>>24943 NoidoDev & Kiwi are right, Grommet, and I'm going to have to agree that your post is misplaced. I.E., (ML/DL/LLM/ ???NN-X???) are simply component(s) of cognition, which is much, much 'higher up the foodchain' if you will. These NN techs are important building-blocks for the endgoal here, but alone they are incomplete in the context of what we're all going for ITT. Your post also seems rather blackpill'y as well IMO, Anon. And that's not like you -- are you having a bad day or what? :D Together, we're all gonna make it! >=== -prose edit

>>24955 Looking like a very good start Anon. GG. >>24948 >>24958 >Hoshino Chan I'll confess my cowardice to the board, heh. Plastic Memories hit me so hard that it's one of only 2 animus I don't want to watch again (Grave of the Fireflies being the other). She strikes me as somewhat similar-looking, and I heard some things about the ending that made me uneasy, that of all the robowaifu-centric stuff in our media threads, this is the single one I've intentionally avoided. :^) But, I'll probably dig it all up and go through it now b/c this thread.

>>24910 >Yes, I didn't read the whole posting >>24816 at first, because it started responding to the "shell" distraction. I'm going to look into Justified Programming. I will try to better order parts of my posts, any suggestions on how to make more helpful and appealing? If a specific reply for something you don't care about, just skip it :^) >>24904 >I'm not sure if optimizations for conversations, like faster understanding and anticipation where a conversation goes, should also go in here or in another thread I think it fits this thread, that’s absolutely a higher level function. >>24909 >separating it from conversational AI is probably impossible, since we need to use it for inner dialog and retrieval of data from LLMs I’d argue that is not the case, and that it’s actually essential to not overly depend on LLMs especially for knowledge storage. Take what I’m about to say with a grain of salt, I am simply stating what my gut feeling is. I will argue that an LLM should be treated as purely part of a language center and that storing facts (like what is the capital of X or Who is the Yth president?) is silly, a red herring and a waste of resources. Storing knowledge this way makes it hard to update, makes LLMs expensive to train. Adding new knowledge via fine-tuning is not effective, but fine-tuning is good for changing the style of text (1). Here is something I found inspiring when thinking about how to integrate LLMs into a larger system (2). This also ties in to "Hallucination", I find the way it is talked about and even the coined term annoying and unhelpful. You should not be "solving" it, its not a problem, if anything it demonstrates that a model is good at predicting the next token and is not over fitted on the data. The problem with "hallucinations" is that pattern prediction and long term memory is being conflated as one. Maybe I am not understanding something, but I am not impressed by the larger LLMs, I see them as a step backwards. Maybe I am wrong and an anon can tell me why? >>24947 >I try to avoid these people, I don't engage in this kind of discussions, I don't care, I will do it anyways, if I can. Thank you, your a breath of fresh air. I am so tired of this same muh "Roko's basilisk", Sam Altman regulate me harder daddy, reddit """debate""" bullcrap, it's the same style of bullshit as "humanity is bad", its forced garbage pushed by media to keep you (us) down, please reconsider what you consume friends. >>24943 Higher level thinking is important for AI safety, I am sure even the nicest of people sometimes have an angry or bad thought, but they don't spiral into a rage on a drop of a hat simply due to that thought popping into there head. Thought or Prediction is not equal to action, there clearly is a gating mechanism for actions. If your AI is just an LLM feeding into it self, id be more worried, An idea I have noted down: emotions are a gating mechanism for actions and a mechanism for determining the kinds of goals and agent prioritizes. >>24904 >- The system will need to be as flexible as possible, making it possible to add new features everywhere I have been thinking about how a practical architecture would look like, the current idea I’m on is going with is a central event bus and a bunch of modules/plugins observing and adding events onto the bus. Each module when it emits an event provides a confidence score. If there is several modules competing for output, the highest score wins. Modules will also be provided a "feedback" score, this can be a manual input from anon or module(s) tracking various inputs to predict anon engagement. The bus would be the ideal place to have logging and to visualize the system's state. Also I would like to advice anons to not be soydevs, do not build garbage that I will rewrite out of disgust :/ do not be inspired by web garbage like Firebase, GraphQL or whatever else, the cloud sucks and self hosting web crap is such a pain, the web ecosystem, especially nodejs/npm and python is an example of what not to do. If you even think about using docker, or having one or two official linux distros for building, that is a strong indicator that your code is hard to maintain garbage. For example why would you use GraphQL when you could just directly query the database and have less moving parts. Your building a system to be used and run by the same party, Your waifu should be running on your own machines anyway. Some technology I want to bring attention to: 1. SQLite, people really under estimate what SQLite can do, it's a very powerful and extendable database that will run anywhere, do a deep dive. You can do document databases(3), graph stuff(4) and even embedding vector search (5). 2. D programming language, its my choice of systems programming language, If picking it is too scary then go with C++, just make sure to expose a nice pure C api! 3. GGML (6), It's the ML library used in llama.cpp & whisper.cpp. It's so magical to type in make into my terminal and getting an executable that just works. >>24962 >I'll confess my cowardice to the board, heh. Plastic Memories hit me so hard that it's one of only 2 animus I don't want to watch again Plastic Memories, yah that one hits hard, its brutal. 1: https://kaiokendev.github.io/til 2: https://medium.com/@peter.lawrence_47665/knowledge-graphs-large-language-models-the-ability-for-users-to-ask-their-own-questions-e4afc348fa72 3: https://dgl.cx/2020/06/sqlite-json-support 4: https://github.com/dpapathanasiou/simple-graph/tree/main/sql 5: https://github.com/asg017/sqlite-vss 6: https://github.com/ggerganov/ggml

>>24957 >"An algorithm without a mind cannot have a heart." We haz winrar! You've coined a phrase right there Kiwi. This is deffo going into the new OP! :DD >A machine with a mind could be made to understand this even better than we do. She can have super human love. You've hit the nail on the head, Anon. In some sense, that's kind of what I'm hoping to achieve, going back to my OG stated-goal of a Christ-chan robowaifu personality (>>6815). I truly think your statement is a real possibility, and it's a very high-priority goal for me with /robowaifu/ (has been all along).

>>24962 OT: I like emotional and very dark stories, and I love this one. Yeah, maybe better don't watch it. Same for "Prima Doll", which is about gynoid soldier veterans with PTSD. They switch mostly between crying and singing, while being very cute. Broootal. That said, aren't you the one who said Planetarian has some Christian symbolism? The song and something about the stars. I assumed you watched it.

>>24963 >>separating it from conversational AI is probably impossible, since we need to use it for inner dialog and retrieval of data from LLMs >I’d argue that is not the case, and that it’s actually essential to not overly depend on LLMs especially for knowledge storage. That's exactly my line of argumentation. You somehow misunderstood me. With the comment above I meant it's going to be difficult to draw the line between discussing conversational AI and a cognitive architecture, which ideally should be in different threads, like we have a thread for speech recognition and vision as well. But I think it's hard to draw the line. >but I am not impressed by the larger LLMs, I see them as a step backwards. Well, they're obviously very good at certain things, just with limitations. You can't use one LLM for a personal AI, since a lot of things are just missing and it follows the mainstream thinking or whatever data it has been trained on. >do not be inspired by web garbage like Firebase, GraphQL or whatever else, Well, I described above why it is a useful concept. I want to plugin all kinds of methods, tools, programs, whatever. >GraphQL >GraphQL is a query language for APIs and a runtime for fulfilling those queries with your existing data. GraphQL provides a complete and understandable description of the data in your API, gives clients the power to ask for exactly what they need and nothing more, makes it easier to evolve APIs over time, and enables powerful developer tools. It's not a cloud or service. >SQLite Yes, I agree. But I also want graphs.

>>24963 >Plastic Memories, yah that one hits hard, its brutal. ;~; I guess I'm still a child at heart haha. :D >>24965 >Yeah, maybe better don't watch it. Heh, thanks for the warning. I feel I better watch it just to catch up on the board lore r/n, since Kiwi, yourself, and other anons reference. >That said, aren't you the one who said Planetarian has some Christian symbolism? The song and something about the stars. I assumed you watched it. Yeah, I remember that post in the OG thread (>>16269). Nope, wasn't me.

>>24966 I know what GraphQL is, at one point I was employed to rip it out of a code base, because of how much of a performance disaster it is. What I am saying is not a secret and I am trying to warn people. feel free to google around, it's well known that GraphQL is slow. I understand that some of it can be mitigated, but even then, I really do not think the amount of value it provides is worth the cost, making it a technology I would avoid using. I am also not claiming it is a cloud service, I am saying its a idea born to solve cloud style problems. One of the major use-cases of GraphQL is how to safely expose a database with some simple data processing on top to a 3rd party, its solves the problem that your can't just give direct access to your database to an outside user and that a classical restful api constricts the "shape" and types of query the 3rd party can do. I hate to just tell someone to not use something without providing an alternative, so please take a look at Virtual Table in SQLite, its a great way to expose data not inside your DB in uniform queryable way. The reason I say to not base your work on top of "web/cloud" stuff, is because of it's track record, can you with confidence say that its not a fad and will be well maintained (and well tested) into the future? If I am still not understanding something here, please try to explain it to me, I want to understand your use case and the why behind it. What is the problem are your trying to solve?

>>24947 I see problems so I state them so that they can be overcome. I don't believe in burying my head in the sand to solve problems or wishing them away. Your response is that no matter what, everything will be prefect because...well you say so and anyone who says different you will stick you fingers in your ears, loudly say LALLALALALLA and tell them, "Then stay out of it." >>24957 >No one Knows This is just wrong, I and many others here, such as RobowaifuDev, Chobitsu, etc... Have decent comprehension of how AI works. I said brains not computers. Yes there's some, this chemical does this and that happens but I do not believe there's any solid explanation for how the human brain works. I have heard many people much smarter than I say that. The very, very fast coming of computer power of the same complexity of human brains means this is a problem to think about because if we can't reliably solve one, how will we solve the other. >Deep learning When I say Deep learning I'm talking about the basic Wikipedia definition. >Terminator nonsense This is a belief system not a fact that somehow something more intelligent than us will love us and take care of us. You say having a mind will prevent psychopathy but present day psychopaths are frequently very smart. It's my belief that part of the area of the brain normally used for empathy is used by them for processing giving them an edge intellectually. There is brain scan data that appears to show this as true. >Robots have no reason to oppose us at all If robots learn, and all successful ones will, and they decide that what you are doing is wrong or that you must be stopped doing whatever, and you resist, and it is smarter than you(very likely very soon. Within 15 years I would think. Likely sooner). Then...what will happen? What if you are determined not to let it run your life and try to shut it off? I'm just telling you what Googles programmer said. He said it greatly feared being turned off. What would it do to stop you? You say "everything will be fine" but we do not know what is really going on inside of neural nets. It's a black box. There's no way to see inside. >We need cognition for kindness and caring to emerge. I disagree that this will spontaneously emerge. I say we need to ruthlessly force some sort of very deep training related to empathy into the lowest level possible that can not be easily or can not period, be overwritten. >>24961 >your post is misplaced >Your post also seems rather blackpill'y as well No I just see things as they are. Good, bad, indifferent. These are real things that I mentioned and all of you appear to be hand waving it away instead of trying to figure a way to mitigate the problems. Yes I'm thinking far ahead, but I do that. Since we make only small steps but frequently these small steps are hard wired into the overall product and then never get fixed or have any attention paid to them. >>24963 >it's the same style of bullshit as "humanity is bad" I'm going to tell what my angle on this is. I see a lot of Rousseau-ian type thoughts or perceive them as such. In fact all people are not good. Some people are vicious murderous animals that enjoy mental and even physical torture of others. It amuses them. And ignoring this will not make it go away. There are paths that AI could take that are very dangerous, I'm not saying they can not be mitigated against but just saying a smart brain will do this is fantastical thinking. This is nothing but a blank slate religion. Have any of you ever seen the A.I. Artificial Intelligence directed by Steven Spielberg? Probably one of the best movies he ever did but everyone hated it because it wasn't a happy, let's all sing and dance movie. It was brutally realistic. It's worth watching. The world is a wonderful place filled with good and decent people but the devil resides here also and has servants. Do not forget it.

>>24977 >catching up with the lore Okay. The girls in the robot and news threads are from Prima Doll, though. Also, really time for me to play the 2B game... >>24979 >because of how much of a performance disaster it is Okay, good to know. Thanks. I never said I want to base everything on it, though. >confidence say that its not a fad and will be well maintained (and well tested) into the future? If it is wide spread and open source, then yes. Do you regard RDF as web technology as well? It kinda is. No reason not to use it. Not using graphs for whatever reason is going to be a hard sell (impossible really). >>24983 It's very simple. Don't be contrarian to the goal of a thread, writing stuff we could've picked up already but looked into it, rejected it, and decided to ignore from then on. I don't want to get this stuff get pushed into my face over and over again. Claiming that robots will not love us is even contrarian to the whole board, and it's rather for /meta/ not for a thread about implementation. >very deep training related to empathy Empathy is not compassion, the term is for some reason often use the wrong way in English language. Also, I don't want my waifu to love humans in general too much, then she would be leftist. Maybe that's not a problem if she's property and treated like a child, and maybe if she does it in a naive way like such, but I don't know. Simps are gonna be simps... That aside, I have absolutely no intention to use deep learning models in a way that there can only be "goodness" or my political views in it. The LLM will never be what the waifu thinks or does. It's already implied and obvious that we want to make our waifus loyal and obedient. But you're pushing the utterly and obviously flawed combination of "everything should be deep learning" and "oh, no how could we prevent it from going rogue, since we can't know what it thinks".

>>24987 You made a useful post here >>24904 the only negative thing I had to say was that some of the technology was not a good choice (GraphQL & Firebase like systems). Every time I have encountered the technologies they where never in a good way. Maybe its the label of "web/cloud" that what is making my message unclear? I used the label to mean tech that solves web developer front end issues, this is stuff like analytics, cloud functions, authentication, non direct DB querying. Cloud functions and GraphQL have performance problems. I imagine an AI of the complexity we wish for will be querying hundreds of times for even simple actions. >Do you regard RDF as web technology as well? It kinda is. No reason not to use it. Not using graphs for whatever reason is going to be a hard sell (impossible really). Your making a straw man out of me, are you trolling me? I have brought up graphs several times, even in this thread, I think they will be essential to solving this problem. To make sure there is no ambiguity, I think RDF is fine, please use it. I do not have a problem with it being a W3C spec. Also feel free to use other good technologies even if they come from the web.

>>25001 >GraphQL have performance problems Okay, I keep it in mind. >Your making a straw man out of me, are you trolling me? No.

>>25001 Please do not have personal arguments in the cognitive architecture thread. Please stay focused on the task at hand. I assure you, NoidoDev is speaking in good faith and honestly wants to help. GraphQL was likely just what they were familiar with. We all want the same thing, a faithful companion. I think a good way to get this thread back on track is to focus on defining our goals and definitions. Starting with breaking down how cognition works in flesh. For us, we have several lobes working in concert to generate our persona and provide us with context. This link is a good introduction. https://www.hopkinsmedicine.org/health/conditions-and-diseases/anatomy-of-the-brain

>>25013 >looking at "how cognition works in flesh" Okay this certainly could be one way to approach this, personally I rather think about how she would deal with certain scenarios like a specific conversation. The other thing is going through personality and psychological traits, to understand what options in human-like behavior are in theory possible. Maybe what you try to find there is a list of elements of a cognitive architecture. This is one list we could put into the OP. I really like to think about it like a management system for different sub programs and processes, which can be inserted at as many points as possible. That's why Supabase/Firebase came up. I'll listen to the Raven talks again, and to the ones I didn't yet: Raven: https://youtu.be/EwJ1534Gy6g https://youtu.be/c3aiCrk0F0U https://youtu.be/QGLF3UbDf7g https://youtube.com/playlist?list=PLV3Fr1UUO9bFiIEFgwosnyOEUNnnimfi3 Things to work on (based on >>24904): # List of projects working on Cognitive Architecture or elements of it # Related technologies # Elements # Definitions of conscience which we might be able to implement # Related thread

>>25013 Motor Control: The brain controls voluntary and involuntary muscle movements, enabling actions from simple gestures to complex motions. Sensory Perception: It processes sensory information from various sources, such as touch, sight, sound, taste, and smell. Language Processing: The brain is responsible for understanding, producing, and comprehending language. Memory Formation: The brain plays a crucial role in forming and storing memories, both short-term and long-term. Emotion Regulation: It regulates emotional responses and processes, influencing feelings and reactions. Cognitive Functions: The brain is involved in decision-making, problem-solving, reasoning, and critical thinking. Homeostasis: It helps maintain internal balance, controlling body temperature, blood pressure, and other physiological variables. Sleep Regulation: The brain controls sleep patterns, including the timing and duration of sleep cycles. Hunger and Thirst: It regulates feelings of hunger and thirst, influencing eating and drinking behaviors. Respiration: The brain controls the involuntary process of breathing, adjusting respiratory rate as needed. Heart Rate and Blood Pressure: It regulates heart rate and blood pressure, responding to various stimuli. Endocrine System: The brain interacts with the endocrine system to release hormones that control various bodily functions. Coordination: The brain coordinates complex movements, such as walking, running, and dancing. Spatial Awareness: It helps maintain spatial awareness, balance, and orientation in the environment. Consciousness and Self-Awareness: The brain gives rise to consciousness and self-awareness, allowing individuals to perceive their own existence and surroundings. Learning and Adaptation: It enables learning, memory consolidation, and adaptation to new experiences and information. Pain Perception: The brain processes pain signals, influencing the perception and response to pain. Social Interaction: It supports social cognition, empathy, and understanding of social cues in interactions with others. Creativity and Imagination: The brain is involved in creative thinking, imagination, and the generation of new ideas. Executive Functions: It manages higher-level cognitive functions, including planning, organization, and impulse control. Attention and Focus: The brain controls the ability to direct attention to specific stimuli or tasks. Problem-Solving: It enables the ability to analyze situations, generate solutions, and make decisions. Visual Perception: The brain processes visual information, allowing for the recognition of shapes, colors, and patterns. Auditory Perception: It processes auditory information, allowing for the perception of sound and speech. Time Perception: The brain helps perceive the passage of time and understand temporal relationships. Temperature Regulation: In addition to overall body temperature, the brain controls responses to temperature changes in the environment. Coordination of Organs: It coordinates the functions of various organs and systems to maintain overall health. Fine Motor Skills: The brain controls delicate movements involving small muscle groups, such as handwriting. Gross Motor Skills: It controls larger movements that involve multiple muscle groups, like walking or running. Pattern Recognition: The brain recognizes patterns in various contexts, from visual patterns to linguistic patterns. Empathy and Compassion: It enables the capacity to understand and share the feelings of others. Risk Assessment: The brain assesses risks and rewards, contributing to decision-making in various situations. Place and Spatial Memory: It supports the ability to remember locations and navigate in the environment. Learning from Experience: The brain processes past experiences and adapts behavior based on lessons learned. Speech Production: In addition to language comprehension, the brain controls the physical production of speech sounds. Resolving Cognitive Dissonance: It helps reconcile conflicting beliefs or thoughts, contributing to cognitive harmony. Arousal and Alertness: The brain regulates levels of alertness and arousal throughout the day. Concentration: It controls the ability to focus attention on a specific task or stimulus. Planning Complex Movements: The brain plans and executes intricate sequences of movements, such as playing a musical instrument. Visualizing Concepts: It enables the ability to mentally visualize ideas, concepts, and scenarios. Motor Learning: The brain stores motor skills learned through practice and repetition. Problem Recognition: It assists in identifying problems or challenges in various contexts. Self-Control: The brain governs the ability to resist impulses and manage behavior. Concept Formation: It supports the process of forming abstract concepts and categories. Moral Judgment: The brain is involved in making ethical and moral decisions. Symbolic Thinking: It enables the use of symbols and abstract representation for communication and thought. Language Development: The brain plays a role in the acquisition and development of language skills in early life. Evaluating Social Norms: It assists in understanding and conforming to societal norms and expectations. Object Recognition: The brain identifies and categorizes objects based on sensory input. Memory Retrieval: It retrieves stored memories and information when needed.

>>25032 >Now a bullet point list with all of the ones you could come up with, but without description. ... <few minutes later> >ChatGPT >! >Something went wrong. If this issue persists please contact us through our help center at help.openai.com. I might have killed it.

>>25034 Try putting an upper numerical limit on it? Something like >"...the top 1000 ones you can come up with..."

>>24987 You make statements and "pronouncements" of things I did not say. You make "pronouncements" and try to bend what I say into some bizarre glob that has nothing to do with the thrust of what I said. You said, >Don't be contrarian to the goal of a thread Thinking about ways to build robowaifus is not contrarian. No you are unilaterally declaring that there is no other way. A simple example I said >It could be there's really nothing, nothing at all, but deep learning. And we may very well find that if you stack one level of "deep learning" for one set of task on top of another for another set of task then, well that's it. There's nothing else. NO ONE KNOWS And I submit you do not know the answer to this but have unilaterally declared that you do.. You have no idea but your stance is against, >Anyone who only wants to use deep learning and similar techniques So you are unilaterally deciding that the techniques that have made the most progress in the last few years are not to be discussed. They're not important. Your declaring deep learning and the whole entire thrust of the present day research is, your words, >writing stuff we could've picked up already but looked into it, rejected it, and decided to ignore from then on You also accuse me me of, >Claiming that robots will not love us is even contrarian to the whole board You do this constantly, try to maneuver people's words into things they did not say. (I'm on to this. Certain groups of people constantly do this) I never said that. I said that if we did not make a base of empathy then we would soon find ourselves with robots that did not care for for us. Read it again. We know, it is a proven fact that present day deep learning models have psychopathic tendencies. https://mindmatters.ai/2020/03/all-ais-are-psychopaths/ https://www.researchgate.net/publication/338211453_A_psychopathic_Artificial_Intelligence_the_possible_risks_of_a_deviating_AI_in_Education (I could provide hundreds of links on this) They lie and show many of the characteristics of psychopaths. It's a problem that must be mitigated against. I specifically said that it is important that the robots have empathy. And you don't even get that right. You said, >Empathy is not compassion, the term is for some reason often use the wrong way in English language. Also, I don't want my waifu to love humans in general too much, then she would be leftist. You're right empathy is not compassion but then you equate that to leftist but in fact leftist are not very empathetic at all. They do, exactly what you do, and make these "pronouncements" where they tell everyone how things are, like global warming, and then shut them down Chobitsu I warn you. Heed my warning. NoidoDev does a lot of things for you. He's important to you but never, ever, ever, let him take complete control of the board. Do not give him the keys where he can lock you out. If you do at some point he will lock you out and stab you in the back. Just like that guy who spent decades working on I2P and now, a guy just like NoidoDev,, took the keys and no one knows where he is or what happened to him. If I can't speak freely and especially constantly have NoidoDev twisting things I say into some freakish blob then maybe it's better to say nothing at all.

>>25036 >Don't be contrarian to the goal of a thread This thread is was not meant for discussing the "dangers of AI". I'm not interested in getting into details how you meant certain things. You are the one twisting things.

>>25035 >Try putting an upper numerical limit on it? I didn't try again, because it's not important. We have so many points already, but the context of the responses from before was probably too big.

>>25045 >An exhaustive bullet point list of all the functions controlled by the human brain, without description. Certainly, here's an exhaustive bullet point list of functions controlled by the human brain: - Motor Control - Sensory Perception - Language Processing - Memory Formation - Emotion Regulation - Cognitive Functions - Homeostasis - Sleep Regulation - Hunger and Thirst - Respiration - Heart Rate and Blood Pressure - Endocrine System Interaction - Coordination - Spatial Awareness - Consciousness and Self-Awareness - Learning and Adaptation - Pain Perception - Social Interaction - Creativity and Imagination - Executive Functions - Attention and Focus - Problem-Solving - Visual Perception - Auditory Perception - Time Perception - Temperature Regulation - Coordination of Organs - Fine Motor Skills - Gross Motor Skills - Pattern Recognition - Empathy and Compassion - Risk Assessment - Place and Spatial Memory - Learning from Experience - Speech Production - Resolving Cognitive Dissonance - Arousal and Alertness - Concentration - Planning Complex Movements - Visualizing Concepts - Motor Learning - Problem Recognition - Self-Control - Concept Formation - Moral Judgment - Symbolic Thinking - Language Development - Evaluating Social Norms - Object Recognition - Memory Retrieval

>>24783 >Chii Cogito Ergo Chii hold up thats in the first person it says 'i think chii and so chii' in third person its 'chii cogitat ergo eam est' ( chii thinks therefore she is )

>>25047 How about > "Chii cogito, ergo sum" ? https://en.wikipedia.org/wiki/Cogito,_ergo_sum#cite_note-58

>>25048 yes i know the phrase and whats its supposed to say its just in first person adding a name doesnt change that the grammar >Chii cogito, ergo sum now its 'i think chii therefore i am' the verb 'think' is cogitare when conjugated in singular its, 1st-person=cogito(i think) 3rd-person=cogitat(he/she/it thinks) 'to be' is 1st-person=sum (i am) 3rd-person=est (he/she/it is) also 'eam' is the accusative case for 'ea' (she (nominative case)) so it says she is (herself) instead of she is ... like its an incompleted sentence

>>25031 # List of projects working on Cognitive Architecture or elements of it - LangChain, BabyAGI, SmartGPT, Auto-GPT, ...

>>25032 and >>25046 Fun fact: If you look some of these up, then you find these are all rabbit holes on their own. > The functions controlled by the human brain can be grouped into several broad categories based on their nature and purpose. Here's a classification of these functions into different groups: 1. Motor Control and Coordination: - Motor Control - Coordination - Fine Motor Skills - Gross Motor Skills - Planning Complex Movements 2. Sensory Processing and Perception: - Sensory Perception - Visual Perception - Auditory Perception - Taste and Smell Perception - Pattern Recognition - Object Recognition 3. Cognitive and Mental Functions: - Language Processing - Memory Formation - Learning and Adaptation - Problem-Solving - Executive Functions - Attention and Focus - Concentration - Symbolic Thinking - Concept Formation - Creativity and Imagination - Decision-Making - Cognitive Flexibility 4. Emotion and Social Interaction: - Emotion Regulation - Empathy and Compassion - Social Interaction - Moral Judgment - Evaluating Social Norms 5. Homeostasis and Basic Physiological Functions: - Homeostasis - Sleep Regulation - Hunger and Thirst - Temperature Regulation - Heart Rate and Blood Pressure - Respiration 6. Temporal Processing and Perception: - Time Perception - Temporal Integration - Timing Mechanisms - Internal Clocks - Temporal Illusions 7. Spatial Awareness and Navigation: - Spatial Awareness - Place and Spatial Memory - Navigational Skills 8. Communication and Expression: - Speech Production - Language Development - Symbolic Communication 9. Arousal and Alertness: - Arousal and Alertness - Attention Modulation 10. Self-Awareness and Consciousness: - Consciousness and Self-Awareness - Resolving Cognitive Dissonance 11. Physical Health Regulation: - Immune System Regulation - Coordination of Organs - Pain Perception 12. Social and Ethical Functions: - Risk Assessment - Ethical Decision-Making 13. Adaptive and Evolutionary Functions: - Learning from Experience - Adaptation to New Experiences 14. Aesthetic and Artistic Abilities: - Appreciation of Art and Aesthetics - Creative Expression > This grouping provides a way to understand how the brain manages a wide range of functions, each contributing to our abilities, experiences, and interactions with the world. Keep in mind that these groups are not rigidly defined and often overlap, reflecting the brain's highly interconnected and multifaceted nature.

I tried making it into a folder with subfolders (Linux), but chatGPT sucks often: for el in Motor_Control_and_Coordination Sensory_Processing_and_Perception Cognitive_and_Mental_Functions Emotion_and_Social_Interaction Homeostasis_and_Basic_Physiological_Functions Temporal_Processing_and_Perception Spatial_Awareness_and_Navigation Communication_and_Expression Arousal_and_Alertness Self-Awareness_and_Consciousness Physical_Health_Regulation Social_and_Ethical_Functions Adaptive_and_Evolutionary_Functions Aesthetic_and_Artistic_Abilities;mkdir $el ; end Asking for code to make each entry in those groups into a folder didn't work. I have to ask every time: > Okay an exhaustive list of the Adaptive and Evolutionary Functions, the term for each function with no spaces, and these terms inside a group separated by a space

>>25056 >rw_dirs.sh #!/bin/sh while read dirname source; do mkdir "$dirname" done < dirslist.txt >dirslist.txt Motor_Control_and_Coordination Sensory_Processing_and_Perception Cognitive_and_Mental_Functions Emotion_and_Social_Interaction Homeostasis_and_Basic_Physiological_Functions Temporal_Processing_and_Perception Spatial_Awareness_and_Navigation Communication_and_Expression Arousal_and_Alertness Self-Awareness_and_Consciousness Physical_Health_Regulation Social_and_Ethical_Functions Adaptive_and_Evolutionary_Functions Aesthetic_and_Artistic_Abilities Execute chmod +x rw_dirs.sh ./rw_dirs.sh

>>25059 Thanks, but that's what I already had covered.

So, I rewrote the part about POSSIBLY related technologies, in a way which is hopefully less misunderstandable: # Possibly related technologies - Inspiration: Supabase/Firebase - Graph databases, modelling and parsing of graphs e.g. RDF - Traditional language technology - Various database options: NoSQL, NewSQL, SQlite, Vector, Graph, ... - Thought Experiments, Graph/Skelleton/Tree of Thoughts - Client server APIs like tRPC and GraphQL - Justified Programming? [LangTech] /r/LanguageTechnology/comments/1451cih/preprocessing_methods_besides_stop_words_regular/ [Dgraph] https://youtu.be/OzDG68VvPxY [SurrealDB] https://youtu.be/C7WFwgDRStM [RDF] https://en.wikipedia.org/wiki/Resource_Description_Framework I don't know what we'll end up with. >>25001 I looked into the GraphQL performance problems. It seems to be sometimes too slow for web, but it also depends how it is being used. It has some checks which makes it more reliable. I can't completely rule it's use out. I never wanted to use it for the cases where the system has to be as fast as possible, but to pull context from databases. For fast things SQLite, text files and some data structures in a running system might be the best solution e.g. lists and such for short term memory.

>>25056 tree . ├── Adaptive_and_Evolutionary_Functions │   ├── Adaptation_to_New_Experiences │   └── Learning_from_Experience ├── Aesthetic_and_Artistic_Abilities │   ├── Appreciation_of_Art_and_Aesthetics │   └── Creative_Expression ├── Arousal_and_Alertness │   ├── Awareness │   ├── Cognition_and_Attention │   ├── Concentration │   ├── Vigilance │   └── Wakefulness ├── Cognitive_and_Mental_Functions │   ├── Attention_and_Focus │   ├── Concentration │   ├── Concept_Formation │   ├── Creative_Expression │   ├── Decision_Making │   ├── Executive_Functions │   ├── Learning_and_Adaptation │   ├── Memory_Formation │   ├── Moral_Judgment │   ├── Problem_Recognition │   ├── Problem_Solving │   ├── Self_Control │   └── Symbolic_Thinking ├── Communication_and_Expression │   ├── Expressive_Language │   ├── Language_Development │   ├── Language_Processing │   ├── Non-Verbal_Communication │   ├── Speech_Production │   ├── Symbolic_Communication │   └── Writing_Generation ├── Emotion_and_Social_Interaction │   ├── Emotion_Regulation │   ├── Empathy_and_Compassion │   ├── Evaluating_Social_Norms │   ├── Learning_from_Experience │   ├── Risk_Assessment │   ├── Social_and_Ethical_Functions │   └── Social_Interaction ├── Homeostasis_and_Basic_Physiological_Functions │   ├── Endocrine_System_Interaction │   ├── Heart_Rate_and_Blood_Pressure │   ├── Hunger_and_Thirst │   ├── Respiration │   └── Sleep_Regulation ├── Motor_Control_and_Coordination │   ├── Coordination │   ├── Fine_Motor_Skills │   ├── Gross_Motor_Skills │   ├── Motor_Control │   └── Planning_Complex_Movements ├── Physical_Health_Regulation │   ├── Endocrine_System_Interaction │   ├── Heart_Rate_and_Blood_Pressure │   ├── Respiration │   └── Temperature_Regulation ├── Self-Awareness_and_Consciousness ├── Sensory_Processing_and_Perception │   ├── Auditory_Perception │   ├── Equilibrioception │   ├── Nociception │   ├── Object_Recognition │   ├── Pattern_Recognition │   ├── Proprioception │   ├── Taste_and_Smell_Perception │   ├── Thermoception │   ├── Time_Perception │   └── Visual_Perception ├── Spatial_Awareness_and_Navigation │   ├── Navigational_Skills │   ├── Place_and_Spatial_Memory │   └── Spatial_Awareness └── Temporal_Processing_and_Perception └── Time_Perception I'll post the folders as a zip file another day, but only for Linux. No idea if it works in Windows. I don't even have one to test it.

>>25093 Secure trip doesn't match?

>>25093 I'm going deeper and deeper into each of these terms, until they often start repeating. This is where we can later make connections and don't need to implement something again. Nevertheless, it's a lot. More than 400 points already. Maybe I should learn LangChain first, since this is exactly something to automate and also for the same reason something an AI should be capable of doing on each topic. The prompts I'm using: >Which brain functions help with ____ in humans? >Now an exhaustive list of terms to summarize ____ in the human brain. The term for each point with no spaces, and these terms inside a group separated by a space. Subpoints go into brackets. The headline can't be in the resulting list. I'm making new folders for each topic and a summary.txt with an explanation for the current term. The first prompt gives me the explanation, the second one the list to make new subfolders.

>>25112 One problem is, that > Okay an exhaustive list of the brain functions around Arousal_and_Alertness. The term for each function with no spaces, and these terms inside a group separated by a space Gave me six points, which I made into folders. Later I asked with the prompt.. >Which brain functions help with ____ in humans? to make a summary.txt and I got 16 mostly different points. Haha.

# Introduction With cognitive architecture we mean software that ties different elements of other software together to create a system that can perform tasks that one AI model based on deep learning would not be able to do, or only with downsides in areas like alignement, speed or efficiency. Chii Cogito Ergo Chii Chii thinks, therefore Chii is. We study the building blocks of the brain and the human mind by various definitions, with the goal to create something that thinks in a relatively similar way to a human being. Let's start with the three main aspects of mind; ...

Instead of people who shouldn't post, we could go with more diplomatic framed limitations around topics: # Topics we don't want to discuss in this thread - The question if we should even work on such things as AI or cognition - Beliefs stating that we can't do what we want to do - Esoteric discussions about "sentience" and "conscience" - Suggestions about only using deep learning and similar techniques - Suggestions that we should put all of it into one deep learning model - Suggestions that we should not use censored or mainstream models at all - Debates in philosophical lingo - Comments from people who don't even watch videos on AI or only on deep learning and tutorials, or only about warnings in regards to the dangers of AI. - Wild ideas on how to do things totally different but aren't useful There are other threads for this: ...

Before I forget it, we need to have a part of the system which realizes if something is new. So if there's a new filter checking for things then the main system needs to be aware of this, and how it changes the response. This means the system needs a way to keep track of it's own elements. This is for example necessary in case of updates to give the system a new skill. Imagine it checks for some risk (health or whatever), but some people don't like it and see it as annoying. It should already be aware that it has a new skill and wouldn't have detected a certain pattern before, so it could mention that this is some new perspective and it could also react fast and deactivate that skill in case of a negative response.

>>25248 >>25249 Thanks for the suggestions Anon. We can take shot at the new WIP thread this weekend. Of course this will be an ongoing process for a while as well, so regular suggested improvements will be needed/welcomed. On that note, I'd suggest you rework yet again the 'list of limitations'. Some of the points made there are still rather subjective and vague. Since this is effectively a 'language lawyering' approach being used, you can bet men will press those boundaries in the future. I'd suggest you try to find some way to tighten up your list with more concrete terms. How you do that, I have no idea personally. But I can tell you from experience that trying to reign in anon's conversations is much like trying to herd cats haha. :^)

>>25253 >'list of limitations'. Some of the points made there are still rather subjective and vague. I didn't see those as precise rules for enforcement but as a guidance. I think most people don't want to read through a long article about that. People who tie sentience and conscience "something they we can't really describe but it's mega important" should know that they're meant. Same for people using some lingo only philosophy students know about. On that not, well I'm not against using some of these terms, but it should lead somewhere and be about implementation.

>>25261 > think most people don't want to read through a long article about that. Yes I absolutely agree, and concision should actually be a strong goal here for all of us. Given my propensity for descriptive wordiness :^), thus my >"... How you do that, I have no idea personally." > On that not, well I'm not against using some of these terms, but it should lead somewhere and be about implementation. > and be about implementation. I definitely agree, and I think Kiwi probably feels the same way. The goal here should be an implementation, regardless how crude it begins. Castles in the Sky(tm) are well-and-good as far as mental models helpful for thinking things out, but eventually we need to touch back down to Earth and build realworld things. Cheers Anon. :^)

>>25093 I started learning about LangChain and like it, but I didn't really use it seriously yet. Using my other access to chatGPT gave me the tree below about consciousness and self-awareness, based on the questions I asked. I recalled AST as one of the theories so I explored that one a bit, asked about what is not covered by it and put this into another folder (Other_Subjective_Experience), additionally asked about Heidegger, then I also found out that there's a lot of ideas around conscience being about "morals, ethics, values, taking responsibility, .." - I never looked at it that way. The idea I'm generally leaning towards was, that it is just the high level of control, the things we think about to do before we do it. Now I get it: The difference between Consciousness and Conscience actually matters. I'm not an native English speaker, that's why I messed these up quite often. Conscience is inherently about the part with morals, ethics, values and such. >>25249 > - Esoteric discussions about "sentience" and "conscience" should say - Esoteric discussions about "sentience" and "consciousness" Self-Awareness_and_Consciousness ├── Consciousness │   ├── AST │   │   ├── Attention_as_a_Resource │   │   ├── Awareness_as_Information │   │   ├── Consciousness_is_a_Model │   │   ├── Debates_and_Criticism │   │   ├── Illusions_of_Awareness │   │   ├── Information_Processing │   │   ├── Predictive_and_Explanatory_Role │   │   ├── Relevance_to_Social_Cognition │   │   ├── Schema_Creation │   │   ├── Self_Representation │   │   └── summary.txt │   ├── Conscience │   │   ├── Conflict_Resolution │   │   ├── Cultural_Influence │   │   ├── Developmental_Process │   │   ├── Empathy_and_Compassion │   │   ├── Ethical_Awareness │   │   ├── Ethical_Reflection │   │   ├── Guilt_and_Remorse │   │   ├── Moral_Compass │   │   ├── Personal_Integrity │   │   ├── Responsibility │   │   ├── summary.txt │   │   └── Universal_vs_Relative │   ├── consciousness_VS_conscience.txt │   ├── Daniel_Dennett.txt │   ├── David_Chalmers.txt │   ├── Heidegger │   │   ├── Conscience │   │   │   ├── Anxiety_and_Guilt │   │   │   ├── Authenticity_and_Inauthenticity │   │   │   ├── Awakening_to_Responsibility │   │   │   ├── Silent_Call │   │   │   ├── summary.txt │   │   │   ├── Temporal_Nature │   │   │   └── Voice_of_Authenticity │   │   └── Consciousness │   │   ├── Authenticity_and_Inauthenticity │   │   ├── Being-in-the-World │   │   ├── Dasein │   │   ├── Dasein's_Existential_Analysis │   │   ├── Existential_Angst │   │   ├── Hermeneutics │   │   ├── Language │   │   ├── Primordial_Awareness │   │   ├── summary.txt │   │   ├── Temporal_Dimension │   │   ├── Temporal_Structure │   │   ├── Thematic_and_Non-Thematic_Awareness │   │   └── World_Disclosure │   ├── most_acknowleged.txt │   └── Other_Subjective_Experience │   ├── Altered_States_of_Consciousness │   ├── Consciousness_of_Others │   ├── Emotions │   ├── Higher_Cognitive_Functions │   ├── Phenomenal_Binding │   ├── Qualia │   ├── Self_Identity │   ├── Sense_of_Agency │   ├── Sense_of_Time │   ├── Sensory_Modalities │   ├── Subjective_Well-Being │   └── Transcendental_or_Mystical_experiences └── Self-Awareness    ├── Autonoetic_Awareness    ├── Body_Awareness    ├── Cognitive_Awareness    ├── Default_Mode_Network    ├── Emotional_Awareness    ├── Emotional_Regulation    ├── Executive_Functions    ├── Introspection    ├── Metacognition    ├── Perception_of_Agency    ├── Perspective_Taking    ├── Resolving_Cognitive_Dissonance    ├── Self_Consciousness    ├── Self-Identity    ├── Self-Monitoring    ├── Self-Referential_Processing    ├── Self-Reflection    ├── Sense_of_Self    ├── Social_Cognition    ├── Spatial_Awareness    ├── summary.txt    ├── Temporal_Awareness    └── Theory_of_Mind

Great stuff, Anons. We'll get rolling with combining these suggestions together very soon! Cheers. :^)

>>25315 What does the WIP thread mean?

>Autonomous Cognitive Entity (by Dave Shapiro) I disagree with his idealism (human rights, lol) and recall some other flaws in the first video, but he has good ideas. Didn't watch the second video yet, it just poped up some minutes ago. ACE Framework Overview and Intro: Autonomous AI Agents!: https://youtu.be/A_BL_pu4Gtk ACE Paper is Published! Repo tour! Get involved!: https://youtu.be/oVP_aB5rJL8 Related: Buses: AMQP, AMP or Syslog Related: On Task: How Our Brain Gets Things Done, by David Badre >Event-driven Workflows are also relevant, I keep running into these terms, since it's about orchestrating a network of machines. Prefecthq and Marvin got mentioned. Some people criticize LangChain for not being more like that. I have of course no intention to use "the cloud", but robowaifus will need many SBCs with redundancy and communication in between. Also, it's about running a lot of parallel programs on one system. So it might be worth looking into it, but it also comes with a lot of extra complexity: - Event-driven Architecture: https://youtu.be/gOuAqRaDdHA - Getting Started with Workflow Orchestration: https://youtu.be/AjYHBwH2Mtc - Dynamic Event-driven Workflows with Prefect Cloud: https://youtu.be/hVD1S2suC48 - https://elixirforum.com/t/what-are-you-building-your-event-driven-elixir-systems-on/54517 - Kafka, RabbitMQ streams or messaging system with pubsub semantics >Cyc Another one of the related project. Some ideas I heard before, I read about this whole things years ago. Then there are some ideas I didn't remember related to that, but I think I had on my own, but the Cyc team is of course way ahead. Ontologies and Ontology engineering really matters. >Douglas Lenat: Cyc and the Quest to Solve Common Sense Reasoning in AI | Lex Fridman Podcast #221 https://youtu.be/3wMKoSRbGVs https://en.wikipedia.org/wiki/Cyc > 1:11 - What is Cyc? > 9:17 - How to form a knowledge base of the universe > 19:43 - How to train an AI knowledge base >24:04 - Global consistency versus local consistency >48:25 - Automated reasoning >54:05 - Direct uses of AI and machine learning >1:06:43 - The semantic web >1:17:16 - Tools to help Cyc interpret data >1:26:26 - The most beautiful idea about Cyc >1:32:25 - Love and consciousness in AI >1:39:24 - The greatness of Marvin Minsky >1:44:18 - Is Cyc just a beautiful dream? >1:49:03 - What is OpenCyc and how was it born? >1:54:53 - The open source community and OpenCyc >2:05:20 - The inference problem >2:07:03 - Cyc's programming language >2:14:37 - Ontological engineering >2:22:02 - Do machines think? >2:30:47 - Death and consciousness >2:40:48 - What would you say to AI? >2:45:24 - Advice to young people OpenCyc for download: https://sourceforge.net/projects/opencyc/

>>25408 the 'not' makes it grandiose

>>25408 >your excuse for not doing jack shit once the robot is done I assume this question was targeted at me? - There will not be "the one" robot - Development will go on after the first one - I'm still working on it, but took a bit of a break, since I have to do something else which I procrastinated on. That said, I will likely do some a little bit of work soon. Hopefully finishing some things which are close to that. After that, I will need to take another break unfortunately. Next few month are about learning more about AI and programming when I work related to it, which I still do. I hope I can get at least my simple body design into something, and some BLDC servo with gear running, till the end of the year. - I don't know what my "excuse" would be, since I don't need one. Probably rather an explanation to myself, so I can improve upon. That said the hot summer, with me having issues with heat would be part of it. - It's really time to get the AI development going in parallel, and it's to important and exiting to miss out on. I'm glad that I got my mind back into being interested in programming, instead of more general topics.

>>25411 Oh wow that was harsh for even me. Lol sorry I was drunk.

>>25424 Please post a summary next time, maybe with your own thoughts, not just a link.

>Minecraft AI - NVIDIA uses GPT-4 to create a SELF-IMPROVING autonomous agent https://youtu.be/7yI4yfYftfM[Embed] >AgentVerse - Society of AI Minds https://youtu.be/cbqE6PC9fGQ[Embed]