died down. I'll go ahead and merge this thread soon. Probably the current /meta.

So, I'm the board's current primary C++ developer, post-wise. I'm currently exploring a robowaifu access-control system project using Anon's ideas (>>12431, >>12455, ...). Till now, we've been basically entirely dependent on using catbox.moe as our 'repo' server for my little projects. While there's nothing particularly bad about this, they have had issues in the past with stability, etc. Who knows? So, I'm going to attempt making multiple ways for us to have the code w/o depending on the speshul-snowflake-accommodating, CoC-loving promoting """repository""" vendors like SJWhub and (sadly, now too) Gitlab. I'm currently experimenting with actually posting the code here on the board itself, and with creating an awk'd merge.file on a place like zerobin. Since Lynxchan (probably rightly so) prevents many types of files from being posted here including zips, we'll need to use a minor (but fairly commonplace) sleight-of-hand to get by the IB-esque file restrictions. As a test, I've performed a typical meson dist on my current v0.1c WIP project. I've renamed the final .7z file I typically post up to catbox as a .pdf instead, which hopefully should allow me to post it here on the board instead. I'v also posted an awk-collected version of the codefiles, etc., that I myself have personally written and made a paste of it up on zerobin. With a little scripting effort, in a pinch this could also be used as a way to reconstruct the codebase. https://0bin.net/paste/QET3EII+#hKdk1Alw4WH7oNxJrmVb2JAEhxtiMEqR5Zv3d7ydycM Note: This paste does not include the two (yuge) external third-party header files CLI11_v2.0.hpp, and json_v3.10.hpp. Those would need to be obtained from their respective sources first if this route was needed to reconstruct the codebase. So, to test this, just download the pdf file, rename it's extension to .7z, then extract it. Then run the sha256sum -c command as usual to confirm the inner .xz archive is uncorrupted, then proceed as typical for my projects to build and run. c7cf0beb0e3f480949e39fbc1274296e81ae3e897bcce1cac70cb57f9756689a *rw_access_control-0.1c.tar.xz

>>12530 Also, since this is a test, I'd appreciate any feedback on success/failure any anons have at it. TIA.

>>12530 (C++ dev here) OK, I tested my own dogfood to completion and it's all working fine for me. Please let me know if anything instead breaks on you.

>>12530 >Note: This paste does not include the two (yuge) external third-party header files CLI11_v2.0.hpp, and json_v3.10.hpp. Those would need to be obtained from their respective sources first if this route was needed to reconstruct the codebase. BTW, the '.pdf' does include the two libraries, just not the zerobin. Cheers.

>>12532 >OK, I tested my own dogfood to completion and it's all working fine for me. Please let me know if anything instead breaks on you. Ulp, spoke a little soon. Everything went fine with the 'pdf', but then I decided to double-check grabbing the paste and found it was truncated quickly at around 450 bytes or so? The copypasta approach worked fine, but not the download. > Can someone here tell me if I'm doing something wrong with zerobin? https://0bin.net/paste/vvKhHDt8#pfYGz17gkj0ANOd5KuDVfFlMhGm1wwibtcK5NCR0YZf

I feel like we could use more meta type threads of a general or /rant nature, building a culture around robowaifus is important even if the discussion isn't directly applicable to the construction or programming of them. We have a team morale and sometimes we need a timeout to just bullshit on certain topics and to redpill newcomers about w*men without it dragging other threads down.

>>12556 >I feel like we could use more meta type threads of a general or /rant nature, building a culture around robowaifus is important even if the discussion isn't directly applicable to the construction or programming of them. We have a team morale and sometimes we need a timeout to just bullshit on certain topics and to redpill newcomers about w*men without it dragging other threads down. I agree that team morale is an important topic for this board (or any other for that matter). Building a good team spirit and esprit de corps is important for any men, particularly those engaged in war (which we are). The difficulty for me is that if unrestrained expression of hatred in any form is allowed, it quickly becomes the exact opposite of 'team morale', and simply devolves into a spite-filled agenda of destruction. We're hardly better ethically than the evil feminists and other Commies ourselves at that point, in my opinion. However, I do agree that a shitposting general is a good idea thus why I created the Lounge. However apparently few Anons want to use it as it's intended re-purposing as an underground speakeasy where they can let it all out. Despite that, I'll go ahead and unlock the roastie thread #2. But be advised this is only provisional. If it again becomes a cesspool of pure hatred and spite, if it once again incessantly dominates the board's frontpage, then I'll lock it up once again.

>>12557 Perfection has been achieved. Ice cream chan ushers in new golden era of world peace and prosperity.

>>12580 >ice cream waifu sir, you are on the wrong board ;-)

Not a robowaifu, but my single-celled companions shall never be forgotten. May they rest eternally in putrid jelly.

>>12623 hope you're not planning on culturing cells to put inside waifu https://onezero.medium.com/lab-grown-mini-brains-have-human-like-brain-activity-2b5eeac9e969

>>12624 there are so many reasons I'm leaving this one off the table aside from visceral disgust

>>12624 If there was some way I could grow an organic computer I certainly would. But I wouldn't want to grow a complete, sentient, organic organism because it would likely be malformed (think of the hundreds of aborted attempts before Dolly, the first cloned mammal, was born to her surrogate). And thus it would be very susceptible to illnesses and it would suffer just as much if not more than many human children. As it is, I can only do aerobic, non-fastidious bacteria and maybe simple eukaryotes like nematode worms...and fungi. Getting hold of anything fancy like human neural tissue is hella expensive. Plus when you start getting specific with what you want to grow, some of the ingredients and lab equipment quickly becomes very pricey (anaerobic chamber, incubator and a proper autoclave rather than just a pressure-cooker). Don't even get me started on any kit related to PCR and genetic engineering LOL! Despite this, even plain nutrient agar-based microbiology for revealing a few of the different bacteria on you/inside you and your local surroundings is absolutely fascinating (as long as you don't mind making PPE and hygiene precautions part of your hobby).

>>12580 >Perfection has been achieved. Ice cream chan ushers in new golden era of world peace and prosperity. LOL. <Unleash Icecream-chan robowaifu on an unsuspecting world >The Madmen >>12584 >>12585 Interesting posts meta ronin. I hope we eventually will see robowaifus and 'she could walk and hold conversation', and in the not too distant future as well. >>12601 Would you two like me to move these 3 posts over to the roastie thread? >>12623 Heh. F :^)

After the chatbot thread, which was unofficially our AI thread, the humanoid robot project thread seems to have reached the bump limit as well.

>>12691 Thanks for the heads-up Anon. Mind linking to the the specific thread you mean here? I'm not seeing which one you mean, thanks.

>>12692 Never mind. I think te forum has problems. My comment wasn't there, so the thread got not bumped.

Ey there, i'm someone interested in making robo waifus possible, what would you say is the biggest hurdle right now? Programming? Modeling? materials? I ask because i want to know what kind of money would need to be put where to have this kind of stuff be an actual possibility, would i need to give money to a hardware engineer? etc? give me some info on what the people here would actually like funding for.

>>12721 This is a great set of questions OP but it's not really something that supports it's own thread. I'm going to merge it soon with our QTDDTOT thread (>>8492). And actually, there are answers to your questions already here on the board in various places. If you can afford the time in your schedule, you might have a look around (or task someone else to do so) and see what's already been discussed here. Also, if no one else steps up to answer you in the other thread, I'll take a shot at it later in the week there myself. Regards.

>>>12721 lightweight streamlined and powerful actuators, power storage and delivery systems, a kinesthetic AI with the processing power to translate all of the small movements and on the fly adjustments into actual functions of a type we would understand (walk over here, pick that object up, turn on light switch, don't step on that cat) and that being said the ability to model it's environment and make sense of it all on the fly.

>>12723 i think he forgot to properly move the post over.... but hold on let me break that down, lightweight streamlined and powerful actuators, what do you mean by this, can you show me examples? how much are they normally?

>>12724 This guy is onto something https://www.youtube.com/channel/UCEVelHmwuCqyje6YTAP2UIA But still to build a waifu from all those parts you might end up with something bulky, noisy, and overheated. Granted there are innovations such as miniature/micro pneumatics, hydraulic transformers (as someone else brought up). But yes, as far as moving like a we do, seems animals still have the market on efficient energy storage, delivery and power per pound. It's theoretically possible, but it seems like rather than duplicating flexible natural movements of opposing muscle groups - robotics is still stuck on robots that move via electrical/mechanical servos, slow, rigid, awkward and often requiring a large bulky battery to power. If we could deliver power enough to create a fully self-ambulatory waifu, in streamlined, feminine form-factor, while via creative use of cabling/hydraulics deliver the massive power necessary and coordinate it properly so the thing doesn't hurt others or itself in the process, this would be a significant leap toward our goal.

>>12724 >i think he forgot to properly move the post over.... Heh apologies. Because of the typically sporadic nature of traffic on imageboards, I usually try to give anons a little advance notice before just abruptly moving things here, and leave it up for a few hours or so before doing so. Just let us know if we can help you with further information (but you already have a well-versed anon helping you out it seems).

>>12725 Well we don't need to jump the gun, but you didn't answer me, how much would moving towards even somewhat okay movement cost? what parts would you need? how much would they cost, would they need to be custom made? give me ballpark ranges

>>12721 >Programming? Modeling? materials? >what the people here would actually like funding for. I could only think of some specialist from the outside building something which would help us. Examples which come to mind: - Model of a simplified female skeleton (probably in OpenScad or Blender), which can be changed in size or in some details as easy as possible. - Face generator for faces similar to some female celebrity names as an input, maybe finding similar looking faces as the first step, transformation of the generated face into 3D anime version of it (bigger eyes, like Alita), then 3D model transformation, maybe generation of a crude skull from it based on how it would look based on that face (I think Disney has a patent on that last part). >would i need to give money to a hardware engineer? etc? Idk. What kind of hardware? For a full robowaifu? He wouldn't know how to do that, just bc he's an engineer.

>>12733 >- Model of a simplified female skeleton (probably in OpenScad or Blender), which can be changed in size or in some details as easy as possible. This sounds feasible and rather cheap, i'll ask around the 3d lads i know for you.

>>12727 I'd imagine in the tens of thousands of dollars are a bare minimum entry level enthusiast garage lab, asking for a price point on such a project is really up for grabs. Maybe someone else here who has an active project in the works might have a better idea. My goal is to eventually budget 20k for a starter garage/outbuilding lab.

>>12721 I think about it's so: If enthusiasts work on it, then it won't cost a lot but it'll take quite some time. If specialized engineers would work on it, it might go fast but it would be immensely expensive. That aside, constraints are the problem. Each of us has a different goal and different approach when it comes to details. Hard shell on the outside is different than something like some animated doll. >>12734 Thanks. >>12735 I have no idea how you come up with such a number. Doing deep learning might be expensive, the rest is rather about learning and trial and error.

>>12746 >I have no idea how you come up with such a number. Doing deep learning might be expensive, the rest is rather about learning and trial and error. you're going to want parts to build the body right? even if you're reverse engineering another robot body you're going to need tools, spare parts and testing equipment. If you're just looking at making a disembodied AI like that anon with the chatbot/waifu, then that's a whole other thing.

>>12746 here's a good thread for that question >>4797

>>12734 >Skeleton model I believe it's way harder than you think. If you actually do that, maybe break it down. Focus on the skull or the hips and legs. I want to flesh out how I meant it: - bones of the legs should be splittable in the middle, so the can be printed, but also something added inside. The ends of such bones, next to the next bone probably don't need to be modeled (sculpted) exactly, since we would ad our own connectors in there anyway. - important are mainly the parts which are visible or recognized through touch from the outside of a human body. These parts should probably feel the same. But it's also about the angle of the legs to the hips, so if she would walk it would look femine. - generally some assembly out of mostly CAD parts with some sculpted parts where it can't be avoided would be ideal. The whole assembly (skull for exampl) should thn be scalable, while the parts still fitting together. Idealy the individual parts would be broken down into printable parts without need of much supports. - If you're going for the skull, let's maybe directly go for extended eye sockets, like Alita would have. Waifu skull, not realistic human. >>12756 I don't see the expensive tool part. Space for building different prototypes and testing parts will become more and more important, though. It's better to live in a house than flat, or even better to have some big hall.

>>12809 So one anon has a workshop with a milling machine, aluminum foundry, welding gear and lathe. Not counting the structure I can't image all of that was under 10k - I'd consider that bare minimum, and I can't imagine not also needing more specialized and durable tools than your basic set at w*l-mart. Plus materials. That's where I got that ballpark from tbh.

>>12809 Can you give me a full shopping list of what you want in the model then? I think if i just get a basic skeleton and then add onto it that would be easier at first, so far i haven't found anybody who can do the job.

>>12866 Well, at the end we need all of it, but it doesn't need to be one big assembly. Just the single parts have to be modifiable so they can fit together. I'm currently working on a simple spine. But this wont have a the s-curve humans have. I strongly assume that a model of a human (female) one would help. I mean in form of an spine assembly, which would mostly or entirely be based on CAD instead of sculpting, therefore be a bit simplified. Otherwise, a female pelvis, which would be printable would be great. Maybe out of around five parts which can be screwed together. At least the connectors and screw holders should be parametric (in CAD). I also mentioned the idea of a waifu skull above. Here my spine early prototype: >>12882 obj bone models: >>12878

>>12884 You want a female pelvis, to be modeled?

Apologies if i'm not understanding by the way, you're just wording what you want in a very confusing way.

>>12886 That goes both ways, I don't understand what the question is.

>>12866 The skeleton needs to be printable and based on a female skeleton. We need all of it at some point, but people (myself) are already working very slowly on modelling useable parts. >>12814 That's someone who does all of it on his own, and working with metal. Also, I don't share your assumptions about more durable tools. Machined metal parts seemingly can be ordered from online services, if one even needs those. Also, such developer doesn't need all of it to get started. Get a 3D printer for ~250€, a PC and a start buying electronics, that's it. Doing deep learning at home would add to the costs, though.

May we move skelly talk to the other thread? >>200 Pic unrelated generated by artflow.ai when I typed the robowaifu ai goddess. was expecting some amalgam of 2b and hatsune miku but this is as close as that probably gets

>>12911 >>May we move skelly talk to the other thread I agree. It's just here because someone offered to sponsor some development and asked for ideas. >https://artflow.ai/ >Due to site traffic, estimated waiting time: 970 minute(s)

>>12920 mine said 578 mins but realistically it was ready in 10 just keep the tab open

>>12921 Lol, I got mine but it didn't know who Cameron Phillips was, so I got some guy back.

>>12922 It didn't work with Summer Glau's roles, have to try again later, but still got something nice out of it.

>>12928 The website is somewhat overrun, currently down, and it seems not to know about Chi?!? Didn't get a result including her after more than a day.

>>12967 I think the idea is to ask it to show you the face of something more general and not specific but who knows, keep trying I guess

NEW THREAD NEW THREAD NEW THREAD >>12974 >>12974 >>12974 >>12974 >>12974 NEW THREAD NEW THREAD NEW THREAD