Humans are far worse than LLMs in many ways: a brief analysis

Comment on: 

wendywee — 9.7.2025 г. 6:41 Discord, MLST, #selfpromotion

While it is generally true that "bigger is not always better...", the "scaling question" was asked back in 2001 (Man and Thinking Machine... - will more transistors, computing power etc. lead automatically to "thinking" etc.) and I myself am a proponent of and work on more conceptual ("symbolic") causal approaches, which don't need to read the whole Internet, it is more complex than that:...

 Tosh — 21.7.2025 г. 23:01

The need to think with world-models (hierarchical simulators of virtual universes at different resolutions of causality control and methods for segmenting/granulating the sensory-motor data etc.) was defined in AGI in the early 2000s, and then rediscovered in the mainstream for the LLMs. However the scaling of LLMs from a long time is not just GPT1-GPT2-"Megatron" -GPT3, it's more than just adding more data. It got hyper multimodal, ensemble models, it includes robot data and physical interactions, RL, RLHL etc. Also the big platforms are unlikely just a single LLM, they are probably multi-component systems, which can include whatever "tool-use", simulators (another branch: Simulation Intelligence), physics-informed deep learning, game engines (=world models). Some foundation models from the last 2-3 years became also "generalist Agents" and they already do have practical "world models" to different precision, span etc.; JEPA-2 as well (although 1 Million hours video is way too much and it is still "dumb" ). 

"World model" is ambiguous, in the classical theories 20-some years ago there was the concept of "resolution of causality control" and the good LLMs actually *do"  have an implicit "world model" with linguistic resolution for many cases. Not perfect, not complete etc. "World modelling", physics, simulations is about predicting what will come next, given particular known preconditions and consequences, that's the causality, i.e. the mechanism is the same, but with LLMs it is not always predicting from the correct premises and don't follow the right laws/reason.

The late popular meaning of AGI about "economically feasible tasks", "better than the best human" (in these tasks),  was promoted by the business opportunists. Originally AGI was about universal learning, prediction and compression in all modalities with a  general algorithm ("next token prediction" was predicted by the AGI); i.e. in many ways this already happened. 

--- END OF QUOTE ---

1. Humans demonstrate the same faults like LLMs, but worse

The LLMs are "vulgarized" and some of the faults of the models are the same in human capabilities and behavior, however the humans don't understand them or they lack integrity in order to admit it.

2. Phonological loop of 2-3 seconds - conscious back context window of just several words

A few humans can remember more than 2-3 seconds in the past in their "phonological loop" - listen to a podcast or video clip, interrupt it and write down what you heard in the previous seconds; on which word you start to hesitate "was it this or that word?... was it a car or a vehicle?"? When you start to lose the correct word order? On which word you completely lose the thread and remember only that "you spoke about processors, the party, the movie, ...".*

The persons who can remember slightly more are yet unlikely to remember perhaps even just 10 or 15  seconds or these ones would be rare "freaks".

So how big is your "context window" then, counted in tokens comparable to LLM's? This is for the past, where the LLMs and computers can remember not just thousands, but even EVERYTHING, billions or trillions of tokens exactly (as memory).

Let's look forward now.

3. Next token prediction for humans - context window for future tokens is even worse

How many "tokens" *ahead* can you predict correctly in your *own* utterances, in plain text and words? How well can you plan your sentences when you write on paper or on the computer?

One word, two words, rarely 3-4-several words?  Can you remember in your working memory a long complex sentence of 15-20-25 words? So it seems that you're a 3-4 gram model. Besides, unlike the LLMs, which have the "log probabilities" and can perform various kinds of "sampling" or "beam search" and sample differently, you usually don't or it would be too hard for you and too slow to write down many variants, so you wouldn't generate even two or three possible variants of your own "next word prediction task", except if you are a writer, you compose a poem, a story etc.

I myself do this, but it takes time: you first generate an initial version of a verse or a poem, then you replay it many times and let the time pass, and in minutes, hours, days you start to find better thoughts and words, ideas which better suit the overall message etc.**

* Perhaps the working memory capacity of 7+-2 items is a common prerequisite for the limitations of this "loop" and "buffer lengths".

4. Hallucinations - intentional lies and cheating and sometimes inability to even hallucinate

Humans not only "hallucinate", or make mistakes - say things about which they lack definite information, or they lack reliable proofs that it is true (or they lack information that makes them believe that the information is true - it can be wrong either), and they are convinced and aggressive in their believe they are correct, and try to convince the others often based on authority and power, they are "bosses", they are "on the right side of the history", they are "the good guys" (in Hollywood movies, LOL).

Humans intentionally and aggressively LIE and try to CHEAT and suggest lies in order to take advantage of someone or the situation and they don't consider this as a wrong behavior if it is in their favor (LLMs learn from humans either, LOL).

Even worse: in many ways and often humans can't even hallucinate, when "prompted", or their attempts are grossly ridiculous, i.e. they have no clue what to say, what the next word would or could be, even randomly; or they cannot generate any coherent continuation of a few words with correct syntax, guided by a topic or more material [Edit+ 1.12.2025: also, most of them don't even understand what is "syntax", certainly not in a deep "meta-linguistic" way - how many humans can do "POS tagging" and disambiguation?)] 

This is true especially if the "HumanLLM" lacks appropriate "pretraining" in the domain. Ask somebody what he thinks about an area of knowledge or event where he has little or no expertise, or in language she doesn't speak. What she would answer? What if you show a "humanLLM" who has never programmed a computer [and hasn't read any programming code] a piece of machine code in hexadecimal  A9 7F 8D 00 20 ... What "reasonable" content would the human generate if she is obliged to continue and return 100, 200, 1000 tokens [and what would be his reaction? = What the h* is that?] 

5. Memorizing and Repetitions of what was already known, "just retrieval"

LLMs are blamed for "just retrieving", repeating existing knowledge, just recombining it.
What else you do most of the time, humans? You can't do even that!

People with slightly better memory than the rest are called "computers" or compared to "computers", and that used to be even in the 1950s, LOL, or these humans have "prodigious memory", if they can remember 50 or 1000 "digits of the number pi" (a meaningless record), because normally humans have poor memory and remember up to 5-6-7-8... digits of Pi (3.1415926...). Even the most intelligent ones need to constantly consult dictionaries, textbooks, books, mathematical tables with formulas, to lookup the code they write or repetitively reread the paragraphs of the posts and articles they write, to dig into the Internet, ... and to rely on LLMs to do the job for them.

In general almost everything in the Universe and in computers, and Universe is Computer*, is a repetition or retrieval from memory anyway, preservation has far bigger weight than development (see Zrim, Creating Thinking Machines).

The creative humans do the same ("Creativity is imitation at the level of algorithms", 2003), while most humans CAN'T DO EVEN THAT well: repeating even 3 seconds in the past from their own memories, their own intentions, thoughts etc.

Yet they blame the LLMs or computers - the real nature of LLMs - for being "dumb", "not intelligent" etc. 

What about you yourself if you apply the same metrics?

.....

* Yes, yet another (short) exercise in "pamphlet against...", see Lem, 1963-1964 and Arnaudov, 2001 (Man and Thinking Machine: Analysis of ... etc., Theory of Universe and Mind - see it also for the Universe Computer, as well the recently published appendix "Algorithmic Complexity"... ).

** These days I did that with a new song, a cover of a Bulgarian rock-n-roll classic, however with new lyrics about something else. Stay tuned.

*** Yes, that belongs to a paper, a continuation of the mentioned one etc.

...
*** See also The intro etc. of. The Architecture of Cognitive Control in the Human Prefrontal Cortex, 12.2003 

• Science 302(5648):1181-5 

• DOI: 
• 10.1126/science.1088545

• Source
• PubMed  
• Etienne Koechlin
• Chrystèle Ody
• Frédérique Kouneiher

https://www.researchgate.net/publication/9010007_The_Architecture_of_Cognitive_Control_in_the_H uman_Prefrontal_Cortex

More and more recent literature in the Neuroscience and LLM-related research in one of the Appendices of The Prophets of The Thinking Machines, currently called #listove.

* Edit, 26.6.2026: correction " Even the most intelligent ones need to constantly consult with dictionaries, " --> consult dictionaries

Read More

The Super Science of Philosophy and Some Confusions About it - continuation of the discussion on the "Strong Artificial Intelligence" thread at G+

On : ...

See the recent previous blog-posts from which it continues: ...

//By the way, I agree with some of your claims (but they should be elaborated with examples - grounds of reason, something that you do not do, unlike me - that's the purpose of the "many words", to build up images and context.)

So I actually do have "a theory" of intelligence and universe, which are going together, published prior to Hawkins' book, the trendy "Singularity" PR, the term "AGI", the "Deep Learning" popularity.

And one additional reason for people not understanding each other is our ape-character - social ranking. That's one reason why a few people would bother to check what the other has written, his theory etc., that is above comments of a few lines, unless he displays high social ranking.

One would read Kurzweil, or Hawkins, but not Todor Arnaudov - my theory is in some aspects "kindred " to Hawkins, however published prior to Hawkins' - who cares. There are even worse cases - see below.

"Randall Lee Reetz
10:02 AM
There is a big difference between philosophy and science. Philosophy only respects the thoughts we like to think. It's a mirror on self interest.
Randall Lee Reetz
10:04 AM
Science is what we have had to invent and work at because of philosophy's obvious blind spots."

Todor Arnaudov:

Sorry, the above shows that you don't have a clue about philosophy, especially the rigorous one. The "mirror of self interest" that's exactly the opposite of the systematic philosophy - it aims at being as objective as possible; even the term "objective" is used for "detaching from the Will" (in Schopenhauer terms, that process is related to Brahmanism and Buddhism terms, "losing yourself");
the motive to be as detached from the Will as possible.

The right philosophy aims for grounded, explained, starting from most basic and provable (as long as its possible) grounds - start with the dissertation "On the fourfold root..." which is exactly about the grounding of the truths as matching to the reality; as I mentioned earlier - 180 years prior a known as a seminal paper about that in the official "science" of AI. The line of Marx, Engels and Lenin is also about scientific method of philosophy and a tight correlation of philosophy and science (see especially Engels), unfortunately it's "polluted" with politics/ideology.

Finally "we" who first invented and worked at science methodologically were namely the philosophers... From the Greek ones to the Renaissaince ones. Of course practice and theory, and philosophy and science go and should go hand by hand, that's another word/way of talking about "sensory-motor grounding", related to the "Faculty of understanding" (Kant, Schopenhauer), the mapping between higher abstractions, lower abstractions and the lowest level data that's empty of intrinsic meaning. Good philosophy is on the top of the sciences.

As of the "many words" of mine - that's one of the problems of understanding - incompatibly different bandwidths. It's not only about time, it's about the size of the buffers at the lower levels - working memory capacity, the Faculty of judgment/top-bottom connection within the cognitive hierarchy; differences in the capacity/access to lower level sensory data in various modalities; capacity to imagine/trace the visualization/materialization of the words into images; and of course - simple knowledge when if missing, and also the attention span in time. (See the other comments for more details)

Generalization


Philosophy is about higher generalization, higher cognitive span - it's steps above science in generality and scope, respectively it's harder to be grasped or held in mind by some scientists/engineers, whose subjects normally require shorter/smaller span - "out of memory"/"lacking grounding data"/"insufficient transitory-buffer-capacity" (see my other comment and the answer after it's published), and it requires to know to what the abstractions refer, so it's supposed that you know the concepts and "mechanics" of the special sciences as well.

The good philosophers are also scientists and engineers and artists in one way or another - you should understand the special sciences/domains and search for the general between them. If you are specialist only in one field (or a few) you can't notice or care about the association to the others, the causal chains between them, that your field is in fact the same as some other fields, how your own field came to existence and why is different, etc. The data to make this inference is missing. Most people suffer from multi-interdisciplinary blindness and multi-modal learning limitations.

That's why many people ask questions whose answer is otherwise obvious - they however scan the world with a spotlight in the darkness, instead of having a sun to enlighten the whole view at once, and if there's a lack of memory to keep the track while scanning - that might be a long journey of trial-and-error within the darkness, until reaching the obvious.

Regarding your claim - there are philosophers like that - it is the sophistry, perhaps some of the subjective idealists, perhaps also some servants of some ideological needs or just "immitators" - they mirror what the audience would like to hear, this is often blah-blah-sophy, not quality philosophy. Kant and Schopenhauer for instance were definitive AGI researchers, aiming at understanding intelligence and creativety completely, as much as the means and knowledge at the time allowed; unlike almost all of the official so called AI researchers for the most of the history, who were mere programmers or logicians or engineers or mathematicians, or combination of some of the above - but were missing the grounding "glue". I would name a few who did have a clue about the glue: Vladimir Turchin and Alan Kay.

...

So what's your knowledge or rather *understanding* of philosophy?

In fact the modern scientifical methods were first understood, proposed and formalized by philosophers - start with that fact; they knew better than the "scientists" what one had to do; many of the typical scientists are rather "pedants", performers of what's prescribed (initially by an authority - the "biological" method of the social ranking of the apes which humans are); most people are like that, and that's why science in Europe was dead for a millenium in the middle ages.

It was dead due to bad philosophy and because, let's call them "scientists" (empiricists), the more "practical ones" couldn't make up a way to understand the facts given the low resolution high generality data they started with and due to too much of obedience to the bullshit of the autorities (and perhaps the more limited amount of working memory, compared to the quality philosophers who then came).

"Angels" and "deamons" ruled the world - it was empirically proven - if you were bad, the sacred forces of Good moved you to the stake and you got burned in order to save your soul! People were bad and sinful, they didn't follow the Commandments, that's why they got ill and died! Who ever needed a better explanation - it was proven empirically.

Then science was revived not without the impulse of good philosophers, who are not pedants, they broke the existing fake dogmas, started to break the obedience to the authority and started to apply and suggest rigorous methods, instead of sophistry and pleasing or serving the religious authorities: Francis Beacon had important works on the inductive method in science, Descarthes paved the way of Calculus and started the idealists school of thought, which reached to Kant and Schopenhauer - their philosophy is a theory of how to build an AGI, so long as you understand it and have enough of working memory to keep their sentences in mind.

Indeed, I would claim that Kant was also an abstract mathematician. "Turing machine" is some 150 years late to Kant's true early definition of a computer - namely the Kant's emphasizing of the appropri conceptions of time, space and causality. This is the most abstract definition of a programmable computer: it needs a clock generator ("time") - there must be changes, and at the lowest level they should be expressible in 1D (the lowest possible dimension, the simplest hardware); it needs memory (state, "matter") and causal laws for the changes to happen in predictable manner, rather than having a random number generator* - that's the computer architecture, the instruction set, the most basic "physical laws", upon which higher forms are built.

* in fact random numbers are also not "random" and follow the law of their distribution; the probabilistic laws are also laws

Schopenhauer extended and made it even more clear - and that's all of the above: a Universe model, a model of a computer, a model of simulation of scientific simulations.

Also, just to mention regarding the talk on evolution here - his otherwise philosophical theory is about the Evolution, which in the beginning was such (that is: more abstract/general/higher science than the direct, purely empirical zoology). He discuss about the evolution (development, emergence, "generatio aequivoca") of the live and of the Universe and mind - and is several decades prior to Darwin - "the scientist".

- He induces his theory from the sciences, strongly refers and supports his claims with biology - zoology and botany - and all other scientific knowledge, available at the time;
- he defines the "anthropic principle" (not using the term);
- the "struggle" between the species, the crude forces of nature (the Will) that care more for the survival of the genus rather than the species and even less for the individual beings - however the individual objectivation of the Will, the individuals, struggling for their own surviving and fighting with the other forms of the Will;
- the fitness of the species to their surroundings and that they are mutually correlated and by the peculiarities of the environment one may induce what kind of animals would live there, and that in different places with similar climate and conditions live similar species, even though they were not exactly the same - because the evolution have lead the Will through similar obstacles and they had to survive in similar conditions);
- respectively that infers the survival of the fittest;
- the humans as higher than animals only in the extent and of the level of "the objectivation of the will", but qualitatively the same (something radical for 19-th century, where the egoistical humans were trained by the Western religions that animals "didn't have a soul");
- that the apes and other of the smartest animals are just one step below humans, they have similar "Understanding", but don't have "Reason";
- that the first human was probably born from the womb of an ape (i.e. not a human);
- that the first humans should have been black-skinned or dark, not white.
Etc. etc.

Let's visit also Marx. He is the proto-father of systematic Sociology, modern Economix; his thought, as with Kant, Schopenhauer and Nietsche is also related to Systems sciences, Cybernetics and together with Engels and Lenin they constantly searched and displayed correlations between the more abstract and the more specific (where honestly, where ideologically).

Notice also one other detail. Only Kant among the above was an academical, but as Schopenhauer mentions - he was an exception, because he lived in time of an enlightened monarch, who was a philosopher as well.

All the others were dissidents out of the universities, "non scientific"/according to the adopted values, some of them were viewed as "cranks" - Schopenhauer or "insane" - Nietsche.

That's it with the "bad" and "self-interested" philosophy, which is in fact a higher level of science and a "spawner" of other special sciences, because it sees more of the landscape**, which allows it to understand, explain and forsee things that normal scientists notice or "prove" decades or centuries later with their pocket spotlights and the smaller scope of view. The latter are driven more of the "blind will", the evolution that doesn't understands its aims and just "works". If they had a clue and understood the "bad" and "non-objective" philosophy they may have made the same "new" discoveries much earlier.


** Multi and interdisciplinary researchers are kind of philosophers as well, they are "meta-scientists". Theoretical physics is also close to philosophy, it's speculative and the most general/abstract of its "sisters".

Read More

Working memory capacity, traces of thoughts ... - continuation of the thread on G+ "Strong Artificial Intelligence" [Continues]

Todor Arnaudov's answer to Randall Lee Reetz at ...


No theory at all? I don't think so. I'm sorry if some of the words/terms sound "embarassing" - it's not "politically correct", but are facts called with true names. That's just measures of intelligence, emotions is what humans often see as the most important though and lose the rest (not the method, but your "existential ..." if I rephrase you).

There's rather no theory in the short statements - where is it in 10 words (what's their purpose at all), it suggests super short attention span - first of all. These ones don't have internal "physics" and structure, to name one, and scarcely refer to anything specific, or it's just a term or two or some general claim.

My comment refered to a bunch of specific concepts and works (I could extend it), and there is even inter-sentence structure/comparisons - for example the difference between the span of the intelligence which is one of the reasons for misunderstanding/impossibility for understanding - Einstein vs Schumacher.

As of the "too many words" - one of the reasons why some people don't understand each other. Incompatible bandwitdhs limitations/impatience/grounding/experience/..., another implied by the above is that they can't follow each-other thoughts.

...

Turing machine is too simple a model, it's "exhausted" of material and obvious at a glance, - more structure is needed, such as:

Randall: "a small intelligence advantage. - ... the question is why?"

Todor:

The comment gives some answers, which at a level of verbal short expressions (...) is what one can get.
So it's about:

- Working memory capacity (in narrow cognitive science sense that's the 7+-2 thing, and in more general sense that's the span of the focus, the amount of data one can process in a chunk/a step/a moment, or can hold for longer for processing without refering to external sources);

- The exent of the connections between the levels of cognitive hierarchy, the lossyness/how much is lost between the levels; the extent of assymetry of bottom-up to top-bottom; the availability of access to the different levels of abstraction of sensory records from the different sensory modalities (one who doesn't recognize musical tones cannot understand a melody or a chord or a symphony as a composer who does);

These connections are related to the defined more than 200 years ago hierarchy of Reason(concepts, most abstract, most general, executive function; serial, small capacity)--Facculty of judgement-Understanding(simulation of the causal forces of the Universe)-Input (external and internal).

All of the above is related to the capability of grounding and understanding of grounding, how much one operates with a model of the Universe or just doing blind calculations without understanding;

-- Facculty for imagination, related to "Visuo-spatial sketchapad" capacity, fantasy; that's connected with the WM capacity, the visual hierarchy and grounding as well.


">>a small intelligence advantage. ... the question is why?"

First of all, it may be not small, but "on/off" - such as missing connections/"facculties" to learn in the appropriate sensory-motor data chain, making you incapable to understand operationally for example painting or dancing or playing the piano. Humans possess "general intelligence", but only humanity as a whole with all the technologies has "versatile intelligence", or as I call it: Versatile limitless self-improvement capability. Individuals do not have truly general (versatile) intelligence, they are bottlenecks somewhere in the brain that make humans unable to learn in some of the modalities or inter-modally beyond most mundane levels, compared to the talented ones.

Second - due to the hierarchical/heterarchical nature of brain/mind, and also the working memory limitations, there might be "fire walls", "valves" where conceptions cease to grow - for example the mind cannot hold all of the required samples at the same time in order to see the causal relation.

Third - you're right that time matters, that's one of the dimensions of the scope/span of the attention, both in narrow term related to consciousness and in wider - a topic that you may investigate and revisit for a year, a decade, for life, and to see from new points of view over and over again. This is related to non-cognitive forces as well such as lower biological needs; to distractability, to the development of the mielin in your brain, to your life situation.

However time doesn't help when there are "interrupted" connections (such as sensory modalities data) or insufficient working memory, which in the human brain is not unlimited.

For example, one reason why some people cannot understand Kant thoughts is that they couldn't keep in mind some of the sentences, or just keeping their vocal representation in the loop overloads their mind so they can't think also about the implications, or see the connection to the following sentences or the past ones.

Again, Schopenhauer has actually talked about that issue in the chapter for the "Essential limitations of the intellect" in "The world as Will and Idea", volume two, ch. 15 - that the highest thoughts/consciousness/Reason is in time, sequential and in order to think of something else, the previous thought should be removed, and only some traces should left, that allow to keep the trace, the direction, the trend, and the rest should go deeper. If one doesn't have enough resources to keep the traces alive, to "bear in mind", she would lose the point and would see the thoughts/sentences/concepts as "unrelated".

If that is applied for a long chain of operations and if the results of the operations can be output and stored outside so that they doesn't need to be born in mind anymore and can run on their own without human understanding (technology, machines) or can be encoded in a cheaper and faster memory such as the premotor cortices/cerebellum as motor programs available on demand without loading the precious consciousness resources) - then a huge abyss grow between people and apes, and between humans who're talented/trained and the rest.

Also there are fields where you can go deeper while still keeping the "trace/connection" foot-print low - that's in the highly formal logical chains where the truth of the rest is not questioned, or if you do calculations. Eventually it has to be mapped/connected to some grounds, like the axioms in Euclidean mathematics and the initial conditions of the problem, which are considered "obvious/proven".

So as long as you have enough of memory to bear in mind, you can run for centuries and produce new results; if you don't - you're left as an ape and the gap grows. Some apes or monkeys (don't remember the species) for example would warm themselves if they found a fire, but they would not throw pieces of wood in the fire to keep it going. Also, apes are known to use tools, but not to use tools for creation of new tools - insufficient resources. Similarly as the children grow, their capacity to bear in mind items grow, as the lenght and the complexity of the sentences they can produce or comprehend.

...

As of the citation from Einstein - let me send it in the previous century. In "Paregra and Paralipomena" (if I'm not mistaken), S. emphasizes the fact that the ordinary people have "very short thougts", in other works also he mentions that the difference between a genius and a "blockhead" might have an endless amount of intermediate steps, but in essence it's only quantitative and in the extent, the ingenious ones see the world more distinctly and clearly and are able to focus/concentrate all of their mental energy in one spot, they can be "objective", detached from the "Will", the biological "self interest" such as social ranking/status, money, sex etc. The average people cannot concentrate and think or analyze experience for the sake of it, they are too much concerned with their personal interest which distracts them and keep them for focusing.


As of some more "scientific" evidences for the working memory stuff:

http://www.bing.com/search?q=Working+memory+predicts+G-factor&src=IE-TopResult&FORM=IE11TR&conversationid=
http://www.bing.com/search?q=working+memory+G-factor+1990&qs=n&form=QBRE&pq=working+memory+g-factor+1990&sc=0-0&sp=-1&sk=&cvid=7a4cc15d3587478aabef5b422e824dad

Kyllonen, P., & Christal, R. (1990). Reasoning ability is (little more than) working memory capacity?! Intelligence, 14,
389–433

THE UBIQUITY OF MENTAL SPEED AND THE CENTRALITY OF WORKING MEMORY
Reply to Conway et al. on Jensen on Intelligence-g-Factor
http://www.cogsci.ecs.soton.ac.uk/cgi/psyc/newpsy?intelligence-g-factor.37

A review of visual memory capacity: Beyond individual
items and toward structured representations

http://web.mit.edu/~tkonkle/www/PAPERS/Brady_2011_JOV.pdf
Visual working memory capacity: from psychophysics and neurobiology to individual differences

http://dx.doi.org/10.1016/j.tics.2013.06.006

In the abstract of the above: "The capacity for simple visual features is highly correlated with cognitive ability."


http://www.cell.com/trends/cognitive-sciences/abstract/S1364-6613(13)00126-5#MainText


Regarding time again - it's discussed in the theories of Hawkins (in On Intelligence), of Boris Kazachenko, of neuroscience - the time needed for the sensory input to go to the PFC - images shown for too short time like in the editing in TV and modern action cinema only "fly-through" mind and do not get critically analyzed or steadily remembered. See also the biophysics, researchers on the "ADD" of current society, which is caused largely by watching television, Indeed, the claim that it is a (legal) drug is made by "official" scientists, there are hypotheses that the rise of the drug-addictions in the 60-ies is due to the growing up of the first TV-breed generations. The fast-changing images promote novelty-seeking/dopamine "shortcuts" and addiction, and people become more susceptible to catch addiction from chemical drugs. If you do not believe that, see for example a summary of the research through the decades by the Romanian researcher: http://www.helikon.bg/books/153/-Телевизията-и-детето_153788.html.

My own theory refers the extent/level of understanding also to the amount of time (for deeper, superficial, so that's "real-time physicist" or theoretical physicist etc.). That's the resolution of perception and causation in the dimension of time, respectively related the level of generalization/the span of data records or the prediction period in the future and the levels of detail.


To reiterate something else on the "small advantage":


I believe "small" has to be defined better and more convincingly as a meaningful concept in order to make sense. In Chaos Theory they use to say that "a small difference in the initial conditions may lead to a big difference in the final state". "Small and big" are too definitive, but they are vague. For example in a textbook they once said "1 mm difference of the position of a sled, may lead to 60 m deviation at the end of a slope".

Big? It's just 60000 times the initial difference.
One may say the opposite: it's rather a small change, bearing in mind that the distance between two molecules is bigger than the difference between the initial and the end condition (in orders of magnitude), and that the position depends obviously on all the path and the obstacles and details that the sled has to encounter until arriving at the end - so it's not just the difference in the initial conditions (relative position from a previous run), it's the whole situation and the initial unknowness of the complete situation with appropriate/sufficient resolution of detail that leads to the apparently "big" difference.


Somebody having wrong/unclear/low resolution model getting excited about being unable to predict the results "as he thought he should have been able", instead of criticizing/correcting/enriching/... his model.


#####

And yet another answer to that arrogant and short-memoried guy, insulting me, but giving food for a good post, enjoy! :D

Todor Arnaudov

4:58


Randall, I would rather suggest you do it. I had a written and published theory far more clear than yours (matching in some points) when I didn't even have a moustache. However yeah - I'm a far quicker typist than you, obviously...

[http://research.twenkid.com/agi_english/ - see the slides
http://research.twenkid.com/agi_english/
http://research.twenkid.com/agi_english/Todor_Arnaudov_Theory_of_Hierarchical_Universal_Simulators_of_universes_Eng_MTR_3.pdf
http://research.twenkid.com/agi_english/Teenage_Theory_of_Universe_and_Mind_4.pdf
http://research.twenkid.com/agi_english/Teenage_Theory_of_Universe_and_Mind_3.pdf
]

As of some more "scientific" evidences for the working memory stuff, let me repost it to your insult:

http://www.bing.com/search?q=Working+memory+predicts+G-factor&src=IE-TopResult&FORM=IE11TR&conversationid=
http://www.bing.com/search?q=working+memory+G-factor+1990&qs=n&form=QBRE&pq=working+memory+g-factor+1990&sc=0-0&sp=-1&sk=&cvid=7a4cc15d3587478aabef5b422e824dad

Kyllonen, P., & Christal, R. (1990). Reasoning ability is (little more than) working memory capacity?! Intelligence, 14,
389–433
THE UBIQUITY OF MENTAL SPEED AND THE CENTRALITY OF WORKING MEMORY
Reply to Conway et al. on Jensen on Intelligence-g-Factor

http://www.cogsci.ecs.soton.ac.uk/cgi/psyc/newpsy?intelligence-g-factor.37

A review of visual memory capacity: Beyond individual
items and toward structured representations

http://web.mit.edu/~tkonkle/www/PAPERS/Brady_2011_JOV.pdf

Visual working memory capacity: from psychophysics and neurobiology to individual differences

http://dx.doi.org/10.1016/j.tics.2013.06.006

In the abstract of the above: "The capacity for simple visual features is highly correlated with cognitive ability."

http://www.cell.com/trends/cognitive-sciences/abstract/S1364-6613(13)00126-5#MainText

You don't understand because you don't care or if you do perhaps have
too short working memory for that kind of presentation - one of the serious reasons why people don't understand/care for each other, explained in details in the posts, as well as in referring scientific publications.

You mistake "science" with "limited working memory" (to the extent that you accept).


Yes, science (as well as philosophy, as well as evolution of the Universe in some of its aspects) is about optimization, compression, shortening.

However there's a limit beyond which you start to lose detail and turn to too general, of which you cannot induce anything more , or lack grounding.

There's no cell made of one molecule.

Some of your claims - see them in my theory (for example) among others - published decade(s) or more ago, by a kid.

However they are too general and confused said that way - that's something that John points out.

The "evolution" should be defined more distinctly. For example I agree about the evolution of the Universe as a whole, however at the same time John is right that there are "sub-Universes" - the individuals whose goals are not always synchronized with the overall trend, not all the time.

The individual is a sub-universe aiming at his own goals. And within an individual, there are other subuniverses which also have conflicting goals and struggle to overcome the others.

Boris Kazachenko has concepts called "Conserved core" and "Adaptive interface", that is related to the concept of "Meta-System Transitions" of Vladimir Turchin. That's about evolution in more articulated form of expression, not in one sentence.

The overall results produce the evolution of life and Universe.

At a higher level of an individual there's the genus - which is also having its interests that are above the interests of the individual, but below those of the Evolution (of life) as a whole, or Evolution of universe as a whole (humans and technology destroying living forms, species and genus being eliminated from existence).

Your claims alone, even when agreed on, are not operational in that form.

You need to add more specific and "physical", that is causal, details to "run" it - something that needs more words than most people use to take in one "bite".

And one "philosopher" such as Schopehnauer had 3500-4000 pages worth of incremental and all-directional proves of one-single thought. You have - how much, a half page with words only.

Read More

Visual Illusion Interpretation - ambiguous changes may make recognition of the "true" change impossible by available data | Visual Motion Confusions | Spatio-Temporal Resolution Trade-Offs

A comment of mine on an visual illusion:

".....Silencing the Change

In the demo below, the colours of individual dots continue to change the entire time, but when the dots move, those changes are much harder to see - the change is silenced. ..."

https://plus.google.com/photos/+AllisonSekuler/albums/5931810534032603025/5931810537230454402#photos/+AllisonSekuler/albums/5931810534032603025/5931810537230454402?pid=5931810537230454402&oid=%2BAllisonSekuler

http://www.cell.com/current-biology/retrieve/pii/S0960982211000303

Well, first of all I can see the colours changing when the circles rotate... :-P So it doesn't work for me...

But I came up with a general interpretation of motion confusions. In few words it's just processing/ambiguity limitation.

If objects are translating on the retina, there are similar objects in close distance that take each-others place (so it's harder to distinguish them) and changing too much in different ambiguous ways, and besides it happens in the periphery vision which is terrible, the visual processor cannot know for sure what was the "real" source of change.


The only "real" data is the retinotopic luma/chroma values, however it could be interpreted as:
-  motion of one "item" from the previous "frame"
- some static items that have changed colours
- "blending" items (the way can be also ambiguous, if change is too fast)

The visual processor can't know for sure what was the "true" change.

In this particular case the speed of rotation is low enough and there aren't clues for "blending" between dots etc., the only "meaningful" transformation that viewers track seems to be rigid rotation.

We see rotation, because that's some of the small differences (motion is slow enough) that can be tracked. But as of colour change, depending on the colour of the dots around, and

Spatio-Temporal Resolution Trade-Offs

From particle physics to video, you can't either move or scan faster and see/measure with the same maximum resolution, using the same hardware.

Error Correction

In general many of these illusions are created with patterns which can't be seen in "normal" environment. There are only special common "objects" that change colour that fast on their own - computer screens and TV sets, and the images generated on them.


Infrastructure

By the way, once I complete one of the milestones of the general infrastructure of my "artificial mind" and imagination, I'd be able to test and create a lot of related-stuff in a blaze. The work is going on, but the details are in "low resolution" for now.


See also my interpretation of the Colour Optical Illusions and 3D-reconstruction with Light sources:

http://artificial-mind.blogspot.com/2012/01/colour-optical-illusions-are-effect-of.html



PS. Thanks to L.J.

Read More

Abstract Addressing and Discontinuous Addressing Impede Processing in Generalizing Hierarchies

The article is related to the series of:  Tuesday, October 18, 2011
Rationalization and Confusions Caused by High Level Generalizations and the Feedforward-Feedback Imbalance in Brain and Generalization Hierarchies

The lowest level of addressing for human mind are:

- Orientation/position of muscles/bones.
- Coordinates in sensory matrices:
  - Visual coordinate on the retina.
  - Coordinate of a tactile input on the body.
  - Pitch of the sound.
  - Particluar tastes or smells.

A bit higher level, but still low:- Spatial coordinates of the body within environment (probably related to hippocampus, "place cells", head orientation cells)

   - These are direct inputs to brain.


Selection, combination and generalization of them in space and time forms higher level coordinates/addresses. The patterns/concepts at different levels of abstraction in the cognitive hierarchy have higher level addresses, where higher level addresses and patterns are harder and slower to access and operate and require more context- and attention-switching.

The phenomenon is displayed for example in:

1. Using multiple monitors and looking right-left is generally more convenient than switching between two virtual desktops with a key or worse - menus

Unless you make a visual comparison between the images on the two monitors, then key-switching is more convenient; but for displaying different data, physical coordinates seem better.

2. Having a "to do list" or whatever materials for look up on a paper/notepad next to the monitor/out of the computer is more convenient for mind to switch focus.

Looking aside of the monitor or turning around feels like a "soft context/attention-switch", you don't get as distracted as you would if you have to travel through menus, open an organizer program, check "month, date, priority....". Take for example big paper posters like ones for conferences - you just have to glance it to find what you need. These are different spatial (physical) coordinates of entities as well.

However if the data from different contexts is on the same screen/physical address, that requires an additional abstraction/parameters to mark the different contexts.

3. Chains of intentional high-level operations needed to access abstract addresses impede processing and are more distracting than a single or chain of low level operations such as adjusting muscles/orientation or changing physical location

For example, if you have to open one place, to search there, then find a key to search for another location etc. (either virtual or physical locations).

Intentional actions involve the longest chains of access of different levels of generalizations. If many actions are needed, this implies it takes more time and it's like "crawling" the hierarchy and sweeping the buffers, then it of course will cause slower context-switch back.

4. Orders of patterns associated with flat numbers are easier to remember than if associated with abstract unrelated names

See "Friday, May 2, 2008 What is better for a multi-effects guitar processor's display - numbers or names?" http://artificial-mind.blogspot.com/2008/05/what-is-better-for-guitar-effects.html

5. Continuity within same coordinate system/address space (related to 4.)


Smoothness allows for simper encoding of the following steps and predictions.


Continues...

Read More

Externalism - or on the External Storage and Tools that Extend Human Mind and Brain and are Indispensable for its Development

Recently I discovered that I've shared believes in this philosophical direction since my teenage works (found how it's called).

One introductory term - "paradoxical universality of brain" - brain is specialized in being universal.

The estimation of brain's enormous computing power and memory are arbitrary/not really practical, because brain power is not comparable to computer power. Computers are in fact more flexible and more "universal" than brain, and the ultimate cognitive system is a brain (generally intelligent agent/core, AGI) + a computer and external memory (neither a brain alone, nor a computer alone).

However, you can use a computer to simulate a brain if you have a precise model and speed, but you can't use a brain to simulate a computer, even if you have precise model - most brains are incapable to simulate on their own even the first mechanical computers (say, those which multiplied), logarithmic rulers and artillery tables with precomputed values, and also the papers to store reliably phone numbers or a line of text.

These external "tools" are in fact extensions of our brain - they are just accessed differently than memory which is internal - the access passes through motion (translation), which in fact is just the lowest level of output that the system produces, it's output of data to the physical reality and translation of data in physical reality. At the lowest level output should be decoded down to the "machine code of Universe", the "language of the particles" (my terms).

That's the same as the memory hierarchy in computers - from the fastest registers, to different levels of caches, RAM, hard disks, Internet to external media which requires to be physically moved in order to get accessible.

Culture starts from the changeable and stable intentionally altered and structured environment which allows artifacts/memory to survive longer than a brain and outside of a brain - that includes the overlapping lives of the different generations allowing language to be transmitted.

A human brain alone, without the external tools, without alterable structured stable environment, without intelligent beings around to teach it stuff and especially to teach it sophisticated language... In this conditions it wouldn't go much further than a chimpanzee's brain in similar conditions.

Read More

DARPA PAL, CALO, RADAR - Cognitive Assistants that Learn and Organize

Other ambitious AGI projects, and these ones belong to DARPA. Scary!

I have a project for a similar tool that I call "Research Accelerator" or "Cognitive Accelerator", that initially would be an integration of various NLP, Machine Learning and Computer Vision techniques which assist and speed up in searching, reading, information extraction and any cognitive taks. This system also would be capable to work with GUI on its own, controlling the mouse and keyboard to do some job instead of user.

Time is ticking away and I need partners...

http://www.calosystem.org/

http://www.darpa.mil/IPTO/programs/pal/pal.asp
http://www.calosystem.org/team/

PAL - Personalized Assistant that Learns
RADAR - Reflected Agents with Distributed Adaptive Reasoning
CALO - Cognitive Assistant that Learns and Organizes

Read More

BOLA - Fluid Concepts open source project

My new friend Sergey has just started a project in Source Forge:

BOLA
Bola is a framework for fluid concepts, inspired by the research of Douglas Hofstadter, particularly his Copycat project.

He suggested me this link, which is conencted to the concepts of Copycat:

Metacat

Good luck! I'm curious, even if I don't join you...

Read More