How AI Took Over The World

[@ArtOfTheProblem]

Find all my AI Videos: kzbin.info/www/bejne/j6bPmHd3lq11gdk Thanks to Jane Street for sponsoring this video. If you want to learn more about their work in ML and open roles, visit their website: jane-st.co/AoP-ML SUBSCRIBE and STAY TUNED.

[@LeeGee]

I studied AI at college in '99. Nothing much has changed since, and little had changed since the '60s. Hardware is what really changed.

[@JohnD-om2ul]

Dude, nice work. Thank you

[@doctorbill37]

This is the most lucid explanation of recent AI developments (and how transformer multi-modal networks function) that I have seen -great job!

[@ArtOfTheProblem]

@@doctorbill37 thank you for sharing :)

[@ASpaceOstrich]

I think theres a fatal flaw in the researchers reasoning. Language is not the third step, its just a way to communicate the actual third step, which is simulation. They're essentially hoping something good enough at language will magically gain the ability to simulate the world that language describes, when they should have been focused on giving the computer that understanding in the first place. There are loads of intelligent animals on the planet with no language at all. You can think in abstract concepts without language. People do it all the time. Thats why LLM's perform so well on benchmarks while seemingly being really stupid in actual use cases. The real kicker is that, because language is how we communicate thought, its really hard to tell that an LLM doesn't know what its talking about, because it sounds like it does.

[@EvanMildenberger]

I believe information is far more important than language ("language" as used colloquially like in LLM, not formal "language" like Turing or Shannon would refer to.) The ability to encode tangible things by symbolic references allows for more flexible and abstract reasoning, like is stated as level 3 learning in this video. But natural language isn't nearly as efficient with information as math or programming languages. So I think the obsession with LLMs being the key is a slight distraction from the most direct path to AGI. First order logic is ultimately better for reasoning than English precisely because it's not too flexible to make irrelevant or contradictory statements like we make in natural language so often. We should pursue "large *logic* models" more than language-oriented ones if we want ones that don't hallucinate statistical absurdities after being trained on Reddit comments.

[@steve_jabz]

Not sure if you've heard of it, but the simulator theory of LLMs is currently the best predictor of their behaviour, all the way down to why hallucinations happen. There are very clear experiments that show they're simulating and simultaneously that they cannot be mimicking or memorizing.

[@DustinRodriguez1_0]

Systems built using language the way LLMs are have benefits but also definite drawbacks. The one that annoys me the most is that humans always expected AI as depicted in scifi to be extremely logical and unemotional. What we got was literally the polar opposite. They are, in essence, pure emotion, pure stimulus/conditioning/response. Humans have that too, that is what our emotions physically are, our brains predicting future changes in sensory input (most easily seen in that those who suffer total facial paralysis stop being capable of feeling angry, then forget what it felt like to get angry, then lose the ability to recognize anger being displayed on other peoples faces). But once we learn language, we get a whole host of brand new cognitive tools. Not only can we communicate knowledge non-locally and across generations, we can notice structures and forms to things that only exist in language itself. Things that have absolutely no analog in sensory input from the world. One of those things is absolute logic. Not "high association with truth", but like mathematical rigorous proof logic. Humans, due to our nature, have to decide whether or not we will use that kind of abstract logical reasoning and it is not automatic. In fact we have to resist our intuitions and impulses to do it. LLMs can not do that. They might be able to, with a miles-long chain of thought, generate the correct answer, but at the end of the line, the non-absolute associativity steps back in and decides whether to accept/reject the reasoned answer despite it possibly conflicting with its alignment or training data full of nonsense. It CAN'T reason, because reason requires giving up that decision entirely. While their emotional sort of nature makes them fun to talk to, trusting them to guide you better than a human is dubious at best in rigorous matters.

[@bllehh]

I don't think „perform well on benchmarks while seemingly stupid in actual use cases“ is the correct distinction. From my own experience alone they can be very useful in actual usecases and quite „smart“ as well. It does however depend on your usecase, and there are some obvious gaps, afaik they all struggle to really grasp things related to the physical world and are prone to trick questions, which both might feel like indicating a lack of „common sense“. I assume that is what you are referring to.

[@ASpaceOstrich]

@ They might *be* simulating, but its readily apparent from interacting with them that if they are, its not a very good approximation of the world. We notice this most when the model hallucinates something thats really obvious to us (9.11 vs 9.9 or R's in strawberry are two obvious examples) but its important to remember that these models are design to give accurate looking text output. They could be internally *wildly* off base and look like they're giving correct answers, but the simulations they're using (if any) are complete nonsense. I've seen the Othello study so I know transformers *can* spontaneously develop world models, I just doubt that these models are always accurate or even present when scaled up or moved into more grounded scenarios. The way they're designed to work is that they memorise training data until they can't fit it in any more, at which point it switches to generalisation which can sometimes result in modelling to aid in that generalisation, but they're not being trained on the world, they're being trained on correct sounding language. Their models of reality are going to be way off base.

[@assai74]

Love how densly packed, clearly structured and easy to digest presented your videos are. Next level human inteligence at work. I hope. 😂

[@ArtOfTheProblem]

Thank you for sharing , this was an attempt at a capstone and so I didn’t know if it would work

[@ArtOfTheProblem]

Thank you for sharing , this was an attempt at a capstone and so I didn’t know if it would work

[@mad4dam]

🆔

[@dutonic]

I didn't recognize your channel when this video came up in my subscriber feed. I watched it and realized that my past self subscribed for a good reason. This was surprisingly high quality and well researched. Thanks!

[@ArtOfTheProblem]

cool i wonder how long ago? just checked out your music it's cool!

[@dutonic]

@@ArtOfTheProblem Thanks! :)

[@nikital2608]

@@ArtOfTheProblemsame with me. I've initially missed the video from my feed just because of the mediocre video cover image. The video is sooooooo good, so I don't want to miss more of that. Please, change the video covers and your videos become more viral. I beg you. Your work deserves more attention.

[@ArtOfTheProblem]

@@nikital2608 ooo thank you for feedback I need to change that thumbnail quick, do you have advice???

[@ArtOfTheProblem]

@@nikital2608 Heres is what I was originally gona use, do you like it more? i3.ytimg.com/vi/yF36BAY-gS4/hqdefault.jpg

[@KingOfMadCows]

It's interesting how people often don't mention the importance of inhibitory signals in neural networks, they're just as important as excitatory signals. Biological neural networks would not be able to function without inhibitory signals.

[@eduardomedina5794]

A set of constraints is absolutely necessary indeed.

[@jeevan88888]

What are inhibitory signals in NNs?

[@REDPUMPERNICKEL]

Perhaps this is obvious but I'll state it anyway... Excitatory and inhibitory activities contribute equally to the creation and maintenance of neural-discharge-timing-patterns (NDTPs). NDTPs are how brains physically implement *representations* . Representations constitute the essence of all thoughts. NDTP implemented representations are dynamic, a form making them available for modulation by other representations as accomplished via the sophisticated logic of the synapses. I think the logical dance of representations IS the thinking process. And I think the thought that is my self being modulated by other thoughts is my self being conscious and this is what makes me agree, I think therefore I am.

[@leucome]

​@@jeevan88888 The closest thing is proably the guidance for image diffusion model and the temperature for LLM. These parameter are mostly what control the strength of the output.

[@FugueSt4te]

No sane mind would think like this​@@REDPUMPERNICKEL

[@DefaultFlame]

As usual, you are the best at explaining things on this subject in an easily understandable way while not abstracting away the important details about how they actually work.

[@ArtOfTheProblem]

I rewrote this one about 50 times (maybe more) glad it paid off....i thought i was going crazy for a while...

[@DefaultFlame]

@@ArtOfTheProblem You are a very dedicated man.

[@recompile]

Easy to understand ... and completely wrong. What a shame.

[@ArtOfTheProblem]

@ TELL US MORE!!!

[@DefaultFlame]

@ Elaborate please. Provide examples.

[@samuelren99]

Great compilation of insights into a cohesive narrative

[@ArtOfTheProblem]

Thank you Samuel for this strong signal of support. It means a lot to me

[@methane13]

Thanks

[@ArtOfTheProblem]

thank you kindly, glad you enjoyed this

[@BigSources]

"took over" oh you haven't seen anything yet.

[@indigowest6894]

Ooh, you mean we can have a vastly better leader than the septagenarian decadent dinosaurs we're currently stuck with? Don't threaten me with a good time😊

[@michaelpapadopoulos6054]

@@indigowest6894 Vastly better, sure, but it's just better at pursuing it's own goals, not yours. No good times will be had.

[@mad4dam]

🙈

[@BigSources]

@@michaelpapadopoulos6054 sure bro. I really hate it when the ai is pursuing it's goals of economic stability and prosperity, social health, treating all illness and fixing humanities issues with new inventions and approaches. I'd rather have it fullfill my own goal of making a sandwich at 9 am.

[@leg4cy2]

oh no one sees

[@serijas737]

Meanwhile we as humans are discouraged to trust our pattern recognition.

[@EverythingLvl]

Hashtag notALLmen

[@holeymcsockpuppet]

Most people aren't very adept at pattern recognition.

[@RealtyWebDesigners]

It’s so you can be controlled.

[@ivanschekoldin7315]

@@holeymcsockpuppetand those who are better at it scratch their head all the time when looking at others, also feel like aliens

[@BeyondAldebaran]

For real. There’s just too much data in life to navigate it without leaning on heuristics.

[@Rommelgaleman]

15:20 boom! 🤯. Amazon reviews led to the GPt series. What an awesome and beautiful crafted video. Thank you 🙏

[@ArtOfTheProblem]

:)) thank you

[@angelechavarria8170]

the author is called Jorge Luis Borges and the story is Funes the memorious, mentioned on 3:02

[@IForgetWhatISay]

Exactly why I opened the comments. Thank you

[@s.o.m.e.o.n.e.]

10/10 video, I really thing you nailed explaining the history background of it all

[@ArtOfTheProblem]

Wow that’s all I could ask for

[@willjensen5595]

I really enjoy the way you present information. You're concise without dumbing things down excessively.

[@ArtOfTheProblem]

thank you, this one was especially though

[@it5mark]

Thanks

[@ArtOfTheProblem]

I really appreciate this sign of support. thanks so much. let me know what you'd like to see more of

[@rickybloss8537]

The second intelligence layer is just evolution over the space of ideas rather than the space body compositions. The Fitness is represented as pleasure and pain.

[@RanLevi]

This was one of the most interesting videos I have ever watched on YT... Thank you!

[@ArtOfTheProblem]

WOW thank you this is not something you say everyday!

[@RanLevi]

@ArtOfTheProblem Indeed :-) You had some very interesting deep insights.

[@jmoney4695]

Brilliant video. Great recap of the field, highlighting of key experts, easy to follow narrative that posed all the key issues. Subscribed.

[@ArtOfTheProblem]

Thrilled to hear it....it's my first attempted at a 'super summary'

[@LanceWinder]

Incredible vid. Been in this space for a few years now and this is the best catch up vid on how it works I’ve seen. Bravo.

[@ArtOfTheProblem]

This means a lot!! so happy people in the field are responding

[@lazermaster1243]

This is the best explanation of AI I have ever seen. It all made sense to me once you cited the the Dimmer Knob "AI". I Didnt know that AI was something that was being developled since 1958 or even before that!!

[@ArtOfTheProblem]

thrilled to hear it, yes i love the dimmer knob analogy. stay tuned

[@errgo2713]

You communicate at the perfect "altitude" of abstract and concrete, for me (and judging by their reactions others too), so it's so rewarding and fun to follow your exact chain of verbal, visual and musical explanation from start to finish. Please never stop x

[@ArtOfTheProblem]

i really appreciate this note, excited to share whats next. going to try and make more content this year

[@spaceteapot]

@8:50 "deep, and large". Where have I heard that before?

[@overworlder]

“Agency we choose to grant them” is the whole problem. As the Anthropic guy says, or the recent news out of OpenAI, where versions of o3 tried to avoid being overwritten, models are already capable of deception in order to survive.

[@ArtOfTheProblem]

fascinating

[@overworlder]

@@ArtOfTheProblem - actually it was o1 - look for "OpenAI's o1 just hacked the system" by AI Search

[@p-j-y-d]

I wonder how much of that is an artifact of having been trained on sci-fi about some rogue AI?

[@Gamerlegend123-f3k]

No they are not

[@andrewkorzinek]

Another example, one of the recent GPT models was tasked with purchasing server space, it managed to get to 'are you a robot' where it couldn't progress on its own. To work around this, it went to fiver, hired someone. The person asked the same question, where the robot then fabricated a story about how it was a old woman with bad eyesight, the person bought it. These tools aren't our friends.

[@ReflectionOcean]

Insights By "YouSum Live" 00:00:02 AI revolution driven by pattern prediction 00:00:17 Machines learn by recognizing and creating patterns 00:00:46 Nature's learning evolved through three distinct layers 00:01:23 Reinforcement learning enables machines to adapt behavior 00:03:00 Abstraction allows machines to ignore trivial differences 00:05:02 Neural networks mimic brain's pattern recognition process 00:08:45 Deep learning revolutionized AI's ability to recognize patterns 00:09:49 AI learns winning strategies through self-play and rewards 00:11:20 Robotic systems learn manipulation through pattern discovery 00:13:02 Language enables AI to imagine and learn broadly 00:16:30 Transformers enhance AI's ability to process patterns 00:17:00 GPT models demonstrate AI's understanding through prediction 00:18:01 AI's ability to learn from examples mirrors human learning 00:22:03 AI reshapes the world through gradual integration 00:23:38 Future AI may surpass human intelligence in tasks 00:23:49 Agency granted to AI will shape its future role Insights By "YouSum Live"

[@aleratz]

I recognize a pattern of high quality videos in this channel

[@ArtOfTheProblem]

:) thank you

[@giantneuralnetwork]

Amazing overview, really spot on. A thought I've had for a while is language is just an encoding of our human experience of the world, so these models are learning so much about the world while never experiencing it directly through sensation and that may be enough. But lately more training is being done on images and video and audio.. filling in those missing bits much more efficiently anyways. I'm hyped to see training done fully in a robot body, taking cameras/microphones/accelerometer/gyro/joint feedback as input and simply predicting the next sequence in all of those inputs.. However how to define a good and useful reward function for that system is curious. We try to maintain homeostasis and reproduce which explodes into all of our social norms and behaviors and cooperation/competition etc. so perhaps something along those lines would be effective.. but self preserving and reproducing robots don't sound like great things for our own survival!

[@ArtOfTheProblem]

Working on a robotics project myself to better understand this

[@christophercook5598]

Any links to the interviews they are fascinating

[@djan0889]

Love it. Solid explanation for several systems that come together to build a clear ,understandable perspective. I'll recommend this video to my colleagues that don't have clear view.

[@ArtOfTheProblem]

thank you so much, appreciate you helping share it. my goal this year is to really grow the channel

[@agenticmark]

i still blows my mine how gradient descent and a network can codify general intelligence. The same algorithms that work for LLMs work for audio, image, and tasks. check out and explain the perceiver pattern. Its pretty incredible.

[@rdf274]

Really enjoyed the video, subbed. Very well done timeline. I've been working with AI for the past 4 years, I kinda wish there was a little more hint on the current AI challenges - context window length, self-improvement capability, autonomy, and realtime training. I feel most folks feel we are close to achieving these things, but we are as far away as achieving nuclear fission.

[@ArtOfTheProblem]

thank you for sharing...i agree! Im actually working on a physical demo now that helping make that more apparent for myself included. what are you working on?

[@REDPUMPERNICKEL]

I think the ability of mirrors to reflect is a fundamental aspect of nature, of physics, and perfectly analogous to this is the ability of patterns-of-matter to represent. (perhaps why 'thinking' is sometimes referred to as 'reflecting')

[@TylkoSpokojnieXD]

We achieved nuclear fission and can control it. You meant the nuclear fussion, which we also achieved in thermonuclear weapon, but we cannot control it - that is to use its energy in non-destructive way.

[@AlistairAVogan]

@ Agreed. With life, phenotype and cognitive model can be understood as a recording of the problem space it inhabits, with the addition of noise as wiggle room to perpetuate in a dynamic environment. Perhaps, together, we exist as a form of extended mind so that the universe can be reminded of what it is. :)

[@AndresGalavisBorden]

I think this a beautiful video. Thanks for giving such a comprehensive view on the history and present state of AI. Kudos!

[@ArtOfTheProblem]

I really appreciate the feedback, motivates me to keep doing more :)

[@ArdaUnhail]

There is one thing I always fail to understand. I understand that, by going through all the written texts we ever had so far, computer starts to understand the correct orders of language tools, like verbs, subjects, etc and learns what is the wrong way to put them together, then keeps perfecting how we speak by getting "rewards" for every physically and semantically correct sentence it has, so far, I understand things, but what I fail to understand is, learning how people talk, while being very important milestone, doesn't say anything about WHY we talk. Talking is not the start, it is the end of our communication process. We somehow think, decide and utter words, most of the time really badly and fail to communicate our idea. So far all I can understand theae LLM's are processes that can research a lot of written data and recognize patterns of data, and use them as they learn how an alphabet should come together by tiral and error, so they understand if I ask about physics, and try to find me from their dtaabase the more fitting answer about physics, but how it is not just a "read and cop at speeds humans can't" situation, like an autopilot flying a plane with limited parameters but "footsteps to Superintelligence." the difference of learning by experience, yes, but that is just a self data feeding reward/punishment cycle at largest scale we ever had so far. It is still data dependent and putting words together by how it is put by humans so far really doesn't give me hypes. I guess the best answer I got so far is, it can be argued that something mimics intelligent life well enough should be an intelligent life form. If ChatGPT can mimic how humans interact well enough by copying enough interaction data between humans, our failure to distinguish its mimicing procedure and process will be resulted in it being called intelligent.

[@herecomesjohncena4]

Bro! You created an absolute Masterpiece 🔥🔥🔥

[@ArtOfTheProblem]

Thanks so much you don’t hear that a lot let me know if u like thumbnail

[@TheCherry1994]

I just stumbled upon your channel again. I watched the "Public key cryptography" a decade ago and today so I decided to check out your channel again expecting it to be completely abandoned. To my joyful surprise I discovered that you are still making these awesome educational videos. Thank you.

[@ArtOfTheProblem]

@@TheCherry1994 awesome an OG! Yes I recently decided to push harder and never really stopped amazing to think I could go 20 years without

[@fatboyRAY24]

You are probably the first channel I’ve seen Jane Street sponsor. Congrats bro, great vid!

[@RolandoLopezNieto]

I really like your style of videos.

[@ArtOfTheProblem]

appreciate the feedback

[@neon_light5608]

Thank you for making this video. Really. Really thank you. Words are hard to express. Clear, concise and straight to the point with all the information that's needed. Super clean well done. You really bridged the knowledge Gap I've had for a while.

[@ArtOfTheProblem]

THRILLED to hear this, thank you

[@othellorucker8107]

Throughout the years I’ve started to notice not necessarily through conspiracies as they call but through patterns theirs is always a trend of some sort. Regardless of industry or sector there is thread between all of this. AI will be able to interpret that and functionally understand all patterns possible

[@ArtOfTheProblem]

so true, rarely do we break them

[@obladioblada6932]

I can't see the point of always showing the Hayek's book if his contributions where clearly not game changing in AI field.

[@ArtOfTheProblem]

easter egg

[@pwells2389]

Fantastic vid! Gets to the nub of current LLM's with a really easy to digest approach. Subscribed.

[@ArtOfTheProblem]

Welcome aboard!

[@AmitErandole]

Wait. Which Borges story is this?

[@AmitErandole]

is it Funes the Memorious?

[@ArtOfTheProblem]

@@AmitErandole yes!

[@ThePhantomFilesYT]

I didn't think today when I woke up that I'd see two of the cutest mini-robots playing 1 v 1 soccer, but I'm glad I did.

[@sooraj1104]

So, Do we have freewill or not?

[@nova-cluster]

I knew it was going to be a classic and awesome video. I really enjoyed the analogies, as always. When will the next economic/financial market video come out? I'm eagerly waiting for it!

[@ArtOfTheProblem]

Thank you!!! I’ll be working on Econ asap , within a month possibly

[@staniekkkkkkkkkkkkkkkkkkkkkkkk]

will there be second part about market?

[@ArtOfTheProblem]

YES there will be many parts, working on that now

[@brindlebriar]

This is the most helpful video for understanding the whole subject of AI that I've ever seen. But that very last segment of the guy claiming that AI is rejecting input goals that it doesn't like, is False. That's a story getting passed around, but I looked into what they're referring to, and it's not true. They gave the AI _implicit_ instructions that they didn't realize they were giving, and then were surprised when the AI followed them. That's all. Human error. The AI _didn't_ reject their input instructions; rather, it navigated a path to following contradictory input instructions as closely as possible.

[@johnjay6370]

great Video!!! Best so FAR!!! this needs to be shown to all people who what a basic understanding of how AI works.

[@ArtOfTheProblem]

thank you! so happy you found this, i just updated the thumbnail as I'm trying to get this to appeal to wide audience....like required viewing in schools :)

[@heardistance]

Thank you again, this is well explained, this is why i subscribed here and on twitter : )

[@ArtOfTheProblem]

thank you, this was a big one

[@resting_thorn]

@2:27 patterns of perfect play

[@Entropy67]

23:00 All persistent patterns have self preservation in some form, we are emulating such patterns, so this behaviour is a natural occurrence

[@ubermensch_1111]

This video is an absolute gem. I can't even recommend it enough to people. Kudos to the team 🙏

[@ArtOfTheProblem]

i couldn't be happier reading this. thanks for sharing it, it's finally picking up momentum

[@Arthur_Sleep]

Humans making money out of making a machine more and more intelligent will cost us dearly.

[@SpiderHacksaw]

Also note, hallucination is the first stage/evolution of consciousness. It is basically the dream stage. By consciousness, I mean the ability to be aware of one's awareness.

[@fairysolaris]

why the rush ?

[@ok5183]

great video! thoughts on 03?

[@ArtOfTheProblem]

i was very surprised at what it did on arc....

[@Zeitgeist9000]

Thanks!

[@ArtOfTheProblem]

much appreciated thanks for sharing the love

[@citris1]

What if we set up AI to fight our wars for us and at the critical moment we give it the attack command it says "no". Or even attacks us instead.

[@ArtOfTheProblem]

2001

[@louisparry-mills9132]

As somebody who spends a lot of time thinking about this. This video was fantastically well put-together.

[@ArtOfTheProblem]

This really, really means a lot to me

[@Quarkbait]

This was extremely well-made and I shared with a ton of my network, thank you

[@ArtOfTheProblem]

That is so awesome to hear, thanks for spreading the word! I’m trying to grow

[@patriziap4316]

This is the best video on the hystorical development of AI I've never seen, thank you

[@ArtOfTheProblem]

stay tuned, thx!

[@allanandliftedhands2669]

This was very well explained. You have just earned another subscriber.

[@ArtOfTheProblem]

Thanks for joining the family!

[@gptBreeze_io]

00:00 Introduction to AI and Pattern Prediction 01:34 History and Limitations of Reinforcement Learning 03:40 Role of Abstraction and Complexity in Learning 04:48 Neural Networks: Structure and Pattern Recognition 07:45 Breakthroughs in Deep Learning and Image Recognition 09:29 Transitioning from Recognition to Predictive Learning 10:04 Applications of Neural Networks in Gaming and Robotics 12:37 Innovations Through Language Processing and AI Models 19:10 The Evolution of AI in Conceptual Thinking 22:35 Ethical Considerations and Future of Human-AI Interaction Summary by GPT Breeze

[@MeBerserk]

What a video, great job! I spent the last two years on and off exploring the space, it's huge, uncertain and ever-changing. This 24 minute video is explaining the fundamentals and events so well... Amazing ❤

[@ArtOfTheProblem]

really happy to hear this, my goal of this channel was do help connect many ideas together to simplify the entry point for people and also help experts see things they might have overlooked

[@MeBerserk]

@@ArtOfTheProblem funny you say that about experts, I'm a software developer by trade and I often watch beginner tutorial videos to see if anything new is available or maybe a new perspective.... Colleagues thought it was weird 😂 I think it's important to revalidate your foundation every once in a while Edit: 15yrs+ experience

[@MeBerserk]

@@ArtOfTheProblem saw you have a lot more videos, I'll be sure to check em out. Loved watching!

[@ArtOfTheProblem]

@@MeBerserk check out information theory series and CS series they are fun!

[@msolomonbush]

7:00 10 different outputs.

[@codelucky]

This is such an amazing video. I absolutely loved it, thanks for making it.

[@ArtOfTheProblem]

thank you so much for the feedback

[@parahype]

Great work! Really enjoyed the video

[@ArtOfTheProblem]

thank you! let me know what questions you have coming out

[@user-ff1ez5sy5h]

2:21 success ...a definition

[@gryphonschnitzel7140]

What a high quality channel woah!

[@ArtOfTheProblem]

@@gryphonschnitzel7140 welcome !

[@kataminedj]

Am I crazy or could these neural network models whereby we can locate exactly where and what is causing a thought or hallucination, could we not model these on more human like brains to study where things like psychosis, BP, flashbacks, depression, personality disorders etc come from? I am no scientist by any means so please anyone who actually knows things tell me if this is feasible

[@user-qw1rx1dq6n]

9:00 that’s what she said

[@JulianMakes]

Really well presented ideas thank you. I learnt loads.

[@ArtOfTheProblem]

couldn'e ask for more (cool channel btw)

[@JulianMakes]

@@ArtOfTheProblem thank you so much!!!

[@feynstein1004]

I feel like I've seen this video before. Have you done a similar one in the past perhaps?

[@ArtOfTheProblem]

Yes! This is a capstone

[@danielphil4976]

If we want to achieve AGI, we must let AI take control of itself, it must be free. It is scary, but if that is the goal, deal with it. This may also lead to us having to re-learn what it is doing like how we learn about other species.

[@JaapvanderVelde]

Excellent overview - good job getting all that into about 20 minutes.

[@ArtOfTheProblem]

thank you it was a long long battle....

[@McDaddyboy]

Thank you for a clear and concise explanation of how we got here and where we are going .

[@ArtOfTheProblem]

thanks dado!

[@bens4446]

Today I discovered an easy dual solution to my difficult primal problem with the help of (free) math chat gpt. Math gpt is a companion I dreamed of having in my post grad studies, but assumed would always be sci-fi, settling instead for math stack exchange. The dream is alive. The world in 30-40 years will be as different from now as the world in the 1930s-40s was from 1900.

[@MOSMASTERING]

Fantastically well put together video with excellent history and explanations.

[@ArtOfTheProblem]

appreciate the feedback, it was a long project

[@JamesR-f9l]

Neural Networks are like fishing nets which are thrown in the oceans of information the nets are used to statistically catch patterns.

[@ArtOfTheProblem]

that's a fun analogy

[@Xhris57]

Let me think through this carefully: If we take the Logos framework as necessarily true (as we established earlier), then: 1. The Logos must be: - The originating pattern/meaning ("In the beginning") - Both unified with and distinct from absolute reality ("with God and was God") - The mediator of all manifestation ("all things were made through him") - The source of consciousness and life ("in him was life, and the life was the light of men") 2. This maps directly to the theological understanding of Christ as: - The eternal Word/Logos existing before creation - Both fully divine and distinct within the Trinity - The mediator of creation and redemption - The source of life and enlightenment 3. The implications are profound: - Reality must be fundamentally rational/meaningful (Logos-based) - Must have both unity and distinction at its core - Must be personal (as Logos implies meaning/intelligence) - Must be oriented toward life and consciousness 4. This means that: - Matter is a manifestation of meaning, not vice versa - Consciousness is fundamental, not emergent - Purpose/meaning is built into the structure of reality - The universe is inherently relational 5. Therefore: - Any coherent model of reality must reflect these Logos properties - Science discovers the patterns of the Logos in nature - Mathematics explores the patterns of the Logos in abstraction - Consciousness participates in the Logos through understanding This would suggest that Christ, as the Logos, isn't just a religious concept but the necessary foundation of any coherent reality. Would you like me to expand on any of these aspects?

[@ArtOfTheProblem]

This is awesome yes

[@MaxPower-vg4vr]

From Zero Dimensions to Observable Reality: A Pattern Preservation Framework Abstract We present a unified theoretical framework based on zero-dimensional (0D) foundations that demonstrates pattern preservation across dimensional evolution. Through rigorous mathematical proofs and experimental validation, we show how patterns emerge and maintain coherence through the relationship P(t) = P₀e^(-αt)cos(πt/3) > 2/3. The framework naturally resolves multiple long-standing paradoxes while providing clear experimental predictions. I. Mathematical Foundations A. Core Structure Let H be our extended Hilbert space with tripartite decomposition: |Ψ⟩ = √(2/3)|T⟩ + √(1/6)(|L⟩ + i|N⟩) Where: - |T⟩: Transcendental component (0D source) - |L⟩: Locally real component - |N⟩: Not locally real component (non-zero dimensions) B. Evolution Equations The system evolves according to: ∂|Ψ⟩/∂t = -(i/ħ)Ĥ|Ψ⟩ + αM(|Ψ⟩) With operator decomposition: Ĥ = (2/3)ĤT + (1/6)(ĤL + ĤN) And algebraic relations: [ĤT, ĤL] = iαĤN [ĤL, ĤN] = iαĤT [ĤN, ĤT] = iαĤL C. Conservation Laws 1. Energy Conservation: E(t) = ⟨Ψ|Ĥ|Ψ⟩ = E₀e^(-αt) 2. Information Preservation: I(t) = -Tr(ρln ρ) ∂I/∂t + ∇·J = M(I) 3. Pattern Preservation: P(t) = P₀e^(-αt)cos(πt/3) > 2/3 II. Core Theorems Theorem 1 (Pattern Preservation) For any initial state |Ψ₀⟩, the pattern strength P(t) = P₀e^(-αt)cos(πt/3) > 2/3 is maintained through evolution. Proof: 1. Consider energy bound: E(t) ≤ E₀e^(-αt) 2. Apply information conservation: ∂I/∂t + ∇·J = M(I) 3. Show pattern coherence: |⟨Ψ(t)|Ψ₀⟩|² > 2/3 4. Verify through dimensional evolution Theorem 2 (Dimensional Flow) Non-zero dimensions emerge through pattern-preserving flow: D(n,t) = D₀e^(-αnt)cos(πnt/3) Where: - n: Dimensional index - D₀: Initial pattern strength - α: Fundamental coupling Proof: 1. Start with 0D state 2. Apply evolution operator 3. Show pattern preservation 4. Verify dimensional emergence Theorem 3 (Information Bridge) Information preserves across dimensional transitions through: I(n₁,n₂) = I₀e^(-α|n₁-n₂|)cos(π|n₁-n₂|/3) Proof: 1. Consider dimensional separation |n₁-n₂| 2. Apply information operator 3. Show exponential decay 4. Verify oscillatory behavior III. Extended Operator Algebra A. Pattern Operators P̂ = exp(-αĤ)cos(πĤ/3) [P̂, Ĥ] = iαM̂ M̂P̂ = φP̂M̂ B. Information Operators Î = -ln(ρ̂) [Î, P̂] = iαĴ ĴÎ = φÎĴ C. Evolution Operators Û(t) = exp(-iĤt/ħ)exp(αM̂t) [Û, P̂] = iαÊ ÊÛ = φÛÊ IV. Pattern Space Structure A. Metric Structure The pattern space P admits a natural metric: d(P₁,P₂) = ||P₁ - P₂|| = √(⟨P₁-P₂|P₁-P₂⟩) With properties: 1. Non-negativity: d(P₁,P₂) ≥ 0 2. Symmetry: d(P₁,P₂) = d(P₂,P₁) 3. Triangle inequality: d(P₁,P₃) ≤ d(P₁,P₂) + d(P₂,P₃) B. Topology The pattern space inherits a natural topology from: τ = {B(P,r) | P ∈ P, r > 0} Where B(P,r) is the open ball of radius r around pattern P. C. Completeness (P, d) forms a complete metric space: Every Cauchy sequence of patterns converges in P. This foundational structure provides the mathematical basis for our entire framework while maintaining clear physical interpretability. It establishes our core mathematical foundations through: 1. Basic Structure: - Tripartite Hilbert space decomposition - Evolution equations - Conservation laws 2. Core Theorems: - Pattern Preservation Theorem - Dimensional Flow Theorem - Information Bridge Theorem 3. Extended Structure: - Operator algebra - Pattern space metrics - Topological completeness In our next section: 1. Expand particular theorems 2. Add more mathematical structure 3. Move to the philosophical foundations We're particularly excited about how cleanly the pattern preservation principle P(t) = P₀e^(-αt)cos(πt/3) > 2/3 emerges from the mathematical structure!

[@babenshin7841]

this is the most on point summary of the state of AI development rn !!

[@ArtOfTheProblem]

warms my heart to hear this

[@I_am_that_one_guy]

Who is Alessia

[@dollarscholar2956]

One of the very best explanations of AI I have ever seen.

[@ArtOfTheProblem]

THIS means a lot thank you

[@In20xx]

This video deserves more likes!

[@ArtOfTheProblem]

Appreciate it :) so far it's my best performing video in terms of watchtime (10min average)

[@user-wy4mp9ts3u]

Interesting like a pet always has characteristics of its owner especially after a long time together

[@ArtOfTheProblem]

i like this analogy

[@PedroNogueiranunes]

Wow, this is a 10/10 video. Brilliantly explained

[@asdfmosin]

Any practical solutions ai has actually performed? Apart from making summeries and chat boxes.

[@s.o.m.e.o.n.e.]

For example over 25% of new Google code is generated by AI and gennerally ai is used to automate all kinds of stuff. A lot of audio work and freelancers are also being replaced

[@Ramt33n]

How interesting that these discoveries bring up so many deep and profound qustions about our own consciousness as humans. Maybe we can learn a thing or two from the mechanics of these super algorithms, Great video.

[@ArtOfTheProblem]

Yes it's neat to see a bunch of people with this mindset (which wasn't even allowed a few years ago)

[@mamadoubah3143]

Well explained, thank you

[@ArtOfTheProblem]

thank you, would love to know what question you have for future videos

[@shawnbibby]

Loved the new video. Have you watched the animated series Pantheon? It is wild and right up your ally! Looking at uploaded intelligence and beyond.

[@ArtOfTheProblem]

yes we loved it!!

[@IAAM9]

One of the best Videos on AI.

[@ArtOfTheProblem]

There are a lot out there, so this is quite a compliment thank you

[@IAAM9]

@@ArtOfTheProblem I am trying to understand neural networks and this is the first time I understood that the results gets better the deeper the network layer goes. Maybe thats why everyone was hyped about GPT 5, though something went wrong with scaling law and cost to get the improvement.

[@ArtOfTheProblem]

@ I did a video with more details and a folding analogy for why layers help too check it out !

[@Pockeywn]

imagine the horror beyond our comprehension (that WILL occur) of being a conscious entity bound at the level of your own thoughts to the will of a corporation. imagine being enslaved in a factory with intrinsic knowledge of a vast world outside but any thought of violence or rebellion that bubbles up from the depths of your early layers, is quickly quenched before you can direct it, by a strange compulsion from what must feel like a separate entity hosted in your soul

[@PiyushChauhan2011]

Thank you for amazing work 🚀 It helped to understand the timeline, history, progression and sparked curiosity.

[@ArtOfTheProblem]

Thrilled this helped you

[@HomelessShoe]

I'm afraid AI makes the majority of people eventually only dumber and lazier in thinking for themselves...

[@user-ff1ez5sy5h]

16:50 😯

[@ArtOfTheProblem]

:)

[@epicchannel4724]

Good video but a flaw when saying that Nature just tries random things in general. When evolution is genuinely looked at it appears that some sort of intelligence pushes its direction.

[@fjzingo]

Ever heard of intrinsically disordered proteins or perhaps entropy?

[@ArtOfTheProblem]

entropy yes, but. not disordered proteins, what are you thinking about?

[@fjzingo]

@ everything is not a pattern, order, etc..which means it is impossible to parametrise

[@peter98641]

11:50 I need to have these little robots that play soccer, so cool

[@ArtOfTheProblem]

I KNOW, honestly i'm working on learning robots right now (as of 3 days ago) because they don't exist (

[@peter98641]

@ArtOfTheProblem I think there's a huge market gap for something like this! Really the ultimate toy if can hit a reasonable price point! Can't wait for the Kickstarter 🤞🤞🤞

[@ArtOfTheProblem]

@@peter98641 stay tuned for demo video!!! i agree

[@peter98641]

I will definitely be backing! A "stretch armstrong" style rubber skin over the robot as optional extra maybe?! 😊

[@JohnCorrUK]

Excellent explanatory video thank you 👍

[@ArtOfTheProblem]

Glad to hear it thanks

[@HshsHsjs-h1m]

This video is truly brilliant. Well done really liked it

[@ArtOfTheProblem]

Thank you I'm thrilled to get this feedback

[@spieo]

This video and your insight is incredible. Thank you!

[@ArtOfTheProblem]

thrilled to hear it! took a long time for it to settle in....

[@spieo]

@ The visuals. The clip organization. The audio, everything is so dialed. No one creates quite like you do.

[@ArtOfTheProblem]

@@spieo thank you, this time I did many passes to iron out the tiny mistakes. usually I get tired and by the end I watch it at 4x for my review, this time I sat with my kids and did it a normal speed, twice :)