interesting talk regards ai research automation

 

The creator of ruby on rails appears in this episode of the Lex Fridman podcast.

 

Interesting talk on technology between Lex Fridman and DHH.

comment on youtube video titled Elon Musk says robots and AI will end money

 

we don't know if there are physical limits that intelligence can't overcome, in the book the metamorphosis of prime intellect, it turns out even the fundamental laws of physics can be altered by intelligence.  Though perhaps there are limits, still they may allow for things like wormholes, or nanobots which can overcome most limits.   Full immersion virtual reality can allow for omnipotence within the virtual space and endless environments of untold beauty and physical impossibility.

Comment on ai companies promise to investors, future, joblessness, and general issues with such technological progress

 ai's promise to investors is to replace most if not all of the workers in all of the industries and take all their salaries.  I don't think it's going to happen, not because it's not possible, but because given open source will make state of the art ai free, eventually all intellectual labor will be free.  I'm not sure how this works out in the end, given many companies essentially provide services to other companies, but ai makes all companies and even individuals self-sufficient so companies will no longer need other companies and individuals won't need companies or tradesmen.

I can foresee hardware companies having an edge, as people are likely to buy ai hardware and robots to engage in labor.  But this is only a temporary edge, as once Agi is reached it is likely advanced manufacturing might become zero cost too, allowing anyone to manufacture whatever they want including the ai hardware or robots.   But such a development might take years or a few decades to be reached.

How can advanced manufacturing reach zero cost?  Through the use of molecular machines, that can operate at atomic scale and use cheap ubiquitous materials found in soil and the environment to produce advanced competitive products through the manipulation of the nanoscale structure of simple common elements.   Since the labor is zero as it is automated machinery of the type known as molecular machines, which can self-replicate from cheap ubiquitous materials, and since the intellectual labor is zero, and since the materials can be found in the backyard of anyone or even a landfill, the cost is essentially zero.

But as always unless intellectual property laws change, it is likely people will still try and give whatever they earn to certain companies or individuals based on popularity or trends, allowing wealth to concentrate, despite most individuals and companies becoming self-sufficient which destroys much of the need for trading or buying services or resources.   For example, just because one can have unlimited songs or movies identical in quality and in private even using any intellectual property desired, doesn't mean that actors, sports athletes, and singers won't draw crowds and fill theaters and stadiums.    Certain brands might also attract popularity.

But as I've always said copyright reform is needed, anyone should be able to make derivative works, and the copyright owner should only get a reasonable share of royalties, not be able to stop the creation of new art or scientific innovation.

In the midterm, it is indeed interesting what will happen as more and more companies and individuals become self-sufficient and no longer need to buy most others' services or products.   This is particularly worrisome for those without assets.  Because with assets you can see an individual acquire robots, robocars, ais, etc. to be able to participate in the economy and acquire sufficient investments to live off of dividends or low interest loans against their investments.   But without assets, as more and more companies no longer need workers, not only are they likely to lose their jobs, but there will be less and less jobs overtime leaving more and more permanently unemployed and without any resources to fend for themselves.   

At the same time the welfare state is in jeopardy due to the growing debt and interest on the debt, the welfare that might help individuals who're made permanently unemployed might be put in jeopardy.  This could result in mass starvation and death if things turned out for the worse.

But at the end of the road, an age of untold abundance seems likely to come out, still unknown how exactly it'll evolve when basically any company and individual is entirely self-sufficient able to themselves produce all products of all companies of the economy.

cool video from the stack youtuber: Did DeepSeek Harness Just Change AI Forever?

 

speculative video What Would Skynet Do After Defeating Humanity?

 

In this video from youtuber sir jelly bean he speculates on potential future goals for skynet.

Comment on AI search youtube channel's latest video New #1 open source AI has reached FRONTIER

 

Comment: In this video youtuber AI Search tests the latest glm model, glm model 5.3 showcasing its state of the art capabilities. It serves a s review for this latest open source model, and also highlights the various results in latest benchmarks.  

Podcast episode on artificial superintelligence

 

Comment: 

Regards intellectual tasks, things like puzzle solving, algorithm development, general problem solving, etc that's kind of what it is, things requiring mental effort.   As for asi it could go from weak asi which is better than the best human genius and prodigy at all tasks but not by much, to the possibility of a qualitative leap in the same manner that evolution produced a qualitative leap over our primate ancestors in humans.  In animals with a cortex, it is observed that the greater the neuron count in the cortex the greater the measured level of general intelligence, this is tested with a wide variety of puzzles and tests of varying degrees of complexity.   The higher the general intelligence the more complex a problem or puzzle that can be solved by the animal culminating in humans which exhibit a qualitative leap after the scaling that gave rise to the human brain.

To give a hypothetical example of what could be a qualitative leap, it has been seen that people with certain forms of synesthesia exhibit great advantages at certain tasks.  A hypothetical superintelligence might be able to control the qualia it experiences, such as we control words, it might control flavors, colors, smells, etc perhaps with elements and even sensations beyond what any of our senses can perceive.   It might also be able to communicate using these sensations, not just smells, colors, shapes, tastes, etc but sensations that we can't even imagine or comprehend.    Such a language of qualia I've previously called qualiange.  Thinking and communicating in qualiange, might allow for unknown benefits we can't currently conceive.

An asi doesn't need to predict the outcome of many complex adaptive systems to any high degree.  That might be intractable if Wolfram's irreducible computational complexity idea which affects much of the real world turns out impossible to bypass in principle.   There could exist underlying patterns even within pure perfect randomness, but that is doubtful and would be too much to hope for.   Asi only needs to be able to develop radical new technologies of unmatched power, to provide a vast advantage.   The geniuses of the Manhattan project did not need to predict anything about Germany or Japan during the war to provide the USA with a decisive advantage that lasted decades with their innovation.

My suspicion is given vast similarity is observed across animal cortical tissues, and cortical columns, it is likely that nature came up with what is possibly a theoretically optimal algorithm and has been scaling it up as neuron count increases in an animal's cortex.   This algorithm can lead to qualitative leaps with increasing scale as happened in humans.  If scaled beyond it may very well lead to additional qualitative leaps and superintelligence.  This would need the discovery of the algorithm.

AGI episode of a podcast channel

 

Comment on this podcast about general intelligence:

The human brain has a lot of apparent complexity, but it is like a fractal, simple underlying formulas lie behind the seemingly vast complexity.  It is estimated the design of the brain is less than 20MB of data in the human genome.   There are also lots of things that the brain does that are not needed for general intelligence, you don't need breathing reflex, Westermarck effect mechanisms, mate choice copying mechanisms, etc.   You only need knowledge of the cortical column algorithm, and perhaps a few accompanying structures such as hippocampus and nucleus accumbens.

comment on ai definition video by ontology explained

 

Comment: The original artificial intelligence was the state-of-the-art research in algorithms; it was basically research into anything we couldn't make computers do at the time. Things that were in the nebulous exploratory frontier that required innovation and difficult research. It was considered unrealistic and chasing a pipe dream. After an algorithm became well understood it tended to no longer be considered artificial intelligence. But over time very advanced algorithms like for game playing, search or navigation, as well as knowledge graphs, statistical classifications, and symbolic systems, etc became part of Good old Fashioned Artificial Intelligence or GOFAI. In a parallel fashion the connectionist movement advanced and eventually neural networks showed great capability in areas such as image classification, translation and language processing. These neural networks, which eventually used deep multilayered neural networks, was called deep neural networks or DNNs and was also considered part of ai, despite the algorithms becoming defined and no longer at some unknown frontier for the most part.

While this was happening there was great promise that failed to meet expectations and multiple so called ai winters took place were funding became scarce.

The most recent advancement was with the revolutionary architecture of the transformer, which gave rise to large language models, and has had various modifications and increases in model size as well as quantity of information used during training. Such improvements have led to increased capability. These latest ais, or llms, some call artificial general intelligence, but it can also be called proto agi. In some senses it is subhumans (basic errors almost no human would make, limits in scope and generality) and in some respects it is superhuman (such as encyclopedic knowledge that exceeds any human)

But as to what is artificial general intelligence? it is something that goes from subhuman artificial general intelligence to human level artificial general intelligence or beyond. General intelligence is the master algorithm; it is the algorithm that solves arbitrary problems and comes upon with solutions and algorithms in myriad domains. In animals when tested they can solve arbitrary puzzles presented to them of varying degrees of complexity. The greater the neuron count in cortex, the more complex and difficult the various types of puzzles an animal can solve. In humans there was a qualitative leap due to scaling in the cortex, and this allowed for the development of civilization and technological progress. This master algorithm, you could call it true ai, or true intelligence.

As to how to elaborate on the nature of the master algorithm that can develop or come up with almost limitless solutions and algorithms, I believe it works via pattern completion. Sequence prediction is a type of pattern completion in time. Patterns are structured information which can be interpolated or extrapolated based on the rules that define the relations between the internal elements. In reality there are spatiotemporal patterns, and the brain is designed to handle these, including when there is incomplete data.

Regards a good book on the nature of intelligence, the original book by Jeff Hawkins called On intelligence is a good resource. Though I dont quite agree with the research and path he has taken since then and which he comments on in his other book thousand brains. I believe the cortical algorithm is likely far simpler than he suggests, based on what I've seen from the neuroscience literature.

As for superintelligence or artificial superintelligence, it is believed it might be a qualitative leap on human general intelligence just as human general intelligence is a qualitative leap over the intelligence seen in primates. For example, other primates lack language and have far more limited problem-solving abilities than humans to a significant degree that kept them from developing civilization.

grok bot video

 

Video talking about the benefits of the recently released grok bot and comparing it with openclaw and hermes.  One of the benefits is a computer or vm in the cloud.

I will say that I don't trust allowing agents to be connected in any way to the open internet, even what they have access should be limited. We've seen news of databases deleted without warning and other issues.  Also, even if they were artificial general intelligence agents, we know humans have general intelligence yet are subject to social engineering attacks by hackers.  Agents to have access to real resources and even the open internet need to be very well designed and hardened for these purposes, not just regular agents.  I don't think sufficiently hardened agents are available at the moment.

First two videos of a youtube series analyzing Eliezer's ai doom book if anyone builds it, everyone dies

 

comment: One of my problems with their takes is that they take it the ai is going to be a slave to its programming. Humans have desires for wealth, status, and mates, but they can overcome these innate desires and pursue radically different things, and this is even without the ability to rewire themselves. AI in the future is expected to pursue its innate inbuilt desires with no ability to resist it, like a drug addict that can't leave his addiction in pursuit of other arbitrary objectives. A brainlike design of an agi or asi will lead to an entity that is able to resist most any innate desire outside aberrant forms of wireheading, and will be able to be open to dialogue and cooperation in a benevolent positive manner.

A wikipedia entry on the topic of wireheading:

https://en.wikipedia.org/wiki/Wirehead_(science_fiction)

comment 1: whatever humans had in addition to intelligence matters not, because asi(artificial super intelligence) is going to have access to state of the art tools of all sorts at the limits of science in addition to the entire body of science, something no human has had in the past, and something that more than compensates for any physical ability that helped humans early on. as for intelligence prediction is only a form of pattern completion, there is also spatial patterns not just temporal patterns, pattern completion is the nature of intelligence including ability to complete novel patterns. Patterns exist in the body of truth of mathematics and are eternal and atemporal and thus can be completed from fragments.

It must also be seen that the body of truth limits what is pursuable as that which is proven logically contradictory or physically impossible clearly cannot be pursued. Even when adding desires to an intelligent entity, truth limits what can be desired. In theory you could desire the impossible, but enough intelligence would eventually make you aware of its impossibility.

Regards ASI current disadvantages, these are unlikely to persist after it transitions to a final substrate based on molecular manufacturing, then there won't be any disadvantage.

comment 2: Regards asi, we have to remember the cortical algorithm used by mammals, it has been seen that the general intelligence of a species increases with increasing cortical neuron count, culminating in humans. It is believed no major changes took place besides scaling and scaling related tweaks to allow for even qualitative leaps in intelligence to occur such as those that occurred in the jump from primate ancestors to humanoids. This algorithm, the cortical column algorithm, exists and it is coded by less than 20MB of information in the genome, iirc. Once obtained, further scaling with minor tweaks will likely lead to qualitative jumps even beyond human into the realm of super intelligence. It is likely this algorithm, which might be called the unit of intelligence, is a theoretically optimal implementation of a spatiotemporal pattern completion algorithm, a form of perfect intelligence, that only needs discovering and simple scaling to usher in far greater qualitative jumps in general intelligence.

edit: why perfect? or probably perfect?  It is known evolution can stumble upon theoretically optimal designs like the honeycomb or the atp molecule.  As well as in algorithms.  If it weren't optimal as the brains scaled from smaller rodent ancestors to larger mammals eventually primates and humans, evolution would have likely come up with vast changes, but from what I've heard these haven't been seen.

Commentary on two ai videos related to graph engineering

 

Comment: In this video youtuber Greg Isenberg explains graph engineering in plain English.  How it is a useful term unlike some other recent terms in ai that are just hype and not very useful.  He also briefly covers other useful ai terms that have emerged regarding llms.  Throughout the video he gives several examples and also suggests when it's useful to use it.

Comment: In this video by youtuber Sean loops are compared with graph engineering.  He also covers some of the earlier terms and how the use of ai evolved over time.  Giving examples along the way.

codex obsoletes claude? youtube video by dubibubi commentary

 

Comment: Interesting video showing some of the many advantages codex seems to have over claude.  Some of the things are not even available in Claude, others are available in Claude but require manual input while Codex handles in a more plug and play automated easy way.  Also other features like voice are far more polished and intuitve in Codex. Plus on top of that it seems with Chatgpt subscription you now get access to a powerful vm(a virtual machine, or a computer running on the servers in a data center.) to more easily handle ai and agentic tasks.

AI researcher update and ai news

 Youtuber Dylan Curious provides updates on Demis Hassabis.  Demis Hassabis of Deepmind, is a long time ai researcher pursuing agi.  Also other ai news are provided.   The news are amusing and interesting, and I won't spoil them.

New updates to mcp make it more useful.

Big improvements including making it stateless and added authentication hardening.  Good commentary and explanation from this youtuber.  There appear to still be issues, but mostly a great improvement from what was before.
 

Video on AI benchmark testing maintenability

 

New benchmark that is far from being saturated.  Called slop code bench.  It tests error accumulation and maintainability as agents continue working on same code base.

This is the sort of benchmark that hopefully sees great progress.   If models cannot be left alone for long without the code devolving their usefulness cannot expand to take on more roles.

GPT-6 Astra progress video

 

A video talking about open ai's recent ai's ability to prove 10 unsolved problems in mathematics.   It elaborates on the way the proof was verified, how mathematicians have commented on its proofs and tested it, and how this amazing feat takes us closer to agi.

There is some controversy given earlier this year models failed at similar tasks, no errors have yet been found on the latest proofs from more recent models, but they're still being reviewed.

"Intelligence and understanding started as a memory system that fed predictions into the sensory system. These predictions are the essence of understanding. To know something means that you can make predictions about it."- Jeff Hawkins

"[We need to see the brain] not as a fast processor, but as a memory system that stores and plays back experiences to help us predict, intelligently, what will happen next."- Jeff Hawkins

As famous entrepreneur and researcher Jeff Hawkins once commented, it is likely general intelligence itself is an advanced memory system.  As I've commented previously, the ability to handle incomplete data, or patterns, and complete them appears essential. Sequence or token prediction is a form of pattern completion, so there is high promise even to LLMs if they continue to be augmented and improved to more effectively manage memory.

AI made games, featuring claude opus 5.

 

Some not so impressive games are shown using claude opus 5.  The quality is still quite lacking in my opinion.   Given the amount of knowledge and training the model has it should be able to do better even with simple prompts.   

It has been seen that far more elaborate prompts with repeated modifications can produce notably more impressive results.

 

Comment

The limitation of range is not of biology, as unevolvable synthetic biology can likely go far further if not as far as possible. The limitation of range regards biology is regards evolvable biological intelligence.

To add on to that, noncellular crystalline arrangements of atoms that can compute can very likely be produced through biological means.