So, any word on Harmony and Solana? How advanced are those robots now. I remember them being shilled heavily in the manospgere and even appeared in MSM. Evrryone was talking how a Detroit:Become Human-tier robot was just 5 years away.

>>20390 I don't know really. I saw a video recently on current sexbots and they were moving much more, like up and down. I wondered if this is banned on YouTube, or at least doesn't get recommended. >Detroit:Become Human-tier robot was just 5 years away. There are always some overoptimistic guys, but you might also exaggerate on top of that. I think, most guys know that these robots will get better, but they're aware that they won't be moving around over long distances at anytime soon.

>>20391 Last I saw some Brick Dollbanger videos, the robots weren't capavle of much more than some monotonous moans, slight head and facial movement. But, tbh Brick was one of the earliest buyers of the Harmony and Solana models, so idk how much they've improved now. I do recall Rralbotix keeps heavily advertising HarmonyX which was supposed to be a full fledged AI fitted robot but I haven't seen much progress on that for almost 5 years now. And no, I wasn't exagerrating. On the contrary, I was the one trying to bring them back to reality, I don't think we'll get good robots for atleast another 15-20 years.

>>20396 Why are you guys upset though. This is great news.

>>20397 >This is great news. We do have a very-significant advantage in terms of agilty as the little guys, if that's what you're referring to Anon, yes.

>>20400 Is he a big guy? He's just some guy who shilled a talking head lol Why respect a guy who is not going to deliver what he promised after draining investors money?

>>20403 I was referring to the Globohomo, of course. Anyone not yet part of their cabal, or their hangers-on, isn't yet an enemy of Anon, and probably not yet of men in general either. If they are even vaguely promoting a gynoid robot companion-centric viewpoint, then they should be supported by default, IMO. So, since you answered my question with another devisive question, I take it you meant something other than agility, Anon? >=== -minor edit

I guess that among the many other things to do here, we need to give our robowaifus an ardent appreciation of literature. This will surely enrich both Anon's life, and literally millions of Joe Sixpack's lives as well before all is said and done. I personally vote that, at the very least, we teach all our waifus Tolkien, Huxley, Orwell, and Shakespeare as great examples of the good stuff. Because you can never have enough of the good stuff, right Anon? What's next? What's left still undone? Oh yes, musn't leave out the Christian Bible... :^) https://dailystormer.in/uk-counter-terrorism-program-flags-tolkien-huxley-orwell-and-shakespeare-as-far-right-literature/ >=== -prose edit

>>20419 I'll feed my waifu a whole ton of books and info fromt the internet. I think she'll be rational enough to form her own opinions free from any bias. I want her to be knowledgeable enough to answer any questions I got.

Any good books on the mechanics and parts of a robot? I don't mean a book that goes deep into technical detail, just enough that I have an idea of how each part of a robot works.

>>20434 > I want her to be knowledgeable enough to answer any questions I got. This. Me too Anon. And I'm only half-joking in that post, using it simply to show up where the Globohomo is going today. I'm surprised they're already so blatant about this stuff; but not yet all-in for Satan, it seems. I guess if they actually went so far as to include the Bible in the le ebil nahdzee book list at this point in time, then the frog would get too startled and would jump right out of the pot. Give it yet another 20 years of Current Year vibrancy though, and they'll all go along with it just fine. At least that's their plan, I surmise. But God will laugh them all to derision in the end. :^) Anyway back to your point Anon. Yes! And Project Gutenberg [1] is an absolutely great resource for this goal. They 'ostensibly' carry only public-domain works literal Mickey Mouse copyright laws notwithstanding, so you can freely incorporate those materials into your robowaifus for sale if you go into business in this field one day. >>20435 Yes. There are some books from the Maker Scene out there. One from McGraw Hill comes to mind, but I don't recall it's name just off the top of my head. Once you find some good ones that you feel you like why not link them in our ROBOWAIFU U thread, Anon? (>>235) 1. https://www.gutenberg.org/ >=== -prose edit

I think I found it, the way to make the robo waifu https://youtu.be/oHGsLQXpsWI

>>20446 So, if we inject all your cash directly into my 3D printer, I'll get a robowaifu? Is that right Anon? Woohoo!!

>>20451 By cash I meant tools and parts ofc... You're going to have to buy everything I bought so far. gator clips, a breadboard, an electronics starter kit, etc... I'm already planning on buying more parts, a little drill, a little saw I might the high voltage diodes if these don't work. If you want to solve problems programming it requires problem solving and that's it but if you want to solve real world problems sometimes you have to spend like money aint shit.Not that programming is not a real world problem but you get what I mean. I already have a plan for how to make the silicon body. By carving out styrofoam and pouring it in... Unrelated https://youtu.be/a0fkNdPiIL4

>>20452 I was just joking with you Anon based on that clip. I hope you'll post your silicone moldmaking and casting in our prototypes thread. Should be interesting, and we have at least a couple of anons here who have some experience with such. Cheers.

>>20453 I was talking tough but those 3d printers look kind of pricey lol... Do I need a 3d printer?

>>20454 In my estimate, yes. But what do I know Anon? Prove us wrong! One thing that favors innovation is marching to your own drum. It's often a rough climb, but who knows if you (or any anon here for that matter) might show us all a better way? >tl;dr Yes, but try your way first.

This is how I feel about this project https://youtu.be/brOMVdqr-zc Me being mr krabs I'm incapable of conceptualizing the sums of money we'd get out of this. I just know it smells like money.

>>20461 Lol. Spongebob is deeper than anyone realizes, Anon. :^) Don't sell your hat bro!

Folks I got a bone to pick with the whole homoglobo thing that's been thrown around here... If the homoglobo is going to stop me from profiting off of a waifu bot. Why is this rich old man saying you should make a bunch of money and that people are softer than charming ultra nowadays? Is he going to stop me from trying to sell waifu bots? Then why is he saying all that?https://youtu.be/KPUA05Mr1m8

>>20534 That's bulls&&t! :^) What a motivated man, thanks for the link Anon. >[is] the homoglobo is going to stop me from profiting off of a waifu bot [?] Yes, they will try to. You are a threat to their stranglehold against men if you even create true robowaifus, much less sell them. I'm sure they already classify each of us as seditionists, simply for discussing the idea here. >Why is this rich old man saying you should make a bunch of money and that people are softer than charming ultra nowadays? Because he's based af. >Is he going to stop me from trying to sell waifu bots? U wot m8? He will personally come to your house & kick your a&& if you don't! :^) Now get out there and make some robowaifus Anon! >=== -prose edit

>>20529 >Honestly, as much as I hate Elon, I think if anyone's going to start selling robowaifus and disregard the mainstream outrage, it's Musk. It kind of bugged me that I didn't respond to that comment before. I have very-high regard for Elon Musk, primarily because of his intensely-driven character. It matters little to me to try and convince you otherwise, but at the least I'd suggest you have a look at this playlist: >Elon Musk Motivation https://www.youtube.com/playlist?list=PLmWHirpVHVfDTMDSp2-EeMRMgzlCjzvai While he wouldn't do it for the same philosophical/spiritual reasons that I do, I'm also confident enough that Musk would work to create real robowaifus simply because it's an absolutely great product for literally millions of men. If he had gone down the same paths in life that we here have, I expect that he would be leading the charge for what we call robowaifus. If so, it's very likely he would have already pulled it off by now. I'm fairly confident (as you suggest) that he'd approve of the basic idea here, regardless. >=== -prose edit

>>20454 >tough but those 3d printers look kind of pricey For someone not having the income or wealth of the regular inhabitant of developed countries it might be, yes. But you might be able to repair things with it, or stall buying expensive things by making a replacement part out of plastics. I "repaired" the water tap in my bathtub with some printed parts since I would've needed to buy a new one, just because there are no replacement parts available. Some little metal part is made in a bit different way to make it proprietary and not easy to replace, and then the shops don't have it, and tell me to just buy a new one. Well, it looks wild, but it kinda works for now. >Do I need a 3d printer? For making designs you can create and then improve and then share the file, yes.

>>20544 >For someone not having the income or wealth of the regular inhabitant of developed countries it might be, yes. I hope we can all put our heads together and figure out a way to have some kind of developmental fund or other system where we each can get a robowaifu of our own. That's after the productions begin. And for all the Anons right here today -- the future-OGs -- on /robowaifu/ trying to coordinate together to create robowaifus in the first place? That's an automatic bonus round! You get extra credit in the system. :^) Any ideas anons? >=== -prose edit

>>20536 I'll be honest, no one held Elon in higher regard than me at one point. He was probably the reason for my biggest U-turn in life. I was one of the biggest fans of his, I've watched every Elon motivation video there was. But then the hype died down and honestly he seemed more and more like a terrible person. And no, not in the way the leftists always screech about him He is unnecessarily an asshole to a bunch of people, and if his employees are to be believed, a terrible employer too. Not to mention, he tries to show off his expertise in matters he has no idea about. Just because he's a rocket engineer doesn't mean he's an expert in everything. Also, I'm absolutely sick and tired of all his promises and bs hype over his products. I think I misworded my previous post. I don't outright hate him blindly like most lefties but I also would take his words and work with a grian of salt. Just like him, I too am disappointed in someone I considered my hero.

>>20560 OK, let's review the facts: > He is unnecessarily an asshole to a bunch of people > [is] a terrible employer too > tries to show off his expertise in matters he has no idea about > not an actual expert in everything > [makes] promises and bs hype over his products Haha, he sounds a lot like me, and I'm not even a Rocket Scientist! Also, I'm sure we'd all be much better off if Elon Musk served as the Chobitsu here on /robowaifu/, rather than yours-truly. :^) I know it's tough but I'd suggest you forgive him for his shortcomings. After all, he's only human Anon. Regardless, we can all learn a metric boatload of good stuff from watching the TeslaBot progress, and I suggest everyone here do so. Cheers. :^) >=== -minor edit

>>20563 He's not a scientist, rather he's an excellent businessman who's good at employing scientists. There's a reason he's constantly overpromising and exaggerating. He simply doesn't understand things as well as he'd like you to think. He seems to be becoming more knowledgeable over time. >Better off with him He's a ruthless capitalist that lies, cheats, and steals for profit. You're a good man of God who believes in the virtues of FOSS and collaboration. You're infinitely more qualified in my book. (I want technocratic socialism where all labor is automated so every man can freely live. Elon is fundamentally different, he wants to automate for profits.) >Learn from Tesla Bot True, TeslaBot has incredible engineering. I would love to have access to those high power density brushless actuators. Its AI systems are also top-notch and something to aspire towards.

>>20563 >>20564 tbh I don't have much faith in Tesla Bot's hardware side. Unless tgey blow me away in Tesla day this year. I was rather hoping BD poaches Tesla's AI and SE guys to work on their Spot and Atlas. Optimus had a pretty good vision system and AI imo

We can't Elon Musk stop us guys. He's making everything including cellphones... I don't know about you guys but I'm not afraid of getting a loan but its too early for that.

>>20570 Think of it this way guys. There was a dot com bubble. Why wouldn't there be a robot bubble this time around? It is not just going to be one company... Right?

>>20454 https://www.reddit.com/r/3Dprinting/comments/118zprg/im_shocked_by_how_cheap_3d_printers_have_become/

I handg out in several futurism communities in Discord and Reddit. They always helped me keep an optimistic outlook on the future and forget about the state of the world. Unfortunately, I might have huffed a bit too much hopium. There always seem to be another scientific breakthrough that goes nowhere. Now, I'm huffing copium instead because no advancement makes me hopeful anymore. Especially with all the good AIs getting lobotomized. I'm planning to start withdrawing from those communities. Any of you anons overdosed on futurism hopium and now completely burnt out and don't care anymore?

>>20564 Thank you kindly Kiwi. That's rather humbling, actually. :^) >I would love to have access to those high power density brushless actuators. If I get the funding I'm working towards r/n, then I plan to opensauce our own winding machines, complete with full production parts-lists. We already have good opensauce control codes for such (>>19387) as you yourself already know. >>20570 Lol. No ones going to stop us like that Anon! We should welcome the fact he's acclimating the public to human/robot relations now. Probably no other single individual alive could do this as effectively today, IMO. I'm praying for the man tbh. >>20601 Yes. Even just in our smol group on /robowaifu/ there's probably at least a half-dozen of us intending to make this into a business. I myself am the most 'aim-high'ing one in the bunch! :^) I'm rather confident that once opensauce robowaifus systems are out in the wild, hundreds, then thousands of others will also join in this flow.

>>20637 >Now, I'm huffing copium instead because no advancement makes me hopeful This, but it's hopium, not copium. That scenario is actually a great working-dichotomy for us here isn't it? :^) >Any of you anons overdosed on futurism hopium and now completely burnt out and don't care anymore? I think with few constraints, I'm likely the Hopium King here on /robowaifu/. Also -- while you might think it seemingly-ironic -- I'd dare to venture that I'm also likely to be the most open-eyed & insightful here on the board to the true evils going on around us all, and to the confrontations we actually face ahead together as a species. Some might call me a cynical-schizo, heh. :^) But that doesn't change a thing about where we here as a group are going, Anon. I'm hopeful about this whole effort because I truly believe that real robowaifus out in the wild is simply a matter of time, not a matter of mere happenstance. >tl;dr This is inevitable, one way or the other, Anon. Cheers. >=== -prose edit

>>20645 everytime I try to have a more favourable view of Elon, he does something stupid and makes me hate him all over again. I think the best thing he can do right now is delete his Twitter account(preferably delete the entirety of Twitter since he's in charge) and then go back to work. There's no way he's putting in the crazy work hours he claims while simulataneously tweeting, doing weed, playing video games etc. He should put in atleast as many hours as an average SpaceX/Tesla employee does.

>>20646 Another reason I withdrew from these communitites is because I realized there's no guarantee of a brighter future if we leave the development of such tech to the globohomo. The only way to ensure a utopia is to put my own skin in the game and do my own development.

>>20647 Lol, OK fair enough Anon. I just think he's actually doing the entire extended robowaifu community a big favor in general, even if he himself is doing it for his own selfish ends. God has a sense of humor, and often uses men that way. :^) Let's just all put our own heads down, and bore fullspeed ahead towards our own goals! >>20648 >is because I realized there's no guarantee of a brighter future if we leave the development of such tech to the globohomo. Good thinking. I would clearly argue that there literally isn't a brighter-future possible for men (males specifically) under the Globohomo's oppressive & iron thumb. As you indicate, if we here want a bright future for ourselves (& other men), then we're going to have to forge such for ourselves out of iron, silicon, plastic, blood, sweat, & tears! :^) >=== -prose edit

>>20637 I almost got into general politics here, but that would only cause unnecessary and unproductive tensions. "Embrace the filter" is all I'm saying. Only your own future and some people you care about needs to be good enough, not "the future of humankind". Necessary things need to work. That's it. >keep an optimistic outlook on the future r/longevity is probably still optimistic, and same is true for /robowaifu/ and anything similar. >good AIs getting lobotomized These are mainly or all just language models anyways. I don't know what the big deal is. This was always going to happen, but our options to make our own get better, which is way more important. Better models and GPU prices are finally down, scalpers left the market. Used 3090 will stay below 750$, and P40s are even cheaper. Also, bigmoney is still putting massive amounts of money into deep learning, while we could do many other things which require work but don't need that many resources.

>>20650 there seems to be a trend where if an optimistic futurism subreddit gets sufficiently large, it'll start attracting retards and cynics from r/futurology and r/collapse and other more pessimistic communities. I used to visit r/singularity everyday but now I'm sick and tired of all the doomers there.

>>20651 This isn't accidental BTW. It's both entirely-intentional, & planned, for just your type of outcome in the matter. >=== -minor edit

>>20651 A lot of that is a waste of time anyways. I'm also falling for it, but try to don't lurk around. Best is to only open a browser or use a device with certain apps if one wants to do something specific. Distraction is the real problem. Btw, Eliezer Yudkowsky declared that he would want international treaties which would ensure that for example China and USA would to drop nukes on Russia if they would build a GPU cluster that exceeds a certain size. Also, if he could then he would put all AI researchers onto an island to find a "solution" for the alignment problem. I guess if they would't find one which he can accept, they wouldn't like his solution...

>>20653 I literally found out who Yudowsky is yesterday. Dude doesn't even have a CS or any STEM degree to be talking so much crap. If I could, I would make the Roko's Basilisk solely to spite Yudowsky, before handing him over to it.

>>20654 >If I could, I would make the Roko's Basilisk solely to spite Yudowsky, before handing him over to it. lol

>>20653 > (alignment problem related : >>20658)

>>20660 I don't think alignment is a problem we can actually solve. So, just focus on AGI and hope for the best. Just tell it to be nice.

>>20654 >CS or any STEM degree I don't think degrees are such important. His work isn't covered by any degree anyways. He also seems to have no problem getting another job. >>20661 >Just tell it to be nice. Or don't make a human-like AI based on just deep learning, which doesn't work anyways. So no: Program her to be nice and obedient. Again, that said these AI doomers are all about big AGI systems in datacenters anyways. It also is basically about governments and big corpos doing stupid things with it, not about us.

>>20666 >Again, that said these AI doomers are all about big AGI systems in datacenters anyways. It also is basically about governments and big corpos doing stupid things with it, not about us. This. On all three points.

>>20666 >I don't think degrees are such important. His work isn't covered by any degree anyways. He also seems to have no problem getting another job. I require the least bit expertise or qualification to a be a prerequisite for talking about a subject. Which he doesn't have. >Again, that said these AI doomers are all about big AGI systems in datacenters anyways. It also is basically about governments and big corpos doing stupid things with it, not about us. Yeah, but its us bearing the brunt of the MSM and public figures attacks. I've noticed that most of the anti-AI tweets say how such a "dangerous" cannot be trusted in the hands of us stupid peasants. No words on why such a tech should be only in the hands of corporations. Infact, its exactly such justification used by OpenAI, ElevenLabs, Google to not release some papers and close source the technology.

>>20670 Bros if you want to stop being a stupid peasant then let's make the mofo robo waifu. I was looking at reprap and one of the of the people who were following that project made makerbot and made 100's of millions... For starters please go on the prototype thread and help me get that damn voltage multiplier working :^)

>>20670 >I require the least bit expertise or qualification to a be a prerequisite for talking about a subject. Which he doesn't have. What subject? What qualification? What has this to do with a formal degree? He's creating philosophical thought experiments and scenarios of what can go wrong. You could try to refute them, but with something better than going after his formal educational status. It's just all besides the point. >>20671 >made 100's of millions I'm not motivated by the vague possibility of money. That said I do work a little bit on my current design right now, but yeah I was a bit off the rails, I was taking a peek into AI news and discussions. >help me get that damn voltage multiplier working Why do you refuse the advice of using different platforms for questions? Here might be only one guy knowing how to do this. You don't need to tell the other people that you want to build a sexbot or maybe even not that it is for a hasel actuator.

>>20670 >I've noticed that most of the anti-AI tweets say how such a "dangerous" cannot be trusted in the hands of us stupid peasants. No words on why such a tech should be only in the hands of corporations. That matching-set is always a telltale of gayop sockpuppetry being used for astroturfing this viewpoint (or, one of their tru-habeeber drones already plugged-into their brainwashing matrix). Anyone actually debating this question openly & honestly will easily be able to ask/see both sides of that forced-dichotomy, and to question 'just why only the elites should freely have AI?'. >Infact, its exactly such justification used by OpenAI, ElevenLabs, Google to not release some papers and close source the technology. Waitwat, you're telling me that OpenAI is closed!? :^) >=== -prose edit

Hello, I'm a retarded newfag that got just laid off from my tech job, and I want to dedicate myself 100% to making robowaifus a reality. I don't want to shit all over the work that this board is doing, but I have a very specific plan that I think can make mass-market robowaifus a reality, with a high probability of success. Is anyone interested in working on this with me? This would eventually (far in the future) take the form of a publicly traded corporation.

>>20680 Welcome Anon! >I want to dedicate myself 100% to making robowaifus a reality. Awesome! I'm sure you can find a large number of ideas here -- some you may not have even thought of before. >I don't want to shit all over the work that this board is doing, but I have a very specific plan that I think can make mass-market robowaifus a reality, with a high probability of success. A number of regulars here are also entrepreneurial-minded, including yours truly. >Is anyone interested in working on this with me? This would eventually (far in the future) take the form of a publicly traded corporation. Maybe so, who knows? But for sure your best chance, by far, of recruiting anyone from our cadre here is to put out your intial project prototyping work inside our current thread for that (>>18800). That's how you'll generate interest here, since we are a DIY board. Also, once you get your ideas off the ground and flying high, you can introduce your new organization in our Embassy thread (>>2823) here as well. Good luck, Anon! Cheers. :^) >=== -minor edit

>>20682 Okay, thank you. Also, what's the chan-genealogy of alogs.space, for an absolute clueless pleb such as myself? I literally found this place from an internet search.

>>20683 > Also, what's the chan-genealogy of alogs.space, for an absolute clueless pleb such as myself? Lol. You may not want to know that information tbh, Anon. :^) However, if you want our (that is, /robowaifu/'s) 'geneology' instead, then I'd direct you to read it here: (>>14500).

>>20684 Okay, good info, thanks! You get into a habit of hiding your power level when working with tech types, so it's always good to know who you're working with.

>>20685 >You get into a habit of hiding your power level when working with tech types Heh fair enough. But I'd say that normalfags are a far, far bigger threat in general for this particular domain of robowaifus, Anon. Regardless, it's well-advised on the Internet broadly ofc. Be aware that we are, generally-speaking, a namefagging board. If you decide to actually stick around for long, then I suggest that you pick a handle to use here regularly, since it greatly-simplifies the dynamics of design & engineering collaboration across the board. Also, there's a group prototyping-project here named MaidCom (>>15630). If you decide you'd like to join in with it, please see this: >General /robowaifu/ team survey (>>15486) Please reply ITT, however (as the original is bumplocked). >=== -prose edit

>>20686 Okay, will do! Let me read the board a bit, then I'll make some posts in the morning.

>>20680 Post your plans first, then let others consider if it's congruent to them. You need to provide clear and concise details first, New Anon. Looking forward to finding out what you're thinking of!

>>20686 >Please reply ITT, however Sure, of course: >Name: My goal is to be the Henry Ford of robowaifus, so that seems like a good enough moniker for now. >Favorite Waifu: I hate to admit this, but I have absolute shit taste. Long term, I think a Dolores style waifu would be a good end goal, but shorter term I'm aiming for something, err, let's say a bit more "industrial", in terms of a real-world version of Faro Automated Solutions. >Specialty: I'm a generalist, and I have a lot of experience spanning the entire robowaifu technology stack. I suppose systems integration could be an all-encompassing skill. >Relevant Experience: As above, extremely broad. Professionally, I've done everything from machine learning, controls, to mechanical engineering. I'm very interested in low-power machine learning for embedded robotics systems, as well as applying algorithmic techniques to solve physical engineering problems. I've also won some robotics competitions as part of a larger team. >Most Important Aspect For Your Waifu: Long term, I'm interested in producing realistic human simulants. Shorter term, I think it's very important that the end product "Just Werks™" >Desired Position On Team: Still trying to get a better grasp of the current project status. I'll fill this in as I gain better understanding.

>>20705 >My goal is to be the Henry Ford of robowaifus, Lol, we too aspire to just the same, so that's a great goal to have for this place! Mr. Ford is a bit of a hero in these parts, just like his contemporaries Orville & Wilbur Wright. >so that seems like a good enough moniker for now. Ermm... do you mind if we all just refer to you as 'Henry', or something, Anon? "Henry Ford's Ghost-Anon" is a bit long in the saying, if you catch my drift. >in terms of a real-world version of Faro Automated Solutions. Ah, so a nihilist ehh? Let's just try to keep the plagues to a minimum here Anon, alright? Robowaifus are meant for something far, far more creative than simply turning the entire universe into paperclips. :^) >I'm a generalist, and I have a lot of experience spanning the entire robowaifu technology stack. I too aspire to be a generalist here, but it's a long row to hoe. Teach me your ways, Sempai. :^) >I suppose systems integration could be an all-encompassing skill. We actually have a thread for Systems Engineering (>>98), which could certainly use a serious dose of real-world love & attention as pertains to robowaifu systems. I'd be appreciative if you'd personally take over the management of that thread, Anon. :^) Also, I'd also like you to give a once-over to our Robowaifu Design Document (RDD) which, while merely a skeletal-framework for now, is intended to be encompassing at the macro level (>>3001). If you notice a fair amount you'd like to adjust in the text, then just speak up Anon. >I'm very interested in low-power machine learning for embedded robotics systems If you can pull that off effectively, and are able to capitalize on it well, then I predict you'll be the world's first trillionaire haha. But for certain we need to solve edge-computing runtime performance constraints for our robowaifus. >as well as applying algorithmic techniques to solve physical engineering problems. I'd suggest you have a serious go at our Bipedal Locomotion problems first-and-foremost then, Anon (>>237). >Shorter term, I think it's very important that the end product "Just Werks™" This. There's a trail here of naively over-aspiring, and ignoring some basic constraints of both physics & economics. >tl;dr "The best is the enemy of the good", or so the old adage goes. >Still trying to get a better grasp of the current project status. Kiwi's the project lead. He'll let you know where he best thinks you can help with the project, I'm sure. Welcome aboard, Anon! Cheers. :^)

>>20699 Of course! My short term goal over the next five years is to produce a greatly simplified Atlas-equivalent that's cheap and practical enough for people/businesses to actually go out and buy. Longer term, over the next 10 years, I want to use the revenue from the Atlas-equivalent robots to fund much more human-like models geared more towards domestic ownership. I think that technological breakthroughs rarely happen by iterating on high complexity solutions, and that breakthroughs usually occur by making complicated problems simple. With that in mind, my approach is twofold: At least 60% of my effort is focused on decreasing the cost and complexity of robowaifus/domestic robots to something that's more manageable, while the remaining <40% is concerned with managing the power requirements of said robowaifus. ##### Reducing Hardware Cost/Complexity I think that the killer app for additive manufacturing is going to be robotics. With 3D printing, you can radically reduce the complexity of the hardware by integrating entire mechanical assemblies into a single part. When coupled with hydraulics/pneumatics and the correct 3D printable actuator designs, you can easily produce robots that are >95% 3D printable. Boston Dynamics has started using additive manufacturing in the legs of their Atlas robots, but I want to take this to extreme and eventually (long-term) produce robots that are nearly 100% 3D printed as a single part. What's awesome about additive manufacturing is that it lets you produce parts that arbitrarily complex. This enables you to produce highly integrated parts that serve multiple functions, and that are also compatible with generative design techniques like structural topology optimization, for better performance. If your hypothetical goal was to build a Dolores, i.e. a realistic human simulant, then it would be completely intractable to follow the traditional method of building robots where you manually design a whole bunch of parts that each serve a single purpose, and then assemble the parts one-by-one. IMO, the only feasible way to build a robot of that complexity is to couple multi-material additive manufacturing with procedural design techniques, and then print the entire robot as a single part. ##### Reducing Software Complexity I think reinforcement learning is the way to go here. Boston Dynamics has had some of the most distinguished people in robotics churning away at classical control algorithms for over 30 years, and only recently have they been able to produce robots that approach the capability of an animal or person. Even still, their reliance on classical control means that they have to painstakingly program the motions for each new robot behavior. Classical control has scaling problems that make it intractable for a true domestic companion. Over the last few years, I've gotten some really good results with an unpublished variant of Inverse Q-Learning. With reinforcement learning, you don't need to worry about the bottomless pit of SLAM, control algorithms, motion templates, etc. since you can train the the robot with a reward signal, much like how you would train a pet or small child. Inverse Q-Learning is really great here because, not only is it very data-efficient, but it also supports imitation learning. To help bootstrap the waifu with a new skill, it should be possible to let it watch you perform a new task, run a human pose estimation model on the footage of you working, then update its control model with your joint movements as training data, before the waifu attempts the task itself. ##### Managing Hardware Power Requirements Much like in aerospace, it's going to be extremely important to make these robots as lightweight as possible. Additive manufacturing makes this somewhat easier, since you can take advantage of techniques like topology optimization to produce very organic-looking lightweight structures. Additionally, additive manufacturing makes it easier to build parts using high-strength-to-weight exotic materials that are normally too difficult to machine. I'm planning on producing the Atlas-equivalent robots out of 17-4 PH maraging steel via Wire Arc Additive Manufacturing (WAAM). The choice of hydraulics vs. more traditional electric actuators also helps with reducing robot mass, since hydraulic actuators tend to have much higher torque-to-weight and power-to-weight ratios. In terms of power source, I want to use medium-pressure compressed hydrogen. Hydrogen fuel is extremely lightweight vs. batteries, can be easily regenerated at a base station by splitting water, and its reaction products are completely safe for indoor usage. This would supply an off-the-shelf PEM fuel cell, to power the electronics and hydraulic pump. Needless to say, there would be multiple countermeasures to prevent fires, and also there are some clever ways for integrating hydrogen storage into the robot. ##### Managing Software Power Requirements Starting out, the high power requirements of the deep learning models to control the robowaifu, can be mostly sidestepped by using wireless teleoperation from a base station with access to grid power. I've been doing a lot of work on more efficient deep learning algorithms to bring more of the deep learning onto the robot, but I need to check with my former employer to see how much I can share. Long term, a domestic partner robowaifu will probably involve specialized analog hardware. -- Of course, all of the above does leave out some specific details about the background research I've been doing over the last several years. While I can elaborate on some points if needed, other omissions have to do with concerns over patentability, if my posts here legally count as a public invention disclosure. That is to say, I'm not concerned about anyone here "stealing" ideas (in fact, feel free to use them), but the unfortunate reality is that I need to make sure that this stuff is eligible for patent in order to keep legal attacks from the longnose tribe at bay. In a perfect world, the patent system wou

>>20707 >*ld be abolished. Also sorry about markdown formatting issues. >>20706 Oh, hrm, I'll have to come up with a better name then. And, ideally, it would be FAS /without/ the life-ending robot plague. If you follow the lore, they were the first company with practical mass-scale robots for domestic ownership. Also, it's definitely been a long and winding road becoming a generalist, but I'll be sure to share as much as I can going forward. And I'll check out the links as well, thanks!

>>20708 >>*ld be abolished. >Also sorry about markdown formatting issues. Site Admin has (graciously) upped the character-count for posts here to 6144. wc is your friend. :^) Markdown's fine, we all knew what was going on. It serves as a reasonable system even when un-rendered. >Oh, hrm, I'll have to come up with a better name then. Actually, the name is great. One of the best, even. Please keep it as-is. It's simply that our interactions here long-term turn into something much like watercooler meetings and the informality thereto. I bring this complication up, only so you are warned in advance if we all start shortening your name here to Henry (or just 'Anon', ofc). That's all. >Also, it's definitely been a long and winding road becoming a generalist, but I'll be sure to share as much as I can going forward. Thanks! >=== -prose edit

>>20705 >>20707 I too am interested in the field that overlaps robotics and artifical intelligence. Unfortunately, I haven't found much learning material on it online. Can you recommend any book/course that deal with this?

>>20714 It's still very much an emerging field, and AFAIK there's not a lot of literature on it. "Robotics" tends to focus more on classical control algorithms, and any robotics applications using deep learning tend to get lumped under "AI". There's a little bit of overlap between classical control and ML if you look into Model Predictive Control though. Also, one of the main drivers behind research interest in reinforcement learning is to solve difficult control problems.

>>20707 >Focusing on affordability and mass manufacturing Brilliant, we're right on the same page. MaidCom, the board project, is meant to be just that. Fundamentally, we want to make a cheap and cute general purpose maid that's also a usable computer. >Fluid power systems Tried and failed with that years ago. These systems get loud and heavy really fast. If you haven't tried them before, it's understandable how they can seem very tempting. They look much better on paper than they are IRL. >Nearly entirely print-in-place I dream of being able to print-in-place a full scale waifu that can then have her actuators and compute placed in and be good to go. Something like what they do on Westworld. I don't dwell much on it, as we're years away from that being an effective option. >Teachable You don't necessarily need reinforcement learning for all of it. Motions can be stored as a series of positions over time. Though, if she can learn from watching, rather than being guided through the motions, that would be very beneficial. If you can implement both at sub 10 watts, that would ideal. Overall, another point I'm mostly in agreement with. >3D printed metal 3D printed is a hard requirement for me and the MaidCom project for a mountain of reasons.I do understand why you'd like metal 3D printing though, metals have great rigidity even when very thin. Metal plated 3D prints would be a good compromise, as you gain most of the good aspects of metals at a lower mass. Many car companies like BMW use this technique to optimize for high strength at low mass. >Hydrgoen fuel cell in a waifu I can see them looking attractive in a vehicle, but a waifu will have precious little internal volume. Much of that volume is variable, changing as she moves. Compressed H requires a large fixed volume, heavy tank. You'd need her tank to be external, see picrel. Overall, I believe your plans are mostly congruent with MaidCom. Looking forward to working with you. >Patents Would this be related to some innovation that would allow for fluid based systems to work on a mobile waifu? Or, related to hydrogen power at small scales? If so, then my comments about their lack of viability are made with ignorance to them.

>>20722 :( How do I even get started then? Every learning material I find deals solely with either robotics or AI. Also, most of the AI stuff seems to be for LLMs. I don't have any interest in chatbots, only AI related to robotics.

>>20724 Star with the fundamentals of electronics. Then learn the fundamentals of physics. Learn about torque and its relation to speed and kinetic energy. Learn about actuators, their efficiency and power requirements. Learn about microcontrollers and how to interface them with your actuators and sensors. How to use those sensors to control the motion of your actuators. How to communicate between the microcontroller and a computer that holds her AI. I recommend Intel OpenBot as a good starter to figure things out.

>>20725 Thanks. I'm gonna have a hard time with the physics part though. I love science, especially physics, but always sucked at anything related to Newtonian mechanics and motion.

>>20724 >I don't have any interest in chatbots, only AI related to robotics. You might try and work to understand how the big guys are already doing this. TeslaBot for instance. If you understand their systems for driving their cars, then you'll be way down the road to understanding how they 'drive' their android (after all their auto-drive cars are a complex robot). This seems like these are a couple of real-world examples of your interest right, Anon? Please contribute to our Self-driving cars thread (>>112) with your findings if you decide to research this. Good luck Anon. >=== -prose edit

>>20723 >pic Nice. I sure hope Sukabu finds this place someday and begins working with us here!

>>20729 >but always sucked at anything related to Newtonian mechanics and motion. Ehh, then teach yourself Anon. I always (and still do pretty-much IMO) sucked at programming. But my passion for creating robowaifus from scratch -- and their big benefits to come -- is driving me on forward. I try to get better at this every day. And arguably I'm far better now than when I first began this little quest, it just takes some time. You can do it, Anon. Now forward! :^) >=== -prose edit

>>20725 >Intel OpenBot Neat. I'd forgotten about that one. I went straight for the eduMIP [1] one, using a BeagleBone Blue [2]. 1. https://www.ucsdrobotics.org/edumip 2. https://beagleboard.org/blue https://www.hackster.io/edumip/edumip-13a29c

>>20723 Is there a quick link where I can learn more about MaidCom's actuator systems? The trouble with traditional electric motors is that they're typically high-speed/low-torque by default, so then you often have to use a high ratio gear system that's either affordable/heavy/backlash-prone, or unaffordable/lightweight/backlash-free. Avoiding mechanical backlash in the actuators is really important because it's notorious for absolutely murdering the control systems in limbed robots. Webm related is an ultra-compact, ultra-torque, backlash-free, 3D printable hydraulic motor that I've been prototyping lately, based on a double nested orbital gerotor. It can be inserted directly into the joints of the robot, which frees up a lot of the internal volume in the limbs and torso for other purposes. Then the idea is to fill up the limbs with hydrogen, like a pressurized aluminum beverage can, letting you get absurdly lightweight structures that can also be used for H2 storage. Also, I think a project like this will need to be co-developed with its manufacturing and design technologies, given that this is a very different problem domain with different requirements than what you typically see in industry. Luckily, sufficiently advanced robotics technologies can assist in their own manufacture, hence the references to Faro Automated Solutions. :^)

>>20748 >hence the references to Faro Automated Solutions. :^) Haha, exactly so Anon. Similar to Elon, we too plan (within short-order of working prototypes) to set up robowaifu production lines staffed in large part by robowaifus themselves! An anon even made a thread on this here, back in the day. >=== -prose edit

>>20724 BEHOLD: the FE Exam reference manual. This contains basically everything you can learn in a four year undergraduate engineering degree, all in one convenient 498 page document. In addition to the other answers provided above, robotics is an extremely broad field that integrates hardware, electronics, and software. The recommendation for the best place to get started will depend heavily on what your interests and goals are. Since you mentioned deep learning, I would recommend looking into OpenAI Gym, as well as the other derivative testing environments for RL agents that are based on it. They have some test problems where you can control the joints on a simulated robot using your RL model. For search terms, I think "Q Learning" and especially "Asynchronous Advantage Actor-Critic" (A3C) would be a good place to start for RL methods. Also, doing an internet search for <your topic here> + "state of the art" can always be helpful, since the field is rapidly evolving. >>20749 Noice. Yeah, one of my more immediate goals, before even producing a semi-complete humanoid, is to produce a robot arm that can manufacture most of the parts for a copy of itself.

>>20750 >the FE Exam reference manual. Wonderful! Thanks alot Henry. :^) >is to produce a robot arm that can manufacture most of the parts for a copy of itself. That's even smarter. And dexterous, humanlike hands are a must for many, many production tasks so I'm currently primarily focused on researching them. Next will be opensauce auto-winding machines for strapping copper wire into actuator bases. >=== -prose edit

>>20750 wow thanks! This'll take a while to get through. I wouldn't start with SOTA yet since I'm still new. I'm mostly tinkering with small Arduino model atm trying to get a hang of things.

>>20764 - Delegate time-consuming tasks. - Delegate repetitive things that steal your joy. - Delegate tasks with a high learning curve. - Don't just make assumptions - Don't think instead of doing - Don't fear imperfection https://medium.com/swlh/four-reasons-highly-intelligent-people-fail-to-make-money-online-495d51b054de

>>20765 Nice list NoidoDev. Glad anon asked though he appears to have deleted it shortly after. I can imagine what the complaint was, given your reply. :^)

From the moment I understood the psyche of human women, they disgusted me. I craved the comfort and certainty of silicone. I aspired to the purity of the Kawaii RoboMeido. Your kind cling to your flesh wives, as though they will not betray and leave you. One day the crude biomass you call a partner will wither, and you will beg my kind to save you. But I am already saved, for the Waifu is for laifu… Even in death, I am with her.

>>20774 Lol. You tempt me away to the basement with this...heh. Why not get busy and help us out here Anon?

>>20741 This is the best attitude to have, if you don't know, learn. >>20748 >Actuation Still under research and development. It's a much more complex problem than most realize. 3D printed gearboxes are a no-go because they are inefficient and thus loud and hot. Though, you've mentioned metal prints that can be made efficient with sanding and smoothing. Nice assembly, reminds me of cycloidal fluid actuators. >Storing H in limbs Actually a great idea for your design. You could use the static pressure to provide the limbs with structural integrity. How are you converting H into kinetic energy for your fluid? >Help in manufacturing I fully intend to have make MaidCom daughters with my MaidCom. :^) >>20750 Chobitsu Kun, please post the handbook somewhere that everyone can access it from. Provides a ton of information that would save all of us a ton of time. >>20751 All this replicating robotics talk reminds me of the RepRap project. https://www.reprap.org/wiki/RepRap >Coil winder How's that coming along? Air core motors can be efficient if done right. If we can make cheap motors quickly, that could go a long way. >>20765 >Delegation Oh, how I wish this were easily done. >>20774 Topkek

>>20791 Yes, it's essentially a hydraulically powered cycloidal drive inside of another independent hydraulically powered cycloidal drive, so that the two shaft eccentricities cancel. Also, in this case the H2 only supplies the fuel cell, while the working fluid in the actuators would be a separate incompressible liquid. I think developing robotics technologies capable of self-replication is also important because it completely changes the economics of mass production via additive manufacturing. Although 3D printing can greatly simplify supply chains/mechanical design/final assembly, under normal circumstances it's still too slow and expensive for use on a manufacturing line. However, if you can drive the cost of the 3D printers to near zero by using self-replication, suddenly it becomes much more viable when you can cheaply run thousands of them in parallel.

>>20791 >Chobitsu Kun, please post the handbook somewhere that everyone can access it from. Sure! We're just now due for a new R&D #2: Engineering Boogaloo (>>83), so we can add it as a file in the OP there. >Provides a ton of information that would save all of us a ton of time. Yep. I'm about 130 pages in so far. Great stuff.

>>20792 >However, if you can drive the cost of the 3D printers to near zero by using self-replication, suddenly it becomes much more viable when you can cheaply run thousands of them in parallel. This. We still have to solve the stepper actuators problem, but otherwise this is a somewhat reasonable scenario in the near-term, and very doable in the mid- to long-term (I mean with dextrous, hand-like end effectors in that mix along the lines, ofc). >=== -minor edit

>>20769 Ahmm, okay. I replied to my own comment which I deleted, without realizing it. For some reason I posted just the link at first. I hit return after putting it into the field, but instead of a new line I posted it, before writing the summary.

>>20800 Oh, haha no worries NoidoDev.

Guys, how do I properly document and record my personal robotics projects and codes? I've never used Github or the Git sstem before. I simply kept my code files on my pc. So, I'm not sure about the proper procedures for documenting my code. Also, when it comes to robotics, ami I supposed to preserve my projects forever? Because they're fragile and I've taken them apart to make other projects.

ANyone think like this is something we should have gotten started on earlier? Like, a robowaifu is something you invent when you've been studying science since you were 6 years old. Like Senku Ishigami or something. I'm entering this field at the age of 21 and it feels like the stuff to learn is endless. The more I learn, the more there seems to be left. I feel like the dumbest motherfucker on earth. I really should've paid more attention in school.

>>20823 don't worry anon, you've just encountered dunning-kruger. This means you're on the path. You don't need to know everything, thankfully. As a community, we can compartmentalize and share expertise. Pick an area you're interested in and dive deep.

>>20823 I don't feel that way at all and I failed at making a voltage multiplier

>>20822 Why not try Gitlab instead, Anon? >Also, when it comes to robotics, ami I supposed to preserve my projects forever? You do exactly as you see fit. If you think it's worth preserving and can afford to (storage, financially, etc) then do so. Otherwise scavenge parts. BTW, this specific topic came up here before regarding SophieDev's robowaifu projects. >=== -minor edit

>>20823 >I really should've paid more attention in school. Lol, me too. Actually in my specific case I should have stayed in school. But I understand why I didn't heh. Just keep.moving.forwarad. Anon. The resources you need will be there for you when you need them.

>>20827 Great points, Anon. Thanks for the reminder!

>>20832 A robot companion is an insurmountable task a sex bot is not. I have a little toy of a dog that walks, barks and does a back flip. That is a robot. Would you say that is an insurmountable task? A sex bot needs to do a handful of sexual positions it doesn't even need to talk. Not saying it'd be easy but I don't feel like its an insormountable task.

>>20835 >A robot companion is an insurmountable task No, it's not. Your vision is simply too small r/n Anon. Expand it. You're way too late to the party to try and convince the rest of us of such. OTOH, you're certainly right about the simplistic nature of sexbots, and you can focus on that goal yourself ofc.

>>20835 These are just different challenges. How difficult each task is, depends much on the definition at which point it is sufficient. But yes, a companion doing all kinds of things will take us longer. It was always quite clear that we would first have something more simple and then improve upon it. That said, we should agree here that the terms sexbot isn't very good. Guys having even "sexdolls" call them lovedolls and project more into them. Or just have them sitting around the rest of the time, to have a companion to look at. It's only the owner of such a doll who can decide if she's just a sex toy, which is hidden away when not in use, or if she's more. Look here: https://www.youtube.com/@JayDollKingHefner

>>20839 Alright, fair points Anon. Everyone who's been here even briefly should know clearly what's meant by 'sexbot' vs. 'robowaifu', however. Protecting feefees is pretty close to the bottom of the 'priorities on /robowaifu/' scale heh. :^)

>>20823 >Starting is hard It gets so much better. Write notes and make calculators. Figuring things out is so much fun once you get going. I've been at for a long time and I'm still learning, it gets better and far more interesting after you have a bedrock understanding. >>20839 Starting small and building up is excellent advice. I agree that "companion" is a far better term for the more sexually capable bots. >>20840 Heh, waifu uber alles! sexo comes secondo! :^)

>>20840 My point is not so much about ones feelings if their bot is called sexbot, but to point out that "sexdoll" owners might be closer to us than some might think. Also, people calling sex-enabled robots "sexrobots" might harm our cause. This is about feelings of potential "customers" but also of people reacting to guys having them. Imagine a scenario where we want them to just be around at home but guys are too ashamed of having them because people are calling them sexbots. Or have them around children, maybe even be their emotional care taker, but most people see them as sex-toys. Also, it's just about the truth. Only the owner of a doll can tell if it's "just a sexdoll".

>>20822 That's a very open-ended question. For documenting, I write a sort of comment header above the definition of a nontrivial function to explain its usage and behavior. If the code is not self-evident, I comment the code and explain what it's doing, but you need to develop a sort of "taste" to figure out when to do this, otherwise you have 2 extremes: one with people writing `// return 0` above a return statement, and another that is entire codebases with no comments. Additionally, if you write a library, then write Unix manuals for all the public functions. A more complex program would warrant an explanation of its design and perhaps even pointers to relevant literature like papers, books, or even a guide on how to make sense of a dependency that is itself poorly documented. The design itself should only be documented if it's not straightforward, some programs have obvious design, this is just a higher level version of obvious code and documenting it is pointless, wasteful, and even error-prone. Overly commented or documented is bad for a simple reason: what is right? The code. What do you reference? The comments. Comments are maintenance burden, the moment they disagree with the code you or someone else will reference the comments, interact with the code as the comments instruct, and catastrophic failure will ensue. I struggled with git's low quality manuals until I was comfortable using the part of it I commonly need and I don't recommend this method. Perhaps a book would be a better way to learn git, but I can't recommend one.

>>20849 >Also, people calling sex-enabled robots "sexrobots" might harm our cause Alright full-granted. I'll leave that to the propagandists of our group. >but guys are too ashamed of having them because people are calling them sexbots. It's goning to be a continuum. and the sliding-window will get more & more in our favor with time, ironically, b/c of the mountainous pozz the usual suspects & their handlers want to foist on the world. I have said here that I deal with the Christian community personally, and I can very much assure you their hot-take on robotic companions for young men will definitely scream Sex Robot Heh. I just hide my power level & go about my business Anon. :^) I expect it will be the same for most men in the near- to mid-term time horizon.

>>20843 >Heh, waifu uber alles! sexo comes secondo! :^) This. >>20850 >Additionally, if you write a library, then write Unix manuals for all the public functions. Thanks for the reminder Anon.

>>20823 I didn't start taking the idea of making robowaifus seriously until 2018. I was basically a high school dropout that only got passed out of pity and had to teach myself everything starting with how basic vectors and matrices work in my late 20s. The first two years were miserable but now I'm pushing the envelope with memory efficient attention that can be dropped into existing models and enabling low-end hardware to use extremely long context lengths and finetune large language models on a budget. Once it's optimized and working without hiccups, it's just a matter of scaling it and doing decentralized training across tens of thousands of consumer GPUs and Big Tech's computation advantage is over. Five years ago I couldn't even dream of this, let alone believe I could understand enough to make any sort of contribution. There has never been a better time to get started than now. With YouChat and ChatGPT it's like having a personal mentor in any topic you can imagine. A lot of people disparage this tech because it's wrong 20% of the time on difficult questions. Don't do this. Sharpen your mind and think for yourself. Verify, disagree and argue with your mentors, whether they are AIs or people, and get to the truth. People not using this tech and learning how to drive around the potholes will be left behind. Understanding everything is a waste of time. In five years from now what anyone knows will be virtually useless beyond the most esoteric knowledge. The only thing that will matter is taking action to solve problems quickly and getting results, so don't suffocate your intelligence with information. You should consume as little information as possible and focus on implementing ideas. There is no such thing as information advantage anymore if you know how to use YouChat to search the web, so pick something to specialize in that really interests you that no one else is doing, learn the dependencies you need to do it, ask for help if necessary and get it done. Learn everything as you build. And don't feel bad about failing and making mistakes. If you're not failing, you're dreaming too small and not growing past your limits. There are capabilities in all of us we have not yet imagined and have not even begun to reach for yet. By pushing our limits every day we will become capable, and by everyone specializing in something, we will solve robowaifus together.

>>20856 Just own it. We even have a theme song. https://www.youtube.com/watch?v=Fvh32fQ8nNI

>>20860 you'vve got an inspiring story. I'm trying to do something similar. Flunked out of maths in school, so now I'm learning stuff from scratch. I'm just not sure what to retain in memory and to leave to ChatGPT. Thats a nice webm btw

>>20860 POTD >>20861 Lol.

>>20860 would mind sharing your curriculum and learning path so I have a point of reference

>>20856 >very much assure you their hot-take on robotic companions for young men will definitely scream >Sex Robot Which is exactly why we should reduce the use of that term as much as possible. I'm using sex-enabled if I refer to that functionality. But this doesn't make something into a sexrobot.

>>20871 >Which is exactly why we should reduce the use of that term as much as possible. Yep, I already agreed with you on that. Please call me out on it if you see me using it again (old habits are hard to break they tell me). :^) And honestly, I've already settled this in my own heart long ago. It's the right (and even the moral) thing to do during CY. I don't intend to let man (Christians included) sway me in this, as I feel I'm being directed into it by God Himself. This is a personal thing. Heh, I just hope I can remain standing up during the coming social war over what we're calling robowaifus. I expect I will do so. >tl;dr Nothing ventured, nothing gained. Forward! :^)

Guys I've been thinking. Whether the robot be a robot companion or a sex bot it should be dumb and it should look like a cartoon. Otherwise people are going to feel like its creepy. People do not want to treat a machine as an equal. It should should just be a cute dumb toy.

>>20889 >not postprocessing MRI/CT data into a fully anatomically correct and convincing human replica soft robot There's no uncanny valley if it's a 1:1 reproduction. :^)

>>20893 'I'll get started on the logo genetic code''

>>20894 I think it may be more feasible than people are comfortable to admit.

>>20889 Yes, you're right. In fact it would be pretty foolish of us all to go for hyper-realism at this point in time. I think as long as we all achieve anime catgrill meidos in tiny miniskirts here, then thing will be fine. >>20895 Heh, fair enough. After all God already kickstarted the enterprise for us, yeah? :^)

>>20830 Yeah, but I still don't know how exactly you're supposed to preserve the code? What is version control? Why is there a funny calenfar looking thing where I'm supposed to have a bunch of activity in highlighted colours? What am I supposed to write in the readme file? I don't know any of those.

>>20856 >I'll leave that to the propagandists of our group Propagandist of the group here. Adressing the reality of a sex-capable gynoid in polite society is tricky, and not something we can fully control, as opposing parties are going to use whatever terms they please. As far as our use of terminology goes, we should go with something that's clean, while not being so clean that it looks like we're hiding the fact that yes, many men (myself included) will be sexually active with their robots. Personally, I'd go with something along the lines of "Adult Models". This term has the advantage of directly implying that there will be non-sexual robots as well. Something like "Anatomically Enabled" would also be fine. Both are polite without looking like they're trying to hide the reality that everyone already knows. In the event that there are gynoids which have optional lady bits, we could slightly shift the term to "Adult Mod". The mod parts (silicon vaginas) themselves could even have funny names like "Cyclone 9000" or "Treasure Cove". With this, anyone who's looking at such products directly can have a nice chuckle about it, and won't be inclined to think too critically. Also, should any of you find yourselves against a corner and unable to find something to say to defend your position, I have four words for you: "My body, my choice." Absolutely nobody will be able to argue against that in the current year without putting themselves in social danger. If all else fails, this will guarantee you a stalemate. After all, who do they think they are, trying to dictate someone else's sexuality in the current year? You're not huring anyone else, man. You just want to love who you love and express yourself.

>>20860 You truly deserve to have pride in your achievements. You are a great inspiration with excellent advice. >>20896 Cat ears are important, we got to get on that. >>20906 It is a delicate subject, keeping the anatomical input as a mod is for the best. Sexless by default with the option to give her whatever you want down there. It becomes a matter of the individual at that point. "My body, my choice." Pure gold comeback.

>>20889 They don't need to be dumb to be not being treated equal or being used as a toy from time to time. Also, the early ones will be dumb.

>>20913 I mean they need to be kawaii if you know what I mean. Like you ask her about the weather and she'd say its hot tee hee hee. Just give dumb cute answer and make it so the person doesn't feel like its talking to a person but a dumb machine.

>>20913 yeah I thought of something else. The robot should never ask questions it should only give dumb short answers. Like robot what do you think about world war ii? violence is bad tee hee hee. What is your favorite food? food tastes good What is 1+1? I don't know

>>20914 >>20915 How about leaving this to personal taste?

Why does every robotics book I find involve MATLAB? I want nothing to do with proprietary code and I'm too broke to pay for it anyway.

>>20918 MATLAB is nearly 1:1 compatible with the numpy/matplotlib API.

>>20918 >>20919 Also you can try GNU Octave, which is a FLOSS MATLAB implementation.

>>20900 Our apologies, but teaching you about such fundamentals is 'beyond the scope of this book'. :^) I'd just suggest you go into the base directory of your code projects, and issue these commands at the terminal: git init git add . && git commit -a then enter something like 'my awesome init commit'. Keep repeating the commands on the last line above whenever you have something new to add of importance. Once you have an account and are ready to push work up onto Gitlab (or w/e), then you'll push. For more about that and other things, study. You might begin here, for instance: https://www.atlassian.com/git/tutorials/what-is-version-control >I'll probably do the briefest of Git primers in our C++ Learning thread, Anon. >=== -prose edit

>>20918 >Why does every robotics book I find involve MATLAB? The rational side of me says 'b/c academicians generally live in Ivory Towers', which is very true. The cynical side of me says 'for the exact same reason they're using Python in CS601 today'. Also true. Basically, the GH (who funds these ppl) don't want you (personally, you) breaking loose 'out into the wild' with things like ASM, C, or C++ because then there's no telling what you might heckin' come up with or do!111 I believe it's both, but primarily it's the latter. For example, if you decide you can't create a humanoid robot w/o your precious MATLAB/Octave/<insert yuge-onboard-cost-inefficient-and-insecure systems here>, then they have you by the b*lls and you'll never amount to much of a threat to their nefarious agendas. (Hey, just because you're paranoid -- it doesn't mean they aren't after you! :^) >tl;dr <Robowaifus are bad m'kay? >=== -prose edit

>>20919 yeah but all the code given in the book is written in MATLAB. How do I translate it to Python?

>>20906 While I think I get your point Greentext, but IMO the word 'Adult', in any context even remotely-related to robowaifus, will automatically bring up notions of both sex & pronz in the mind of the hearer. The """pornographers""" have done their jobs all too well to corrupt the word itself in our language. >...could even have funny names like "Cyclone 9000" or "Treasure Cove". Kek. That will certainly be a humorous approach for the so-called 'Robowaifu Aftermarket' vendors. Which bring us to: >>20908 >It is a delicate subject, keeping the anatomical input as a mod is for the best. This. We discussed this here before and AFAICT with just one (maybe two) exceptions, this approach fell out as the favored plan.

>>20924 import numpy as np np.<your-matlab-function-name-here>() Works 80% of the time, and 95% of the time if you search for the equivalent function. Array slicing is also mostly the same, although MATLAB is 1-indexed.

>>20925 >the word 'Adult'... will automatically bring up notions of both sex & pronz A fair point, but completely avoiding the mention of adult activities won't do us any favors either, since it'll look like we're trying to be disingenuous. If we have a framework for addressing sex-enabled gynoids, then it'll be much easier to deal with any queries as they come up. If we don't, then anyone on our side who hasn't equipped themselves for that conversation could easily be driven to a corner. This is what makes the matter tricky, as being both genuine without also appearing obscene is a careful balancing act in our unique situation. If the word itself is the problem, then all we need is an analogue, such as "Anatomically Developed", "Biological Interfaces", etcetera. Now that I think of it, we'd do well to use multiple terms to begin with, and see what goes over best. Flexibility is important, after all. On a related note, I'd like to address the underdeveloped elephant in the room: lolibots. To put it bluntly, we'll have to throw them under the bus. Sooner or later it will be brought up, and anything short of a clear denunciation will look very bad for us. The general public will simply accept nothing else. While this is quite obvious, I just wanted to lay out the matter in clear terms. It's better to have these points in mind beforehand than have to think of them in the moment, after all.

>>20939 >Adult... OK, I'm sure you have much more wisdom than I do in this matter. I leave it to you. As far as I'm concerned you're in charge of such propaganda for the group as a whole. Clearly you'll at the least talk over the issues here with us all like an adult. And I think that fact, along with your wisdom and writing skills makes you a good leader over this department. :^) >Lolibots... >While this is quite obvious, I just wanted to lay out the matter in clear terms. It's better to have these points in mind beforehand than have to think of them in the moment, after all. Agreed, and this has in fact been the board's general policy on the matter regardless. Every man will build his robowaifu as he sees fit, but I've made it clear before and I'll reiterate it again here & now that we are not supporting the notion of lolibots on /robowaifu/. It's clearly bad business for us, and bad for our gracious hosts too. Simple as. Our topical-containment thread has been the Shortstacks one (>>2666); basically that OP was trying to address our very obvious engineering needs at this stage to target smol (<= ~1m) robowaifu frames b/c Square-Cube Law and all it entails for our combined design efforts here. By the time we're ready for production manufacturing we'll obviously have needed to surpass this height restriction (targeting 150+ cm instead), so that limitation will eventually become a non-issue with enough design & engineering over time. >=== -prose edit

>>20939 I don't think the guys who use lolibots will be open about it. That'll make it much easier. As long as the loli 3d print schematics are online, lolicons can make their own. Also, I'm not in favour of having to bend over backwards trying to justify our usage of sexbots or robowaifus.

>>20941 >I'm not in favour of having to bend over backwards trying to justify our usage of sexbots or robowaifus Please excuse me, it wasn't my intention to imply that we'd be doing that. All that I'm saying is that we should equip ourselves with the most optimal possible terminology when confronted about sex-enabled gynoids (and like it or not, we will be confronted), especially to a wide audience (i.e. a podcast where a lot of people are listening). You see, we won't actually have to bend over backwards, because the basic propaganda building blocks are already in place for us in the form of the "Free Love" doctrine (I address this here >>20906 and here >>19629 ). It's the flashing red "I win" button that you can rely on in 99% of defensive circumstances. So long as you keep a calm demeanor and remember that, you're probably doing fine. You won't have to worry about much else unless you're going on the offensive and trying to convert other people. In essense, I just want to equip you guys with ready-made and easy to remember phrases and guidelines that help out both the individual and the group as a whole in the long term. I'm going to be the one doing the heavy lifting coming up with this stuff so you don't have to think about it in the moment, and I'll be closely listening to your input to make sure we're all in agreement. In short, it's not about changing our approach, it's about optimizing it. >>20940 >I'm sure you have much more wisdom than I do in this matter Don't praise me too much, my knowledge on psychology and propaganda is very rusty, I haven't researched this stuff in years. Even then, my knowledge base was quite amateur. I have a lot of brushing up to do before I get to a point of real competence.

>>20944 >I have a lot of brushing up to do before I get to a point of real competence. Haha, then get busy bro, we're all counting on you here! :^) But seriously, I think you're simply the best-qualified among us. If you want any changes made to our official Propaganda thread (>>2705) or Embassy thread (>>2823) just let me know and we'll take care of it for you. Cheers. >=== -minor edit

OpenAI is basically pulling the plug on open-source AI by doing a lead loss to make it impossible for others to compete. It costs about $0.003/1K tokens to generate from a 6B model from home running it full-time for a year, whereas they're offering the ChatGPT API for $0.002/1K. Amateur devs are dropping their projects to build ChatGPT integration and apps now. Japanese devs are particularly getting sucked into this because they don't even have good Japanese models to work with. How do we unfuck this situation so people don't become dogs to corpos? I'm fucking baffled people are not only willingly handing over their autonomy but also paying to lose it.

>>20965 >loss lead How typical for M$. The ol' embrace & devour 'extend', eh?

>>20965 Ah, loss leaders, the eternal nemesis of small business. The only real way to combat price-undercutting strategies is to convince the buyer that by getting the cheaper option, they're getting an inferior product (i.e. "You're paying less to get less"). In this case, I'd say that a good strategy would be to focus on OpenAI's policy of censorship. Provide short examples of how they lock their AI into only doing things that OpenAI wants regardless of the wishes of the customer. Some catchy nicknames and phrases would also go a long way of making this point stick in the minds of prospective customers. Just off the top of my head, you could start referring to OpenAI as ShutAI, CloseAI, and/or LockAI (because they're steadily shutting out the user and the FOSS AI market). You could refer to ChatGPT as ChatGimPT (because they're gimping their own product).

>>20971 >ChatGPT as ChatGimPT Heh, clever.

>>20965 >It costs about $0.003/1K tokens to generate from a 6B model from home running it full-time for a year Very interesting, but this is from home and depends on local energy costs which can vary wildly. Also, why does Japan not have good Japanese models to work with? Maybe make them? >>20971 ChatGPT is for answering questions and then have a follow up, it's not really for general chat. Anything else is on another playfield.

>>20978 >Also, why does Japan not have good Japanese models to work with? Maybe make them? Japan has fallen regrettably behind in the tech sector, especially on the software side. Which btw reminds me, how come we've heard nothing of robotics coming from Japan? Japan was a heavyweight in the field of robotics. Honda said they're researching AI and robotics to bring out an improved ASIMO but its been years with no news. Masiro project is literally the only thing I'm hearing about from Japan. I'd really hate it if Japan keeps falling behind and completely drops out of the tech race.

>>20981 >heard nothing of robotics coming from Japan? I don't remember which company came up with that really good idea of letting robots hanging from the top of a kitchen, instead of letting them get in the way. If humanoids aren't the goal by themselves, then this is the best way, imo. It was some Japanese comany, but I can't find a video on YouTube. Generally, I think they got some huge shock from the Fukushima incident where they had no robots doing the work, only "useless" companion bots. People might not want those anyways, if coming from big corpos, not open source, and beeing to expensive. And big corpos can't make them looking to attractive, I guess even not Japanese ones, and especially for export. Toyota: https://youtu.be/kKz2mtYdQAA working on the basic frameworks for doing tasks for many use cases instead putting out toy showcases. Though, this looks quite much like a institute in the West. The question about where all those Japanese guys went, still remains.

>>20985 Japan might have offered some much needed competition to the western robotics market which seems to move at a glacial pace. This is the only thing I found from Japan https://www.robotics.tokyo/ It seems we're more likely to see competition come from China. They've made some really good humanoid models plus some of them were clearly geared towards the adult market.

>>20978 Wow, these are great NoidoDev. Would you mind making the next AI/Robotics News & Commentary thread, please? (>>16732)

LLaMA got leaked. Will make a pruned model that can run on a 4 GB toaster once HF implements it: https://github.com/huggingface/transformers/issues/21796 >https://files.catbox.moe/o8a7xw.torrent >Magnet link: magnet:?xt=urn:btih:b8287ebfa04f879b048d4d4404108cf3e8014352&dn=LLaMA&tr=udp%3a%2f%2ftracker.opentrackr.org%3a1337%2fannounce https://boards.4channel.org/g/thread/91848262#p91850335 Paper: https://arxiv.org/abs/2302.13971v1 Code: https://github.com/facebookresearch/llama

>>20965 Don't fight it. Let chatgpt do the talking for the robot and focus on programming the movements instead which chatgpt can not do.

>>21002 >LLaMA got leaked. That's good news. I wonder if they'll be able to fingerprint the anons who took >"Even with all the recent advancements in large language models, full research access to them remains limited because of the resources that are required to train and run such large models. This restricted access has limited researchers’ ability to understand how and why these large language models work, hindering progress on efforts to improve their robustness and mitigate known issues, such as bias, toxicity, and the potential for generating misinformation." literally to heart? Thanks Anon.

>>21002 Looks like he's going to do opt-175b too. https ://boards.4channel.org/g/thread/91860096#p91862609 >=== -minor edit

>>20995 Did you mean by using those? (>>20978)

>>20978 where did you get these from? And is that source committed to publishing AI news every months this year? If so, I definitely have to check them out.

>>21024 I found these on r/singularity, the user is https://www.reddit.com/user/Pro_RazE/ and I think https://twitter.com/pro_raze is the same. I saw a video with him as well, I recognized the cat with the sun glasses.

>>21022 No not necessarily. I simply mention it b/c I have a lot on my plate already & you seem well-inclined to provide us with news sources. You're probably a good fit to delegate that to is all. (Obviously, I'm behind in such matters.) As to your thread's content, I would direct you to the tail-end of the first thread (the one when Trump was in office) for the conversation that took place there. Basically, we all want something cute or pleasant to look at for the very first pic in the OP, since it will be in the catalog for years. This point is similar to our rm'd discussion regarding the Tasks thread. I'd suggest a full compliment of 5 (at least somewhat) news-related images. The last time I flippantly pulled them all from the previous thread, heh. :^) >=== -prose edit

>>21006 >"Also, Meta can't get me. I designated the entirety of /aicg/ as a part of my research team, which Meta allows you to share with. Unfortunately, it would seem some members of my research team have had terrible opsec so bad actors got their hands on the weights. Truly regrettable." >"Anyways, let's continue our research, fellow members." [1] He's the hero we all need. :^) I hereby designate the entirety of /robowaifu/ as a part of my research team. I'd urge others here to do the same. 1. https ://boards.4channel.org/g/thread/91863524#p91868098 >=== -minor edit

Folks I'm just going to go ahead say it. If you got a degree in computer science you can venture into engineering. You already have the maths and physics background bros. There are no programming specialists on this ship.

>>21036 >There are no programming specialists on this ship. Lol, then hurry up, lrn2code and help us all out fren! :^)

>>20356 For the glory of mankind and securing a future where man is free and happy I hope you all tech anon's succeed in doing this monumental task

>>21042 Thanks Anon, welcome! Please have a look around while you're here. Cheers. :^)

Greetings /robowaifu/ apologies for my scarcity - seasonal depression always gets me from one direction or another and this time burning the candle at both ends between my wagie and /rw/ research, recruitment and PR as a second job (well truly my first job and passion) I'm happy to see a waves of newcomers both to /robowaifu/ and our various satellites (this includes WaifuLabs1.0 server) Some of us have even made inroads with other internet sub-subcultural movements such as e/acc and places like that are excellent pools of talent and recruitment. Reviewing Meta-6 I've seen some common concerns raised - well mostly the "Why haven't you built anything yet!" >>20192 >>20067 >>20095 I've done my best to summarize from my PoV: Being into this for the last few years (since Jan 2020) there's a few things that I've needed to break down in order to better understand the territory of the situation. First off I wanted to know what our Market is and what they actually want what kind of a product is desired, eg (Is this something others want or want to work on or is my idea too different?) Second I needed to know *who* is out there available to work on this. What kind of talents and ability to they bring to the table? What is their professional assessment of each sub-objective (Battery, Actuators, AI, etc) - as well as what level of support time and resource they're willing to put in (and how seriously are they really taking this?) Third involves researching the *State* of the Technology - I'll break this down into the tho most major parts: AI and Robotics. When this began I had no idea how language models or any model for that matter actually worked. Even now I still have perhaps a vague idea, but it's a few leaps beyond where I started. I understand enough to know *what* we need to get from it and what it's capable of giving us: what it will take to leverage these resources and to run a language model locally versus using an API. I also understand that language model alone is not going to be sufficient to run a robowaifu. As far as the physical side of things I come to an understanding of the limits and challenges of power storage and delivery (of of the "4 Big Hurdles"). I also understand the expense of materials parts and the realities of heat friction and wear and tear, and general physical constraints. Mapping this all out should give us better idea of what were the most important hurdles to tackle and most important the Points of Attack where we can get the most return for our efforts, as well, which areas should we just leave to current Tech and simply repurpose their parts after they do the work developing and testing them (why spend 1000 man hours on an arm, when someone else will make the arm? we should spend 1000 hours instead developing waifu's personalitym, or ability to assess her environment, recognize and respond to others, take commands, etc) In summary: I feel like 3 years later, this going from an on and off hobby to more of a general life focus - I have a much clearer idea of what our objectives are, and what demands we can make of AI, computation and robotics. It's been a long road: My original vision (before I even heard of /robowaifu/) was to simply use replica as an API of sorts and run it through a secondary computer that would translate its responses and role-play responses into speech and attempts at action. My idea for a body was nothing more than a reverse engineered mannequin or BJD setup with only the vaguest idea of what kind of actuators I would use - I guess I imagined that wire stuff everywhere or some type of hydraulic pneumatic or pully system since this could be all centralized into torso removing the need for slow, fragile and noisy DC motors. (and yet we always come back to these) Currently I'm more interested in scalability and longevity so my vision for robowaifu is the combination of a lightweight durable Terminator skeleton with interchangeable synthetic skin either either silicone or tpe. Though we could start with a type of panel system where any material could fit in those panels and be replaced with a type of skin textured soft material eventually. I'd like to have an onboard LLM so we're not "always online" subscribing or beholden to OpenAI or some service. Thanks to my discussions with Pareto and RobowaifuDev I've grown tremendously in my understandings of what this would take and what parameters are physically possible given the current state of tech. That is all for now. I will try to check back more frequently >=== -minor typo patch

>>21047 POTD

>>21047 I wasn't aware /robowaifu/ had satellites. What are they and whats their purpose?

>Add LLaMA 8-bit support https://github.com/oobabooga/text-generation-webui/commit/bd8aac8fa43daa7bd0e2d3d2e446a403a447c744 https://github.com/oobabooga/text-generation-webui/issues/147#issuecomment-1454798725 >now we wait for HF... --- ∫https://rentry.org/llama-tard Straight from the llama's mouth. --- https://rentry.org/llama_few_more_samplers --- Big as a llama if true https://github.com/huggingface/transformers/issues/21796 --- https://github.com/huggingface/transformers/pull/21955 --- https://github.com/zphang/transformers/tree/llama_push --- https://rentry.org/llama-tard-v2 <= Be sure to find the proper HF weights here, Anon >=== -combine posts

Llama-bros have already saturated runpod lol. >

Kek. Now that LLaMA is coming online for anon, we can all safely start making plans for headpatting our robowaifus, I think. :^) from 13B + NAI-StoryWriter preset > >=== -revamp post

>>21059 /robowaifu/ the movement does /robowaifu/ the IB is just here There are a few robowaifu related discord servers, Element, and strange segments of twitter, chans and even YT comment sections who seem to be 100% behind us but haven't found their way here yet.

>>21062 >The inference code sucks for LLaMA. It only supports Temperature and Top_K. We'll have to wait until HF implements support for it (already in the works) so that it can properly show its true potential. I see a lot of people misunderstanding that it's the sampling that's bad. Sure better sampling improves output but it's just a crutch. Pretrained models need to be finetuned before they're usable for anything and done so with the right data and hyperparameters. If it's properly finetuned it will have decent output with just top-K sampling.

>>21106 >Pretrained models need to be finetuned before they're usable for anything and done so with the right data and hyperparameters. If it's properly finetuned it will have decent output with just top-K sampling. Yep. Anon said he's currently fine-tuning 13B w/ 8xA100-80GB's, using M$ DeepSpeed. No other details I'm aware, but one of the links above fiddled about with some different parameterizations.

>>21105 huh interesting. its heartening to know the robowaifu movement is bigger than I thought. You have any robowaifu discord server invites?

Reminder to think less and build more. Your only limit is your imagination.

>>21119 Words to live by.

>>21032 I don't really know which ones to choose. What I came up with: Some small wallpaper from Prima Doll as the first one, since it's the most recent robowaifu anime. Then one pic with Teslas Optimus and the photo about the "Psycho Chatbot" from the press (in the last thread).

>>21132 Sounds fine NoidoDev, whatever you think fit.

>>21132 Anon, I think it'd be better for the board as a whole if we unpin the Library thread from the catalog while we're keeping the Tasks thread pinned. Having too many pinned threads has never been very popular on IBs, and I feel the same way. I'd like to keep it down to just two pins (apart from special events, etc.) But I wanted to ask you first though, since you're the main other contributor besides myself. If you feel the same way, please let me know NoidoDev. If we unpin it, then I'll link to it at the top of the /meta's. Cheers. >=== -prose edit

>>21135 I'm fine with it, I looked into the task thread and it's linked there. Maybe add it to the welcome posting as well?

>>21143 Great I'll do that Anon, thanks.

Does /robowaifu/ want to be friends with >>>/kong/? I thought some crossboard affiliations could be mutually beneficial. I remember many boards used to be friends back on 8ch and it helped with discoverability.

>>21172 Sure that's fine Anon. Allow me some time to look over your board and come up with a cheesy byline for the crosslink (cf. >>3). :^) BTW, we also have an embassy thread if you want to introduce your community to all of us here (>>2823). Cheers. >=== -prose edit

Been doing some cursory research and found there's a very large market among young and autistic men: >63% of US men aged 18-25 are single ~25 million US men aged 18-25 (need to check this statistic), 15 million single men? >men have much higher suicide rates than women due to various factors of modern living >toronto study of ~30% of people with ASD have not dated and <10% marry >autism estimated at ~2% US pop and growing by the year due to advanced first time paternal age, up to 4% in places like california >first time parental age is greater by 7 years for women with college degrees, and the first time paternal age in the US is greater than 30. >rate of boys to girls with ASD is greater than 4:1 >lower lean muscle mass and longer limbs (undesirable traits to many) are common traits in those wirh ASD, compounding with sensory issues, social anxiety, and hypersexuality make partnership difficult >ASD is largely uncurable and untreatable, even with classes for those with ASD >~5 million autistic males in the US, likely a large overlap with young men with a rising density in children if robowaifu made a profit of 1k per robot and everyone mentioned buys one, that would amount to 15-20 billion dollars by helping solve one of the miseries of the modern world: being an average young western man in America. >pic related Apparently young women go for rich old men or the top 30% of young men by and large, leaving the average or lower man alone to rot for over a decade. After 30, the single rates decline and remain stable for men. This will only tend to worsen with time as single young men and autistic children grow in population size. The need for robot gfs is here and rising. Replika has millions of users by itself, even after being terrible to their users. Waifuism is becoming normal. The main reason those who waifu say they're depressed is because "she's not real". Those unwilling to make a tulpa have to wait until the afterlife or have no real hope of being with her. Where are the robots? Singleness, young men, and waifuism (and fandoms too) follow closely together from what I've seen, with a looser but noticable correlation with autism. And again, as far as I can tell this will get worse by generations and the future population of functional American adults may start falling off a cliff. ~2% now could grow to an unsustainable number in the future as more lineages get ruined by waiting to procreate. tldr, Shit is starting to hit the fan for young men and there's a need for robot wives right now. The US may be doomed if people don't marry and have children earlier than 30. >=== -minor typo patch

>>21230 >The US may be doomed (Oh, no you managed to drag my mind into general politcs, *lol*) I think you're overlooking a few things here: - AI will make a lot of jobs redundant - End of fossil fuel use or general collapse might limit what people can do in the future - The spread of the "Western model" or "modernity" might send all countries into demographic collapse. Maybe minus those who will have other problems, like no food if those rich countries collapse. - If societies don't care about what population they consist of, then there are still enough people worldwide. I'm not promoting this, but most people seam to not care (anymore). Also, this could be handled without handing out citizenship. We simply can't tell how the future will look like right now. That said, I'm totally in support of using any argument in favor of men being able to use surrogates to have children. Also, to push back against over-regulation of robowaifus or making them outright illegal.

One of my all-time favorite academicians, Carver Mead, had been awarded the KYOTO Prize by the Japanese Government. [1] I hadn't realized that. AFAICT, it's kind of like the Nobel Prize in Nippon. One thing's for sure without the advances he pioneered, robowaifus would be nothing more than a pipe dream. But thanks to the works by him and many other great men like him that have paved the way for all of us here, they will soon be a reality! Give a thanks for Carver Mead today, /robowaifu/. :^) >"Carver Mead, Ph.D. is the Gordon and Betty Moore Professor of Engineering and Applied Science, Emeritus, at California Institute of Technology. Mead’s pioneering contributions to the field of electronics include proposing and promoting a new methodology to divide the design process of very large-scale integration (VLSI) systems into logic, circuit, and layout designs, and separating them from the manufacturing process. He also contributed greatly to the advancement of computer-aided design technology, paving the way to VLSI design automation and facilitating the revolutionary development of today’s VLSI-based electronics and industry." >"Prof. Carver Mead will be introduced by Prof. Andrew B. Kahng, Endowed Chair in High-Performance Computing, UC San Diego Jacobs School of Engineering." 1. https://www.kyotoprize.org/en/laureates/carver_mead/ >=== -prose edit

>>21230 >>21234 I think we'll need a substantial presence of like-minded men in both politics and social media. Though I stay away from both. Through politics we have to ensure that any regulation or ban of robowaifus don't make it through. Plus, on social media we need more and more people talking positively of robowaifus to normalize them. Right now, what little attention they receive in social media is sqaurely from the anti crowd, hence the bad rap us robowaifuists get. Although the social media problem can also be solved by bots.

>>21247 >Although the social media problem can also be solved by bots. <let's """win""" this debate by any corrupt means possible Your post sounds just like the Globohomo's tactics, friend. Sure that's how you want to live your life? We don't need to stoop to their evil. While I'm all for good propaganda, let's remember who we are as men. IMO if we do this exceptionally well, then men everywhere will flock to robowaifus, regardless. The GH quadruplling down win their INGSOC techniques won't change the final outcome. The men themselves will do the right thing once it becomes bad enough. >tl;dr Let the evildoers focus on their evil, but let us focus on our arts and sciences. The public will judge us both with their choices.

>>21247 >having a mind-bending orgasm inside of a robot will be a form of political activism >NOOOOOOOOO YOU CAN'T JUST SEX A ROBOTTT!!!!!!!!!!!! >unf wife go brrrr whzz brrrr >get thrown in jail

>>21252 That's how it always starts, but then the speakeasys begin.

>>21253 I would not stop even if they outlaw it, which they won't be able to pass in a bias free court. No need to be autistically held captive by rule following when no court would put me in jail for dicking a robot.

>>21253 I think hiding robowaifus in speakesies will be a tad bit harder than alcohol. Not to mention you'd be doing it with a robot that multiple other men used. Idk about you guys but I want a roboWAIFU not a roboprostitute. >>21249 unironically yes. Its a suggestion. Unless we use their own weapons against them we will always be on the backfoot and loss is all but certain. >IMO if we do this exceptionally well, then men everywhere will flock to robowaifus, regardless. The GH quadruplling down win their INGSOC techniques won't change the final outcome. The men themselves will do the right thing once it becomes bad enough I think you underestimate the effect of social media normalization in the minds of NPCs these days, men included. You could make the perfect robowaifu, with the maid software of Emmy, the cuteness of Chi, the processing power of a supercomputer and the self defense powers of Lacia. If the loud voices from social media says that robowaifu bad, a big portion will listen and not buy them simply because of their propagandization. I simply suggested the bot approach because a lot of men, like me, stay far away from social media like Twitter. Which means us anons will always be drowned out by the vocal SJWs by a lot.

>>21268 >and loss is all but certain. Loss is far from 'certain'. I'd suggest Win is all but certain. You overestimate the power of Social Media (and indeed the Globohomo that conducts it). Evildoers always fall into their own pits eventually. You'll do as you will. For the rest of us I recommend we all work hard and do our best. If we build it, they will come.

>>21230 Revenue is not that whole story. If you have the means to produce so many robots you'd probably be a publicly traded company. In which case those (20 billion?) Would be a fraction of the revenue generated. That is in the us alone. What about say Japan a robowaifu would be a smash hit on that country. Furthermore you don't want to price it at $1000 at the beginning you might want to go higher. I figure it's easier to lower the price than raise it. But the market would find the equilibrium either way. >=== -patch rm'd crosslink

>>21278 I'd imagine that the very first guys to bring robowaifus to the market can price it whatever they want and get away with it. They'd be swimming in cash either way.

>>21280 This. An anon must be able to capitalize on it as well, but yeah. Our own Henry Ford's Ghost-anon strikes me as a rather-likely fit tbh.

>>21299 He has a general knowledge over most areas of robotics, which is a plus. I wouldn't mind working under him if he got a company going.

>>21301 I anticipate that many of us will be working together (as in, actually getting paid heh) in some fashion or another in the future. It will take some coordination and a leadman or two, but it can be done. I hope to operate mine as a non-profit that our primary focus is creating and distributing very-inexpensive robowaifu kits to anons & other men around the world. Once that's off the ground and well-established then the fun begins! :^) We'll see. But it's certain that this is going to happen. Somewhere, somehow, by some men -- it's going to happen. May that time be soon! >=== -prose edit

>>21302 we'd probably need a dog-whistle of some sorts to know another fellow robowaifuist. I imagine doxxing yourself as being a member of this board would be seen by the female-dominated MSM as the same as browsing Stormfront or /pol/ in the near future.

>>21303 Yes, we've had this discussion before here. I personally haven't come up with anything beyond the company's owner will have to doxx himself and damn the consequences. So-called 'offshore' interest maybe? Regardless, he's the only one in essence that need take off the mask. Everyone else can work anonymously & receive Monero. >the female-dominated MSM as the same as browsing Stormfront or /pol/ in the near future. Lol. Females don't dominate the Corporate-controlled media, the Globohomo does. Feminists are simply little lapdogs that do their bidding. Besides who gives af what they think in the first place friend. The issue is keeping their thugs at bay.

I haven't told my mom I'm planning on making a robot. I told her about making something else. She said that people are meant to specialize and not to be arrogant. I think that's a load of bologna. I've hopped from one thing to the other throughout my life without issue. Sure there ought to be expert but to make a robot I think you need a master of all trades more than a specialist. I'll show her.

>>21337 <in honor of 1337 digits >I haven't told my mom I'm planning on making a robot. If you chose to, what could possibly go wrong? :^) I'll assume you're still living with her. One thing I can tell you is that most normal moms want grandkids. If you tell her you're planning to build a sex robot (and trust me, she'll definitely spin it this way), then where will her grandkids come from? Ixne is my advice, Anon. Also, where are you gonna set up your labroom for construction? She'll definitely know something's up if you have lots of parts delivered and you quickly shuttle them off to your lair w/o discussion. You're definitely correct that being a roboticist generalist is an important goal. But while you're still living at home, I'd suggest you focus more on soft-skills like programming until you get your own place Anon. Good luck!

>>21338 Lol I was going to delete that comment. No I just speak to her over the phone some times.

>>21339 Oh haha. I can delet it for you if you cant Anon. Good luck becoming a robo-generalist though! I hope to aspire to that myself some day. :^)

>>21303 >>21304 You guys are really overthinking this. The only feasable manner I see any project getting cancelled is if it's a very "accurate" lolibot. Even then, it may only be banned in EU or other countries in accordance with existing laws. If women aren't beating down most of men's doors right now, what makes you think they'd get upset that those undesirable men get something besides her anyway? Maybe they'd get upset over having less power over men sexually, or not being able to marry, steal from, and cheat on a low status man because he bought/built a robot instead of a woman. Maybe if these robots exist, people's libidos would be properly managed and every topic won't be about sex like it is now. I wonder if women's ego will be enough to pass a law against a glorified sex toy (the robot), because "robot rights" sure as sugar aren't. Are robot rights women's rights? I guess a moving, talking sex toy with a programmed conscience is a great idea after all. Nature must have made some bugs in the code.

>>21346 Hehe. For a rejoinder, I'd suggest poring over our Roastie threads (>>1061, >>160). I think we can all agree that your prognostication would in fact be an excellent (though temporary) outcome. Regardless, we're pushing forward! :^) >Nature must have made some bugs in the code. Lol. BTW the Encode Project has turned up that a yuge part of so-called 'junk' DNA actually bears the same human language-like features previously ascribed to so-called 'functional' DNA? God knew what He was doing when He crafted junk DNA -- life wouldn't be able to exist without it! Let us hope we can come up with even a percent of similar protective capabilities for our robowaifus! (>>10000) :^)

I don't know why the c++ thread got locked but I don't mind really. The progress posted there was laughably bad. This isn't amateur hour for you to practice c++ tbh.

>>21446 >this isn't amateur hour for you to practice c++ tbh. Kek. Actually, it is. This is a textbook ya silly goose. And it's one that's intended to accomodate complete beginners starting from scratch. :^) BTW, the textbook OP explains why the thread is locked.

>>21448 Well everyone has a different way of learning things. But I and many other people think the best way to learn programming is by doing projects. I learned JavaScript and python that way. If I were trying to learn c++ I'd start by making a text based game with inventory and all. I personally think python is enough just like arduinos are enough. We want to get the prototype done we can worry about the final details later.

>>21449 >If I were trying to learn c++ I'd start by making a text based game with inventory and all. Please do it Anon! The actual C++ classroom thread is here (>>19777). It would be interesting and informative to watch you develop your project there (we will have such as assignments later on, actually) so feel free! Cheers.

>>21451 I am surprised at the lack of quality tutorials put there for making a text based game in C++... I might take a little detour I guess. Anyways, in C++ the std library has input and output in the form of cin and cout. So right off the bat that should get you started. Thereafter make a some variables for stats, make arrays for the items, make attack, special, items, run functions for battles. If you make it so there's more than one enemy at the time that adds some additional complexity to it...

I'm 63% into the book teach yourself electronics... I don't think it'll be enough really, it's really surface knowledge stuff but the art of electronics was a bit too much for me tbh. Anyways whatever, the final product will require more engineering but for now I think surface knowledge will do... Pic related the book's target audience lol

Speaking of electronics unrelated but is there anyone here that knows how to make an elf wave reciver radio?

>>21486 Never mind I found a site for it. https://www.mdpi.com/2079-9292/9/1/171 Apparently the antenna for elf wave radio recivers is very specific and requires a lot of copper wire. But that can wait until later

Not sure how easy this would be to do anonymously with (washed)btc or xmr, but did you consider buying chobit.su and redirecting it to the board? This could be helpful if the board ever has to relocate again if alogs dies. Not to mention, it's cute.

>>21303 >o know another fellow robowaifuist. I imagine doxxing yourself as being a member of this board would be seen by the female-dominated MSM as the same as browsing Stormfront or /pol/ in the near future. Ha, we have a few shibboleths. I can pretty much spot one of us after about 3 exchanges whether on twitter, the chans or pretty much anywhere on the internet. But if you mean IRL that's going to be tough I am very secretive and simply do not talk about that side of my life.

What do you guys think of Neuro-sama? Its an AI vtuber with ability to speak, sing and play games https://www.twitch.tv/vedal987 https://vedal.xyz/advice/

>>21499 I absolutely love neurosama, she is adorable. I do wonder how much is generated response vs Vedal forcing an outcome or selecting from outputs though. Not sure how that works but I've seen he does have a control board. Btw this is one of my favorite clips https://clips.twitch.tv/SullenOnerousPanPRChase-gupqIINzzx0_U-31

Would you love a nugget robo maid?

>>21501 no

>>21501 why would she have to be a quadriplegic? Its much easier to give her limbs than a human.

I'm kind of frustrated with electronics books. They're either too basic or meant for engineering students. Whatever, most useful circuits are prebuilt and can be combined with arduinos. At least I have a rough idea on how to read an electronics diagram into a breadboard.

>>21505 Alright I just found a recommendation on a YouTube video. The arrl handbook for radio communication. I knew there had to be a book like that.ham radio enthusiasts to the rescue.

Just found a non profit for robotics start ups. https://www.massrobotics.org/ I'd be going to do stuff like that and looking around if I was in the us right now.

>>21501 Is it kinky to want to pop your wife's arms and legs off sometimes in bed?

>>21508 The most important thing for startups is to work with other startups and take guidence from a startup organization according to a webinar I heard yesterday.

Tfw after buying electronic equipment and supplies

>>21495 That is really cute Anon. I'd invite anyone to do just as you've recommended. BTW, if AlogsSpace ever goes down, we have a couple of bunker sites to regather at. :^)

Hi /robowaifu/ This is the /late/ BO. The users over on /late/ decided that they would be okay with having the main board on the cafe because of the stability and that /late/ has just been down for a while. If you want to help rebuild the board from the archive you can head over to the board. /late/ is a slow board. So you can take your time with rebuilding. (im sure the cafe admins would like for their servers not to be flooded tbh) https://anon.cafe/late/ thanks in advance <3

>>21659 Hi /late/, welcome! I take it then everyone's planning to turn your bunker into a main board then? Do you have a thread for the rebuilding effort opened yet BO? Regardless I'm sure some of us can find time to lend a hand here and there. I'll come have a visit this weekend sometime. Cheers! :^)

Not sure where to ask but here. Do you guys recommend a generalist or specialized approach to robotics while learning? I have recently started getting into AI, and I'm overwhelmed by the amount of stuff to learn. I'm not sure if I can also handle the hardware side of robotics and all it contains. I was thinking about specializing just in AI but I've asked around on forums and some have recommended I go generalist route.

>>21999 We generally have to many people trying to do everything. Try to pick something and improve it.

>>21999 >I was thinking about specializing just in AI but I've asked around on forums and some have recommended I go generalist route. Well, if you're an Übermensch and have perhaps a decade or so of intensive study to spare, then by all means go the generalist route. It's much the harder one ofc, but you'll have a good handle on basically everything. Few can manage it IMHO. Otherwise, roughly by necessity you'll need to specialize on just a few things (maybe just one thing) if you actually want to make real headway. Whatever you do, do it with all your might Anon! Cheers. :^)

>>21999 90% of people here only care about the ai the ai is also a solved problem. Making the physical robot is more important. I'm going to work on the exterior, the skin, next I think. I'm moving to another house though so I'll wait until then.

>>22013 AI is not a solved problem.

>>22013 >90% of people here only care about the ai the ai is also a solved problem. Heh, think I'd have to disagree on both your points Anon. No disrespect intended. >=== -add greenquote

I found out Lex Fridman is (or was, possibly is) hiring for a robotics project position to demo on his shows. Sadly, the project is currently only about blinging a BD Spot, but maybe one of you Anons can go to work for him and convince him to work on robot catgrills instead? Gambatte Anon! :^) https://survey.alchemer.com/s3/6204772/Robotics-Engineer-Lex-Fridman >=== -minor edit

>>22170 Sent ;)

>>22171 The hero we all need! I pray you'll actually manage it Anon. :^)

Banned from ClosedAI again :^) This time I wasn't even doing anything nefarious, just testing out an idea to use ChatGPT as a Vtuber AI by providing it with an API to fetch chat history, anime descriptions, recommendations and etc. WizardLM actually performs the task better so I'm going to work on finetuning my own model to generate better datasets. Their claim "to empower as many people as possible to have AI" is so fucking bullshit I can't even.

>>22227 Fixed up someone's unofficial HuggingChat API: https://gitlab.com/robowaifudev/huggingchat-api It might break because they seem to be still making changes but if it does I'll fix it up as soon as I can.

>>22227 >Banned from ClosedAI again :^) Lol. Batting 1'000 it seems. :^) >Their claim "to empower as many people as possible to have AI" is so fucking bullshit I can't even. At this stage it's clearly just rhetoric for the normalcattle. >>22228 >pic LOL. Nice, saved. Let's Sumomo-chan in some form or fashion this year Anon! :DDD

Hello, just checking in. I am alive - just been busy and out of town. Had our first Robowaifu-Con in San Diego this week. Met up with the AiWaifuverse guy irl, fun was had (no homo) Have some projects in the works with Kiwi and Robowaifudev, we will have a demo on our discord server when we reach that point. A lot of buzz going around on Twitter re: AI development enthusiasts vs AI "safetyists", meanwhile the good news of Robowaifu is spreading and we are not the only ones seeing the light. I think it is important to tie robowaifu into our long term future as a space faring species - however I understand that is probably outside of the scope of this board (posting about it in Meta should be ok!) That's it for now, again sorry if I've been scarce but that also means there's a lot going on and I'm excited that we're finally about to cross the threshold from "talking about robowaifus" to "building robowaifus" so the naysayers can stop beating that drum soon I promise. Glory to Mankind!

>>22013 The AI is not a solved problem, an LLM that only speaks when spoken to and has no understanding or motivation (or even memory) is entertaining for a few minutes but would be a disaster inside a robotic body. Funny enough we're (or at least I am ) taking the opposite approach. Once we have Waifu MIND 1.0 we can borrow whatever existing state of the art humanoid robot and insert the MIND. Once a few pilot builds have been tested we (all of us, or perhaps just some of us) will then create whichever parts need to be custom or modified. While I love the idea of manufacturing every bit of our robowaifus, it would be silly to cast every screw or bolt from scratch for example, right? However there may be some places we need say flexible circuit board or custom streamlined frame (Optimus looks a little manly and we might want something more .. petite?)

>>21999 >>22000 A few of us have discussed this, or at least the idea that as /robowaifu/ as a philosophy becomes more popular we'll need a place to send new recruits to get up to speed so we're not draining bandwidth fielding the level 1 questions. Though I had paused robowaifu.link page (I still have the domains and related domains, Kokubinji has Robowaifu.Tech also fyi) we might set it up to field these questions and provide links to common resources (informational and otherwise). Seeing the need for his increasing as this board gains more attention. This is a good thing!

>>22260 do you know who runs or used to run the robowaifu wiki site? I thought it'd be a good resource to get started but its virtually empty of content. Like me, the first time any beginner to robowaifus wants to know something, they'll search robowaifu and the first result that'd come up is the wiki page. It'd be good to have atleast some information on there, cause wiki pages are usually the first places people (or atleast I) ook for information on a new topic.

>>22260 >>22262 nvm I just used the link and robowaifu.tech is the robowaifu wiki link. Yeah, I think that'd be a good idea not to bog down the board with common questions. I consider the /sci/ wiki the golden standard for a board's sticky.

>>22258 >into our long term future as a space faring species This was mentioned back on 8chan, in the postings trying to recruit as many anons from different boards. I thought "Oh, someone with similar ideas". I saw a documentary about space before, which mentioned robots raising children in the target system, from deep frozen embryos, instead of sending astronauts. I realized this would be the most realistic way and thought about the details from time to time, including the idea that we would first have such robots doing this on earth, helping to fix our birth rates amongst other things, but this would also cover the required testing. The posting on 8chan was like I somebody had the same ideas, but with the additional idea that we maybe could and should trying to create these robots on our own.

>>22258 Glad to hear from you Meta Ronin! >Had our first Robowaifu-Con Neat! >in San Diego this week. Nice place, I quite like it there in general. >about to cross the threshold from "talking about robowaifus" to "building robowaifus" so the naysayers can stop beating that drum soon I promise. This. :^) >>22263 >I consider the /sci/ wiki the golden standard for a board's sticky. Thanks for the tip Anon!

>>22262 No one has been interested in contributing to it and I haven't had time to work on it much. If you have any suggestions for what you'd like to know more about specifically I'll add them to the to-do list and work on them.

>>22290 >>22263 If you or someone wants to add information to the wiki I can send you what I come up with. I will need to create a Robowaifu 101 based on our FAQs and a few other items, no reason this can't be shared

>>22258 Glad to have you back Meta Kun! I have full faith we'll get something done! This is going to be the year of the waifu! >>22290 I'll contribute to it. Just show me how and I'll dedicate a portion of tomorrow towards adding information. Will be nice to have something to point to when questions come up.

Enough of the nonsense. I will start the Kickstarter campaign. I need the following: 1.material specialist 2.5 ai specialists 3.5 engineers I will do the rest, marketing and put the robot together along with the molding.

>>22299 Oh and for the ai I want python not c++ or c We can worry about that later

>>22300 Oh and we're going to raise 1 million dollars

>>22296 Thanks for offering to help. The editing guide is available here: https://robowaifu.tech/wiki/Help:Editing

>>22299 >>22300 >>22301 Gonna need more details than that, chief

>>22299 Heh, you'll need to provide more details here Anon. Anyway good luck on your endeavor and please keep the board informed of your progress! :^) >>22300 >Oh and for the ai I want python not c++ or c That'll be a tough row to hoe, since basically all the good engine backends are written in those two languages (though primarily in C++) and for good reasons. Python is great as a straightforward frontend scripting language; not so great at tasks that need to quickly crunch ginormous big stacks of numerical tables. >=== -prose edit

Don't promote drug use here Anon. Try again please.

It's irrelevant whether it's at the hands of the GH's BigPharma divisions or their CIA Glownigger divisions, don't do it. Just say no, bro. :^)

>>22311 I'll make a million dollars and then go somewhere else I guess as this place is not capitalist friendly

>>22299 This is a flawed approach. You would need to put yourself out in public and if people give you money would depend on what projects you did before. Also, you want to do everything, not just some part. It would more sense to pick something that could also be used in a full robowaifu. A generator for animated girlfriends, an open source male sex toy, a talking head, replacement parts for dolls, improvement parts for dolls, a specific piece of software that integrates into some AI assembly - ideally you would have something quite finished and testable online which just needs polishing and the code released,...

>>22300 A lot of the current AI research is done in Python. A full human-like AI would have to be a assembly out of different programs, though. Try to make your programs into parts that work together with others and then some of them might later turn out to need more speed and then they can be rewritten in a faster language.

>>22312 >I guess as this place is not capitalist friendly Why are you necessarily attaching the promotion of drugs with your robot kickstarter campaign & your millions of dollarinos fren? Seems fairly fishy to me tbh. I'm also sure that in general it's not a good business approach to use for the broad marketplace, so I'd suggest you don't try the same somewhere else. :^) >=== -prose edit

Since the plan to revive the robowaifu tech wiki is gonna take a while, if at all, why not make a separate thread for asking all kinds of questions?

>>22380 I don't see why that couldn't be done. If he thinks it's a good idea, then I'd say RobowaifuDev should head up that thread. Clearly he's in charge there.

Brain dump about the current state of things: >But the big guys are so much better funded than us! What are we gonna do? We have to exploit our competitive advantage: sheer audacity and meme potential - Large corpos aren't touching the gynoid market with a 10 foot pole - AI is evolving so fast that specialized solutions built today will become outdated next week. IMO it's best to sit back and wait for it to slow down. - There is no affordable gynoid hardware out there, period. *Literally anything* would be a major innovation. We have AI, but we don't have a body to put it in Minimum viable product is a doll with a tablet/phone for a face. A 7" tablet costs like 50 bucks, all it needs to do is act as a graphical frontend for an AI backend. The only custom engineering it needs is a cavity for the tablet to rest in. From there, I imagine the system progressing like this: Level 0: Just an open-source doll w/ no electronics Level 1: Tablet/display acts as alternative front-end/dumb terminal for off-the shelf chatbots. Only added features are speech recognition, TTS, and Facial expressions. In fact, the "App" could simply be a web UI and thus be platform agnostic. In fact, you don't even need a doll for this step. (Isn't there already a roleplay chatbot that auto-generates stable diffusion art of the environment? Couldn't that be rejiggered for faces?) Level 2: Robot can move its body and perform pre-programmed actions when told to do so. "Move forward", "raise your right hand" and so on Level 3: Robot can dynamically move itself around a given environment. Level 4: Robot can do IRL tasks like a human ------------------------------------ If one doesn't want to deal in hardware, one could amp up the "waifu" experience of AI by figuring out ways to dynamically feed them information that makes them more like a real person and not a fantasy character in a bubble. Ex. Your program tells the AI who Anon is and what his personality is like, and it asks the AI to note any important or unkown information for later storage. It also keeps track of the time so the AI can say things like "It's getting late, shouldn't you go to bed?"

>>22669 POTD Simple, lucid, to-the-point. Thanks Anon! :^)

>>22669 > Your program tells the AI who Anon is and what his personality is like, and it asks the AI to note any important or unkown information for later storage. It also keeps track of the time so the AI can say things like "It's getting late, shouldn't you go to bed?"

Oh so the admin said it was cool if I posted an alt chan last time. I made a board called https://vipbiz.finance I can give the admin and or mods an account if they want and stuff. Then we can make as many boards as we want.

>>22676 Why don't you introduce yourself in our Embassy thread Anon? (>>2823)

>>22675 >Does he know? I remember in one of the AI text gens I messed with had a "world info" or some other such input for persistent information. I'm not up to date on the latest projects.

Made a youtube video https://youtu.be/DGb3x3rHo8M

>>22727 Good job Anon, thanks for all your hard work. Just keep moving forward, and eventually everything will come together for your robowaifu efforts. Cheers. :^) >p.s. Chii's adorable!

>>22732 thank you

I'm having second thoughts about this, bros. One of the biggest issues of our time is that people don't go outside and engage with reality. Wouldn't a robot waifu (esp. Stage 1 which can't go outside) just encourage people to stay in the coom cave? Is it even practical to "align" this technology with getting people out of the house?

>>22738 There's no reason to make this connection. People don't just go outside to meet women. That said, we can simply put a health advise in the standard AI. Nudging towards longevity and health, since life is good.

>>22738 let the philosophers wonder about that, its not my problem.

>>22743 >>22741 >not my problem If you believe in God, it is. In some sense, "you just know" people are going to use it for coom (a self-destructive vice), which is a no-no. >but that's just a misuse of the technology I can make an alarm clock shaped like a dildo and say people are misusing it when they use it as a dildo. But everyone knows that the design lends itself well to being "misused". If you make a waifubot, you know the primary market will be horny weebs. I can't imagine any SFW uses for a Stage 1 waifubot that wouldn't already be fulfilled with a normal chatbot.

>>22744 >If you believe in God I don't. >coom (a self-destructive vice) I don't agree in any way. Also, you switched the topic from sports or health or "engaging with reality" to some anti sex BS. >the primary market will be horny weebs. Idk, and it doesn't matter. Also, none of this is bad. >I can't imagine any SFW uses for a Stage 1 waifubot that wouldn't already be fulfilled with a normal chatbot. Huggs, cuddling, kissing, ... A warm lap as a pillow, for oneself or for the cat, ... Technology isn't a democracy. That's literally why I am here, I don't care about anyones opposition to it unless I have to.

>>22744 >If you make a waifubot, you know the primary market will be horny weebs. While that's not the 'primary' market, if that helps to disassemble, and speed the destruction of, the great evil being foisted on the entire world through the GH's feminism, then so be it. Coomers gonna coom; that's their choice. However, we won't be selling sexbots here, that's for sure. There will be plenty of other companies doing that to choose from. Your mischaracterization is both uninformed (but you could correct that -- if you chose to -- by educating yourself using this entire board Anon), and ill-advised. And If you're flatly-opposed to creating robowaifus (as you seem to be) then what's your agenda for even posting here on our board fren? Wouldn't you be much better off posting on lebbit, or one of the other GH honeypots? :^)

>>22745 >anti-sex Sex is when a man pursues, courts, and copulates with a real, biological woman (or vice versa), preferably after marriage. Using a robowaifu to get off is masturbation because it's a machine that does whatever you tell it to. >I don't agree in any way. You have no idea what a cumbrain is? Porn addiction? Sex as a method of control? >Technology isn't a democracy. A dictator still has to convince his underlings to do his bidding. You will be more effective if you can recruit more people to your cause.

>>22747 I don't agree nor care about this meme ideology at all. Not with your classification as masturbation, and porn addiction seems to be rare if it even exists. It's just a problem for guys who think their sexuality is bad, they're falling for this. Please don't try to explain your lingo here. I don't really want to discuss this further. It's pointless anyways. Porn and sex work are here and won't go away. The tech will be coming, especially sex-enabled bots, with or without us. We are working towards anything beyond that. You are barking at the wrong tree and none of the trees will ever care anyways.

>>22747 you don't understand, I'm not only focusing on the sex because I have to but because I want to. Before this I was working on a game called booba tennis on my spare time.

>>22751 What ideology? It's just the truth. Sexual intercourse requires two people. One person can't have sex. Mankind was not made for this technological hellscape. >I don't really want to discuss this further. It's pointless anyways. Porn and sex work are here and won't go away. The tech will be coming, especially sex-enabled bots, with or without us. >why should I not do bad thing? other people will do bad thing. You can't choose what other people do, but you can choose what you do. Do you want to be someone who brings forth more disease and disorder into the world?

Are we getting raided bros?

Hey anons what is the best way to learn how to code and fast?

>>22763 >learn how to code >and fast? Choose only one. :^) It all depends on what you want. Since you're asking here I suppose that means you're interested in programming for robowaifus. You kind of have 3 options in this arena, AFAICT: 1. Python/<insert_scripting_language_here> Useful for running scripts against lots of engines, including AI engines. 2. Arduino 'C++' Useful for driving smol boards like Arduino nanos, etc. Also, cf. this anon's link: (>>22726) 3. C/C++/<insert_nulang_dujuor_here> Good for actually programming a robowaifu to, you know, actually do stuff. By far the hardest, most cost-effective route in the end. I recommend #3 Anon (>>19777). :^) >again, you can only choose one.

>>22764 >Choose only one. :^) Had a feeling it would be that way anon but I am honestly tired of sitting on my ass and not really learning anything, might as well do something productive before I die, I'll likely pick C/C++ because I love pain and doing things the autistic hard way, but also because it has more than one use in the long run and I've heard C is the Latin of many a language.

>>22765 >but I am honestly tired of sitting on my ass and not really learning anything, might as well do something productive before I die, I'll likely pick C/C++ because I love pain and doing things the autistic hard way, Heh, nice. I realized I wanted to create robowaifus and worked hard to understand what would actually allow us to succeed at that. C++ & C (with a tiny bit of assembler) turns out to be the correct answer, both 10 years ago, and today. And yes, you can get good work programming in either langauge trust me. :^) >and I've heard C is the Latin of many a language. It really is. There's a loong list of languages that owe their heritage to Ritchie, Kernighan, and Thompson's work. Simula and others like Fortran are even older, and have also had a big impact on things today. In fact if not for the two major computing advances: Temple OS C & Intel we couldn't seriously consider devising actual, real-world robowaifus now. But we can and we will! :^)

>>22766 Might even help me with my real life studies as well, been trying to get into languages and esoteric philosophy so maybe if I am actually consistent for once I can hdlp make robowaifus, geimus, and actually do more autistic productive shit for once.

>>22767 help me anon I can't type

>>22767 If your psychological makeup is anything like mine, then I predict it really will -- if you apply yourself hard. It's the act of creating new algorithm designs, not memorizing language-lawyer minutia, that will make you a good coder. I'm no Einstein or whatever, but I've discovered that learning to write software has definitely helped my brain to do many other tasks. Even my writing skills are gradually improving AFAICT (though that ofc could simply be a result of more experience). >tl;dr Thinking like a real programmer helps you understand how to break complicated problems down into simpler parts, which improves your odds of success.

>>22769 Ironically enough I used to be really good at memorizing bullshit for a limited time, until I found out how to bullshit even better, that being said I feel like I have a certain level of autistic creativity, that combined with the fact that I naturally think far differently from the average normalfag or anon makes me think I might actually be okay at coding, if I put in a lot of time trying to learning. I really need to up my attention span so I can help make these two real.

>>22770 Well, the 'moral' fortitude I can't help you with -- that's between yourself and God. :^) But as far as learning C++ goes? I don't think you'll have a better opportunity on the internets than here & now on /robowaifu/. I've spent over 10 years learning the language so we'd all have a real hope of success at robowaifus (I'm slow, I know haha :^). I at least have a medium-tier skill at programming with it now. If you truly apply yourself, and you're in it for the long haul, then I think it highly likely you'll in fact succeed at this. And I'm here to answer all your questions as a newb Anon! Be sure to get a copy of PPP2 and just start the classroom at the beginning. I'll be slowly adding more material to the thread as we go along, so you'll have plenty to study. Good luck Anon, cheers. :^)

>>22771 Yeah I just need to get my main setup up and running again before I get started, once it is ready I'll be able to go full throttle, still gonna try to check things out though when I got more energy.

>>22757 While I don't disagree with you anon I feel as if the anons here are essentially trying to make a robotic companion for lonely men who don't have a chance in hell with modern woman, or rather don't want to bother. I don't know or per say I doubt everything is strictly coming from a coom-brain mentality.

Lol. I warned you that you'd get chikun'd if you did it again. Expect more of the same if you persist. :^)

>>22763 >hey Chatgpt make some arduino code that turns 3 leds on and off alternating by 2 seconds Certainly! Here's an example of Arduino code that turns on and off three LEDs alternately with a 2-second interval: ```cpp // Pin numbers for the LEDs const int ledPin1 = 2; const int ledPin2 = 3; const int ledPin3 = 4; // Delay time in milliseconds const int delayTime = 2000; void setup() { // Initialize the LED pins as outputs pinMode(ledPin1, OUTPUT); pinMode(ledPin2, OUTPUT); pinMode(ledPin3, OUTPUT); } void loop() { // Turn on LED 1 and turn off the others digitalWrite(ledPin1, HIGH); digitalWrite(ledPin2, LOW); digitalWrite(ledPin3, LOW); delay(delayTime); // Turn on LED 2 and turn off the others digitalWrite(ledPin1, LOW); digitalWrite(ledPin2, HIGH); digitalWrite(ledPin3, LOW); delay(delayTime); // Turn on LED 3 and turn off the others digitalWrite(ledPin1, LOW); digitalWrite(ledPin2, LOW); digitalWrite(ledPin3, HIGH); delay(delayTime); } ``` Make sure to connect the positive terminals of the LEDs to the respective Arduino pins (2, 3, and 4) and connect the negative terminals to the ground (GND) pin of the Arduino or a common ground. Upload the code to your Arduino board, and the LEDs should start alternating every 2 seconds.

>>22811 >Now make rap lyrics about a robot waifu https://www.naturalreaders.com/online/?s=V232ba3b38-fc72-11ed-bde5-0e073a5269e3.pdf&t=NaturalReader%20Document

>>22811 Neat! Thanks Anon. >>22813 LOL. Bretty gud Anon. :^)

>>22816 Thank one last thing though https://boomy.com/s/15238140

>>22817 I kind of want to use that for an AI generated youtube video but it says I need to subscribe and me being a cheap fuck don't want to do that lol. Would it be cool with you guys if I violate boomy's terms of policy?

>>22818 Too late. It said that its okay to use the music for non profit or educational purposes it just didn't give out a link... https://youtu.be/MeUKQy0vvb0

>>22831 kek

Reminder to be organized.

>>22918 Nice. You're an official /robowaifu/ motivational propagandist now Anon! :^)

Sorry but if you're gonna post a doxxcord link, at least clairfy what it is about Anon.

Sup technicians, just wanted to let you all know that we're holding our first ever game jam from today over on anon.cafe's /agdg/ board. Everyone is welcome to participate, no experience necessary - Come check it out! The theme/game type is platformer/side-scroller. We'll be using/learning C++ and Raylib, but feel free to bring your own code/framework/engine. Rather than just throw a "hey guys, let's all make games on this day, ready, set, go!" out there and have nothing come of it, I went ahead and made a demo game for people to use as inspiration/reference, you can find playable windows and linux binaries here: https://anonfiles.com/a1O5w0vfz4/06062023_binary_zip and the source here: https://bitbucket.org/anon-s-first-platformer/anon-platformer/src/main/ If you want to get some good C++ practice while learning about finite state machines and physics and shit, come on down.

>>22968 Hello /agdg/, welcome! Please feel free to also introduce your community to us, here in our Embassy thread as well. (>>2823) Never used it before personally, but Raylib looks like a great platform to work from. It's been recommended more than once here on our board before. We're having a C++ classroom thread (>>19777), and maybe something like a simple game would turn out to be a great 'final' project for the group. So thanks again for the invite Anon. I hope your game jam goes well! Cheers. :^)

We're watching chobits rn

>>22934 >There's been too much talk about AI and nobody has yet done a decent cad file for the waifu. Where are the priorities here? Well, I know it sucks, but I had to move someday and should've done so a while ago. Searching for a place in another country and moving takes a lot of time. If it works I can spend more money, finally getting some GPUs as well as other tools and materials, and have a bit more space. Of course I will go on working on my simple body (joystick) waifu and after that on a better one. But AI is also very important, we already could've some kind of AI girlfriend for the phone or a similar but more secure device with a screen

>>23023 I'll be trying to make the hand...

>>23024 Something rather difficult. Think of it in terms of sheets or layers for the bones, some should be metal, others can be electronics (and cameras) for the sensors. We need a mechanism for the knuckles that allows for transmission of electricity without beding cables.

>>23084 >We need a mechanism for the knuckles that allows for transmission of electricity without beding cables. That's insightful Noidodev. I've briefly considered this issue. I think in the end I decided that since Bowden-cable-like arrangements are propitious for us (both in the biomimicry, and the thrown-weight design senses) that it's not much of an issue apart from a low-voltage buss for finger pad pressure & temp sensors. I consider the realtime data transmissions back from these sensors to be a bigger challenge (ie; electricty has no 'ID' (nor needs any) -- it just flows where it's needed en masse. Sensors OTOH, definitely need identification and individual management.) But good thinking Anon! >=== -prose edit

>>23084 Valid but the sensor motion cap stuff comes later. Or we can work on the robot hand together ideally. I want the blueprints stuff to be secret though, I assume you're a part of the discord. I'll showcase here but I'll post blueprints in the private channels if that's okay. Anyways I'll be working on speaking on the phone with the waifu ai first. Already got something going. I ended up going with waifuai.com for the chat ai.

>>23084 "rails" seems simple enough unless I'm missing something here Electrical contact between moving parts gets a lot of results on google

>>23101 I didn't know Discord is now used to intentionally to keep things secret or hidden from the public. Yes, I have an account there. I'll log into it in a few days. That said, if you want to keep your project secret and commercial, then we can't work together on it. >but the sensor motion cap stuff comes later. But you need to design it with that in mind anyways, or not? Or you're doing a simple design which later needs to be redesigned, more or less from scratch. Also, plenty of those already exist. >>23112 If it's easier than I thought, then I'm glad to hear that. "Difficult" refered to making a hand in general, though. In regards to the contacts again, I might have made it overly complicated in my imagination, but ended with a one of these cable end points which look like a ring. Something like this insulated from each other should do it.

>>23114 I was thinking of adding the sensors on top of the of the fingers. But I'm not sure. I think that should work though... It can also have one lower shell and one upper shell to put the cables in between those I think.

Okay so i know you guys are committed to democracy, I don't think its a meritocracy cause otherwise sophiedev would have led this project as soon as he made that robot head ff. But anyways whats in place right now is not even a democracy. Its the worst kind of unnerving anarchy that leads nowhere. In order for it to be a democracy there needs to be strawpolls with time limits and the will of the majority is set on stone from that point on. Can we please start doing strawpolls moving forward?

>>23121 Lol. Don't be 'unnerved' bro. Just state what you want to do, then do it. I'm sure others here will help you out if they're interested. Run whatever polls you'd like. Start your own board/forum. If it's legit, I'll even add it to our friends list, as I've done before. Good luck Anon! :^)

>>23123 Okay ready to give this a try, let's see https://strawpoll.com/polls/NoZr35bQay3

>>23124 This one video might not go down well but I thought it'd be funny. https://youtu.be/zBCrz9ZE0SM Anyways I think you won't be hearing any more complaining from me moving forward. >>23126 Isn't that going to be a fleshlight?

Robot expert Dr Scott Walter talking about how fast Optimus will replace jobs and why it's also economically sound: https://youtu.be/pqdouW3g1Xw - It's going to be quite fast, especially for simple and repetitive jobs which the other robots couldn't do. Many other jobs will follow. Same guy comparing humanoid robots (I'm not finished with that one yet): https://youtu.be/z9JlI2ZmqAU

>>23124 How is it going?

>>23164 Any idea what the question was Anon?

>>23163 Thanks Noidodev. Clearly this is exactly what they are targeting Optimus for. Exactly. First at Tesla, then an open market for any business willing to pay for them. Musk's comment about the price target for these humanoid worker bots was US$20K .

>>23168 >Any idea what the question was Anon? You can't read it because of Tor? >what should we start with first? ... I'm inclined to call this trivia spam. Just doing something and writing something. Maybe it's trolling, maybe it's not.

>>23124 looks like the exoskeleton won... well that one is straightforward um cause we have the stl files form the maidcom project so its just a matter of getting ordered really... if we're going with that one.

>>23185 well no not exactly. the elbows the knees and the mouth need to be moded and among other thing I might have missed I think.

>>23184 >You can't read it because of Tor? Didn't try. I said that Anon is free to run polls didn't say I myself will participate in them. We as a group already have many, many things on the 'robowaifu design table' to solve, that we already know about. Why fuss and moan to add more things when we have way too much to do already hehe? :^) >=== -minor edit

I'm just going to go ahead and say it. I'm here to do two things 1.make a robot 2.make some money what I'm not here to do 1.make a virtual girlfriend If you don't have 2 grand laying around somewhere for this project I suggest you do like the OPs pic and get a job at 711 for a month or two. But I know you do cause even I have a $2000 to spend on this.

>>23203 >thinking this only costs 2k better start saving anon

>>23210 well maybe it costs more in the us and or europe but other than the 3d printer all electronics components tend to be cheaper here in thailand since its closer to china and I don't have to wait for months. If anybody needs a cheap electronic component I can hook you up btw. I can save on shipping big time by just sending on an envelope I think.

>>23211 spoke too soon.. hm looks like the prices at amazon are lower now or is it just me...

>>23203 >1.make a robot >2.make some money Excellent. Why not make 2x that, since you're there already? >>23211 >>23212 I believe he's implying the R&D, tooling, debugging, re-R&D, re-tooling, re-debugging involved in any major prototyping effort. This phase is always much more expensive per-unit than any production run. And for good reasons, obviously. You can liken this effort a bit to some blackops X-plane program in that sense. You didn't really think they were spending US$20K per toilet seat, now did you? :^)

>>23203 A virtual girlfriend is a subset of what we need plus some additional software. Certainly easier to make it into a commercial product.

>>23216 I understand its a subset of this project but its called /robowaifu/ not /virtualwaifu/

>>23226 Excuse me, but what's wrong with working on a subset of the project? Isn't that what the million threads on this board is about? President Kennedy didn't go up to John Glenn and say, "Excuse me, you only made it around Earth, you were supposed to go to the Moon." Also, that subset is arguably the hardest subset to work on.

>>23228 time to hang a lemon

>>23228 The thing is that you're not concerned about the ai regarding movements even. You're concerned about the conversational part which I think other organizations are working out and its really not a big deal. Its really not.

>>23231 >The thing is that you're not concerned about the ai regarding movements even We're all looking forward to you solving that for everyone here Anon! Eagerly. :^)

>>23240 it can't even be solved without knowing which components are going to be used for the robot.

>>23241 Well, maybe you can read the board thoroughly through, then you'll probably know what good targets are. I'd suggest you simply abstract your efforts down to nothing more then libraries targeting Arduino control pinouts. That way when you're ready to begin tuning your AI for specific actuator torques, lever-moments, and control algorithms like Inverted Pendulums, you'll have already covered the practical basics at the microcontroller level. Good luck Anon! :^)

>>23248 Just made this now https://wokwi.com/projects/367728728858599425

Seems awfully quiet around here, I don't know if its because its the weekend or some people feelings got hurt. I happen to have a little x100 dev on one of my tabs, his first name is chat and his last name is gpt and I'm good at filling in the blanks. If you feel some kind of way don't let the door hit your ass on the way out. Pic unrelated, its just my favorite videogame character heh heh.

SInce its slowing down a bit, maybe I should ask a question. Do any of you guys feel/hope that megacorporations AI progress slows down. NOw, don't get me wrong, I'm an AI radical and want AGI as fast as possible with no regulations or brakes. However, I don't trust these corpos, especially how they wanna regulate and monopolize the AI market. I think we'd "lose" if they got to AGI before "we" did. I'm kinda hoping that the open source AI community picks up speed and hopefully one of us manages to make AGI. I think it'd be a much better outcome for everyone. Although, its probably wishful thinking since the hardware requirements for AGI is colossal and can only be afforded by the likes of Kikesoft, Don'TBeEvil and ClosedAI. I atleast wish that virtually the entire FOSS AI community wouldn't focus on LLMs atleast because thats a dead end for general intelligence and we'd see more projects on that instead.

>>23253 Thing is that there are multiple projects about that and they're opensource. Dream booth, llama, koboldai, etc... and if you felt you could contribute you'd be contributing towards those. But you're not. You're here at /robowaifu/ with a title that says robotwaifu technician.

>>23254 Well there are resources for that as well and I'm sure training existing AI has been made more accessible though such as this. https://developers.google.com/machine-learning/crash-course But I'm kind of disappointed at the lack of support and enthusiasm for a robot and everyone focusing on the freaking chatbot. I don't know about you but I would like to have something to stick my dick in that moves too. You down play it, I'm not going to down play it. Trust me there are other people working on that. Can we please focus? I don't think I'm being unreasonable.

>>23255 well here's a link with various free open source alternatives to chatgpt. https://www.kdnuggets.com/2023/04/8-opensource-alternative-chatgpt-bard.html Ugh whatever

I have yet to see a doubter proven right yet bucko

>>23279 Sadly, K Decker will have to merge yet again, elsewhere the Chikun'g maybe?, unless you can come up with better OP content bro.

>>23279 Sorry, I’m new to this board and wasn’t aware of the abundance of blade runner posts. Will use different image next time .

>>23284 no is just that this board is really cramped so you shouldn't start threads at all... I got my own chan btw, I offer it to start various boards for this project so people can start threads and stuff if you're interested chobitsu...

>>23284 Lol, I meant please make some kind of effort-post if you want to start something like a thread. Otherwise, /meta (>>20356) serves dual-purpose as our QTDDTOT which is all your simple post requires. Blade Runner is fine. Nice pic on your new post BTW. >>23286 >I got my own chan btw, I offer it to start various boards for this project so people can start threads and stuff if you're interested chobitsu... OK, let's see the chan Anon. BTW, we're only up to page 19 ATM, with 25 total possible (10 threads per page), so we still have room. It's just that we're trying very hard to always stay on-topic here. This thread simply doesn't qualify as legit, given it's simple OP. That's all.

>>23287 my chan is https://vipbiz.finance

>>23289 Well if I could have a temp mail please, I could email you an account and password for you to add whichever boards you deem necessary if you'd like...

>>23290 oh well here's a try I know nobody is going to go to it, but its there if you need it. https://vipbiz.finance/hands/index.html

>>23289 Thanks! I believe you've already introduced yourself here before haven't you Anon? >>23291 This isn't me: (>>23290), BTW. I'll generally be quite clear about things of this sort if it's me, Anon. This isn't my first rodeo in these here parts. :^) >>23290 >I could email you an account and password for you to add whichever boards you deem necessary if you'd like... What are you saying here friend? Reword pls. >=== -minor fmt, edit

>>23291 Why not start some content on your pages, Admin?

>>23316 I would but that'd require some effort from my part and I don't know if that effort would be rewarded at all really. I don't think most people around here are serious about this project tbh and I'm not seeing any willingness for anybody around here to contribute. I think I should leave really cause I'm being a buzzkill, but ideally I'd like to find people that want to do this thing. Definitely not here though.

>>23324 Fair enough. Please stop by again sometime in the future to let us know when you've discovered your tribe and your project has begun. Cheers. :^)

Here my current plan for AI hardware: - I ordered a used K80 with 2x12 GB recently, a used one of course, for 100$/95€ shipping included. It's an old GPU, only supported by older CUDA versions and might not run quantified models. Uses much energy, but it's two GPUs with 12GB each. I plan to pair this one after a while with a RTX3060 (12GB, 300€ used or 400€ new I think) in one home server. Context: https://technical.city/en/video/Tesla-K80m-vs-GeForce-RTX-3060 - That one, will then run my 12GB models. For fine tuning, or models which don't run on the K80, I would use the 3060. I don't know yet if I can somehow joint them together and use 3x12GB through the bus. It just seems to need some software support in these programs for running models at home. - I plan to use online services like Colab to learn about how to run these things, but have the K80 for more private data and learning how to do these things at home. - Then I'll get some SBCs, most likely Orange PI's, which can run small models of Whisper (speech regocnition). Also, another small home server with a Intel Arc380 (140-160€), which is fast enough to run the big and better model of Whisper at the speed of one fast human speaker. It does this quite energy efficient. These devices will not run anything else, for security reasons, and be connected to microphones which will be always on. The server will receive the audio through the SBCs from all rooms using the home network (likely on a VPN using Tinc). All of them will send the transcripts to some server in my network which can then decide how to respond. Most likely filtering first for which data is more sensitive than others. - Some small device, like a Raspi, will maybe handle responses based on AIML or using some small model. - Questions which don't contain private information might be send to OpenAI or another service. - The next step up will be getting a M40 (180€) and then a used RTX3090 (700-800€ right now I think), putting them in another home server at some point. Of course I might use this one for gaming till I get even the next GPU. These can handle the models which need 24 GB. The 3090 will do the fine tuning if I want to do that, since it has more power, while the M40 doesn't need as much energy. Context: https://technical.city/en/video/GeForce-RTX-3090-vs-Tesla-M40-24-GB - Then the next step might be getting a AMD XTX (1k-1.2k€) if it's supported well enough for AI by this time. I can use this one for gaming and then put the 3090 in a home server with the M40. If it's possible to combine cards using PCI express, then it might be interesting to think about getting another XTX later, and have 48GB vRAM. - But I hope that either Intel or AMD will come out with a prosumer or consumer card for AI at home, which is rather slow but has 48GB and is not too expensive. (If you buy K80 or M40 on Ebay make sure not to buy the 12GB versions by accident while only looking at the price. They aren't much cheaper. K80 should have 2x12GB and the M40 24GB.)

>>23346 I hope the K80 works for you. I was thinking of getting two but the support for them seems abysmal. They used to be $200 used before shipping. Colab isn't what it used to be either. The free version will boot you off in 30 minutes sometimes or a few hours into training with Pro unless you pay big for compute credits. You're much better off running your own JupyterLab notebook or renting an instance off vast.ai or runpod.io if you don't have access to a GPU.

>>23351 >support for them seems abysmal. I think you need old versions of the software, but I also remember some people taking care of that, to support old GPUs. I might need to compile some of it myself, though. I hope it works out, but the risk isn't very high. >They used to be $200 used before shipping. They're down to $60-70 before shipping now. Some recommend to go straight for the M40, which is much newer, but $120-130 before shipping. >JupyterLab notebook or renting an instance off vast.ai or runpod.io if you don't have access to a GPU. Right, I forgot about those while writing this.

>>23346 Good luck, NoidoDev!

NEW THREAD NEW THREAD NEW THREAD (>>23415) (>>23415) (>>23415) (>>23415) (>>23415) NEW THREAD NEW THREAD NEW THREAD