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: 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.
comment: Youtube channel Learn Meta-Analysis goes over first impressions the unsloth studio ui. There are many benefits to the ui, and according to the youtuber it can do video, images, and text, iirc. Of course that depends on the loaded model. Can also work with web based models not just local models.
The local ai video generator can run even with less than 16GB of video ram, but for full precision needs over 60GB of video ram. Highest quality yet, and with easy setup through the popular comfy ui.
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.
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.
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.
Molecule that enzymatically reverses glycation damage has been tested and been successful in the lab. In the near future might become available after human testing. The video mentions potential interventions that might reduce damage in the mean time while waiting for this to become available.
