Douglas Hofstadter: Analogy as the Core of Cognition [video]
youtube.com
youtube.com
From what I recall, to Hofstafer analogy making isn't some higher level cognitive process, certainly not a language based one, but basically is THE cognitive process all the way from perception on up, and is the mechanism by which we form object categories in the first place.
As always Hofstader's ideas are interesting, but I can't say I agree with them. It seems that the key evolutionary benefit, and function, of a brain is prediction, which is the superpower that moves us from being stuck in the present to being able to "see" (predict) the future, and therefore from being merely reactive to being able to proactively plan and predict future outcomes (what will the sabre-tooth do, where is the water supply?) based on our experience.
Given the never-same-twice nature of sensory perception, before you can predict you need to be able to generalize/categorize, which I think Hofstader would regarded as analogy making (how is this thing I'm seeing similar to what I've previously seen?), although it seems the actual mechanism involved is embeddings or embedding-like representations where similar inputs have similar representations, and what might more simply be considered as associative recall provides the generalization from view/instance to identity/category.
So, is it really analogies all the way up, or are our perception and cognitive processes better regarded as generalization and prediction, which seem not only seem to have direct and obvious neural realizations, but also match the evolutionary needs that we would expect to exist?
If he had accepted the "Bitter Lesson", I think he would have been at the forefront of LLMs.
The path forward all big llm providers ("ai" labs) have gone is neuro-symbolic (even though they publicly would never labeled it as such to not admit critics like Gary Marcus were right - even though all their actions actually point in that direction).
Smolensky's latest paper posted here the other day has some thoughts on how modern neural networks might beconsidered neurosymbolic, or rather "gradient symbolic processing," from another perspective entirely.
I wouldn't say the bitter lesson has given out! If you haven't noticed, these things keep getting bigger and bigger.
This is so funny to me, because as many people know, sharing an analogy with another person is the fastest way to LOSE an argument with someone, or otherwise spiral it into an unproductive place.
I think it’s Scott Adams who used to say analogies work well for explaining. They work terribly for persuasion.
I just experienced this in a conversation. An analogy offers an opportunity to engage with the straw man and miss the forest for the trees.
If I say, your face is like the moon, I mean it is bright, not that it is a million miles away.
Ex:
Explainer: "Getting a spleen means cutting open the patient and taking it out. It's just like how I'm going to unzip this section of the patient-shaped doll, and remove this little purple bean. In both cases a hole is necessary in a similar location."
Hostile listener: "Nonsense! I can just buy beans at the store! So just buy a spleen! No hole!"
I’m not sure if Hofstadter puts it this way, but to me even the core aspects of your sentences in this post have roots in analogies. What does it mean to lose an argument or to spiral it to a different place? There is no place, there is no lost item, but we talk about these abstract ideas in ways that largely depend upon understanding things like physical objects and space and movement.
Perhaps the common occurrence of analogies in language is because language is distinct from the things it is referencing, so some degree of abstraction is unavoidable?
This is true if the arguer is hostile, but as I've gotten older, if I get the sense that someone is entering an argument with the primary goal of "winning", I'll try to avoid that framing or just look for an offramp entirely.
Sure, the other party might think they "won", but they were going to think that anyway. For those more inclined to feel that the point of talking to each other is to learn from each other, I'll continue to use analogies and other things reasonable people understand.
Absolutely. The best kinds of arguments are those where you both have the shared goal of reaching consensus, and treat reaching consensus as a collaborative activity of finding the correct answer, even though you disagree with the starting point. There are a few techniques for this, which work well when operating in good faith, and can backfire when dealing with hostile counterparties. Most things work badly when arguing with hostile counterparties.
Yes; here is one instance: <https://www.youtube.com/watch?v=n45D_zi3O5g#t=14m13s>
It's notable that when we do use analogies there are really two use cases
1) Trying to use them to communicate an alternate or simpler POV
2) To make a joke, by pointing out a ridiculous comparison
The joke case is interesting, because this seems to be the basis of half of humor - breaking expectations (the other half perhaps being taboo breaking?), which really isn't how you'd want an adaptive cognitive system to work. You need to see the predictive regularities in the world, not see absurd similarities that make you laugh.
If the workings of those circuits are obvious to you, I'd really like to learn. Do you mean the level of analysis at https://transformer-circuits.pub/ ? (That looks like good work but not a deep understanding.)
Hofstadter referenced this back in the day as a promising beginning: https://en.wikipedia.org/wiki/Sparse_distributed_memory which sounds kind of similar in style to the embeddings you bring up.
Our cortex is connected in a loop with our thalamus (cf recent talk of looped transformers!), which can be separated into thalamic core and matrix. Basically our senses feed into the thalamic core, and cortical outputs (predictions) feed into the thalamic matrix. The thalamic connections/loop seem to act a bit like a patch panel connecting different regions of the cortex.
The cortical microcircuit is a real physical circuit - a pattern of synaptic connections between layers and inputs/outputs. Those Transformer circuits are more functional than physical. We know how a Transformer is physically connected, with attention heads etc, and this "circuit" analysis then describes how it functions, e.g. with attention heads in adjacent transformer layers effectively combining to form "induction heads" that copy data across embeddings.
If you think about it, this requires recognizing something common in things which are very different: e.g. book and towel have different texture and color, and you somehow need to separate what's depicted from the background. So that's pretty much innate, core brain function. I mean, any animal with vision can recognize an object from the background - otherwise vision is useless. But for humans (and some animals) this translates to depiction of object on a flat surface very easily.
So, yeah, analogies-all-the-way-down seems plausible. Even object-vs-background and depiction-vs-paper is itself an analogy.
Our cortex (all of it - visual, auditory, association. motor..) is remarkably simple and regular - it's six layers of neurons with a specific circuit design, that appears to make it a predictor, with top down predictions meeting bottom up sensory feedback. As noted, it makes a ton of sense that evolution would have equipped us with a predictor (and the more powerful the better). Prediction directly supports reasoning and planning (multi-step what-if prediction).
So, my question for Hofstader, or you, would be given that the brain seems to be literally built for prediction, then what is to be gained by regarding this as analogy making instead? Is there any evidence for this? Even as a philosophical viewpoint, doesn't that make for very convoluted explanations of simple things like reasoning and language? (I explained prediction-based reasoning above in one sentence - the analogy-based explanation takes a whole dense book!).
One could certainly use the language of analogies to describe perception (Pooh's ears on the towel are analogous to his ears in the book, etc), but what is the benefit of doing so, rather than the simpler view of this just being generalization (which is specifically what the animal needs), and where is the evidence in the brain?
Analogy-making points to more specific mechanism: ability to identify features within representations, and modulating associative lookup using those features. I.e. it's not as simple as D(x, y) i.e. distance between embeddings, but something like D(f(x), f(y)) where f projects representation to a specific feature space.
The evidence can be found e.g. in LLM interpretability - people were able to identify features as something concrete. Also in our cognition - we can tell _how_ two things are similar, or find Y similar to Y in context of F.
"Prediction" by itself seems like a black box: if we are trying to predict sensory input, that gives no explanation to our ability to focus on specific details, or explains why some prediction failures are more likely, etc. OTOH if we say that cortex might be 'disassembling' sensory signal into high-level features and is trying to predict those (features themselves might be identified as "something useful for prediction", i.e. something which explains a lot of variance), it's much easier to connect low-level "prediction" to our high-level cognition.
- The never-same-twice nature of sensory perception
- The need to recognize objects from different directions, distances, partially obscured, etc
- Because the nature of the world is that predictive categories are equally or more important than individual identities, especially to basic survival, and for basic organisms
Rather than "generalization" being unspecific, I'd say that as far as perception goes there is a very concrete realization of generalization in our brain - embeddings and associative recall by partial embeddings (see half a face, recall the whole face).
Analogies don't seem to offer any benefit over prediction in terms of what to focus on - there is always a need to separate foreground (object) from background, and to learn what is instance-specific vs generalizable. Is is irrelevant or coincidence that sabre-tooths are near the water hole, or is this part of a predictive pattern (or analogy of some sort?!) that should usefully be learnt?
When we are not sure about what we are looking at, we often then switch from the gestalt "looks familiar" to looking for specific points of confirmation. Is that car in the parking lot really ours? Does it have our air-freshener hanging from the mirror?
Needing to focus on details can also be part of generalization and exception making - when we recognize that two situations that seem similar are in fact different, and we are trying to isolate what the differences are.
I think that the strongest argument for our brains being built for prediction, not analogy making, are that:
1) Prediction is exactly what is needed to survive and thrive - to take advantage of the external world, and not be killed by it. We directly need to predict. We do not need to make analogies, unless analogies can also be used as a way to predict.
2) Our brain appears to be built for prediction, and prediction gives rise to the very obvious learning signal of prediction failure. One of the lessons of LLMs is confirmation of the power of prediction as all you need to drive learning.
The overall way our perception seems to work is indeed as much by prediction as sensory input - in a way prediction really comes first, with sensory inputs narrowing/correcting the prediction (and when you turn it off then we predict/dream without the constraint of sensory reality). The environment you are in already leads you to predict what might be there, and memory of what specifically was there previously will shape those predictions. Sensory feedback corrects/narrows those predictions, and then - if necessary - we look for specific memorized features to confirm what we are seeing/sensing.
Considerations of evolution, and what would be the sensory and cognitive "needs" of organisms of ever increasing complexity, all the way up to ourselves, can also point to what brain capabilities are older and more fundamental, and would have been there to be refined, adapted and built upon later. Even the most simple organism needs to never-same-twice generalize, but will be stuck being reactive until evolution has either equipped it with "reactive predictions" such as the fly being startled by the fast-moving visual input (predictive of being swatted by the horse's tail), or eventually developing a predictive brain like our own, especially useful to an animal that is a generalist living in an environment that is more difficult to predict.
He is still an AI skeptic in as many ways that he can reasonably be, but doesn't deny the raw ability. Actually he thinks it will eclipse humans and he is a doomer.
He sees it as being an extremely empty type of intelligence though.
But I hope that people who have an intuitive understanding of contemporary machine learning (not me) will sometimes watch videos like this and think about things at a higher level. LLMs have a LOT of assumptions built in.
I think the Hofstader's view of modern LLMs is actually a deeply human and touching one. Looking at his work over the years, his curiosity has always veered towards human thought. He could have written GEB with a focus on completely different examples of self-reference, but he chose three striking humans from history. When he's describing modern systems as "empty intelligence", I think there's a little bit of heartbreak in his perspective, because he sees them as fundamentally different from humans in a way that leaves the part he loves--the "I" in the loop--out of the equation. He gave an interview a few years ago where he explains his feeling as being "diminished" not in a "What will I do if I'm not the best at math?" kind of way, but more specifically as he puts it, that humans are "imperfect, flawed structures".
My instinctive reaction to claims like this is, is it really any different from humans? Why?
Current LLMs have a lot of limitations. The pretraining without the ability to learn beyond their context is a big one. Every prompt is literally instantiating a new instance of the intelligence. If you take these limitations into account, it's not at all clear to me that they're all that different from humans when it comes to the fundamentals: we learn from training data, we respond according to what we've learned.
In fact, many humans resemble the pretrained model quite closely - they do most of their learning in their early years, and don't update much after that except in relatively limited ways. You do get lifelong learners, but they're the exception.
Taking all this into account, why is the one kind of intelligence "extremely empty" whereas the other is not? Is this not just anti-machine bias when it comes to intelligence?
The drive toward the formation of metaphors is the fundamental human drive, which one cannot for a single instant dispense with in thought, for one would thereby dispense with man himself.
It's akin to the amino acid interactions in proteins that hold biological matter together, and determine it's shape and active form. Protein folding and narrative/storytelling have strange homology :)
(I work in this area via collective intelligence, and these ideas are very dear to me during the past decade. It's neat to see the intuitions seemingly becoming validated in language models)
Hmmmmmm. While I get what you mean, and I don't disagree, please be cautious when making analogies between biological systems and more distant fields.
Yes, folding is driven by hydrophobic collapse due to interactions between residue sidechains. Really, though, we are just describing two 'complex systems', where large numbers of diverse interactions between elements leads to diverse and emergent structures.
The fact of a hole in an analogy is not in itself information. Of course there's a hole.
> I get unreasonably upset when people nitpick holes in good analogies…
What you’re describing is exactly why I strongly recommend avoiding analogy when communicating: Don’t tell me what a thing is like, tell me what the thing is. This forces really thinking about how to describe the thing precisely and succinctly. (Edit: The listener will make their own analogy.)
But you do have to choose an analogy that your listener does not need to consider first. They need to know the subject of the analogy without having to question your description of it.
narrative -> trope + theme -> memes
I'd be very surprised to meet someone who learned that lesson by comparing protein folding (suggesting a high level of intelligence and probably education in order to understand) to storytelling, rather than having learned that stories are made up of parts in grade school or as a child just hearing, reading, thinking about, and making stories.
Very much agree. Every abstraction loses something and so is wrong in some sense, but some shave off in clever places, and allow new shortcut paths of thought and insight, sometimes to an existing place... but sometimes to a new place that was previously unreachable, or not sufficiently reachable by the necessary type of attention/minds.
> The mechanisms and math behind protein folding and an analogy between that and storytelling doesn't necessarily give us any mechanism for or deep understanding into storytelling.
Respectfully, I actually disagree (as someone trained in biochemistry).
For one example:
proteins = 1D chemical structures, peptides popped onto end one-at-a-time, that in the right chemical environment, fold into a low-energy 3D state that transforms its environment through work
stories = 1D semantic structures, words popped onto end one-at-a-time, that in the right cultural/cognitive environment, fold into a low-energy HD mental model that transforms its environment through work
This has led me to render embeddings of sliding windows of narrative streams (to investigate shapes of broadcast stories), and look for analogs to "active sites" in the way that a story interacts with the cultural medium, measured through dimensional reduction of aggregate valence reactions amongst a viewing audience.
Obviously there's more to the metaphors I use than that, but that's the one I'm leaning on to inform my work with sensemaking and map-making from valence response data. Aggregating crowd-sourced valence reactions to linear narratives gains a lot of insight by reflecting it off how we process and think about linear peptides in protein folding.
I've been developing tools inspired by these metaphors, and my research and prototypes are directly informed by what knowledge of the biological domain inspires me to port into social domain. I'm lucky to have interactions with interested parties associated with federal government, practitioners of democratic reform, and academics investigating platonic space hypotheses. I'm definitely on the edge of an ecosystem, but anything I've done of value has been guided by the metaphors that direct my attention to potential applications and research. Someone could get to my conclusions another way, but the metaphors are my cheat codes to arrive there early :)
And it goes both ways along a good metaphor -- People who develop tools in the biological sciences have told me that my porting their methods into my own distant domain has inspired them to improve their tools to ask better questions of their own biological data. So it's very fruitful work that I wholly credit to travelling both ways across an insightful metaphor. https://i.imgur.com/AeYxx7p.png
He described a piece of AI research (in the 60s?) where they set up a rule system - like Prolog or similar - with the mapping 'sun <-> nucleus' and 'planet <-> electron'. Then they ran the rule engine and lo! produced an analogy between the solar system and atoms.
Now this is a shallow analogy, as there is only a very weak correspondence between atoms and planetary systems. Electrons do not actually orbit in the same plane, but in more complex 'orbitals', and they are better described as probability clouds anyway.
So while yes there is a _homology_ (I re-read the parent and realised they used homology instead of analogy) between these two systems, it does not tell us much. Is there some mapping between how water molecules scaffold the folding process and storytelling? Does narrative structure tell us anything about local minima in the folding surface? I doubt it.
I do not want to be too harsh here - reasoning through analogy is fun and can be useful, it is just limited in what it can do, especially the further apart the systems are.
https://george-lakoff.com/books/metaphors-we-live-by/
Took a course with Lakoff as an undergrad and it was compelling.
(The post primarily emphasizes self-referentiality rather than analogy, but I suspect similar things could be said about analogy.)
I’m confused why there seems to be a dismissal of the most basic ‘strange loop’ of the LLM - the fact that it’s evaluating a context to choose the next word, then reevaluating in a context where that word has been appended.
That always seemed to me like the essence of a Hofstadterish strange loop, so the emergence of Hofstadterish phenomena (self rep, etc) doesn’t seem surprising.
And if anyone's reading this and hasn't read Hofstadter, you're making a mistake, it's utterly perspective-changing stuff. Well, it was for me, at least.
Then, after I graduated from college, a friend of mine who was finishing his physics degree mentioned that he had just finished it and that it was life-changing. I read it again and deeply regretted having put it down for those 12 intervening years.
At first, I flipped through the book looking at the Escher prints. My second time through, I read the dialogues. Next I had a go at reading the chapters and got lost. Maybe on my fifth read-through (years later) I actually attained an understanding of Gödel's Theorem (at least for a brief moment before it vanished again).
I credit GEB (and its dialogues like Little Harmonic Labyrinth) for helping prepare me to deal with the kind of multi-level abstraction involved in developing a language interpreter or machine simulator. Especially that time I had to debug a crash that only occurred when running my simulator inside a simulated instance of itself (i.e. when three levels deep, but not two).
First, LLM AI systems have incredibly huge blind spots despite their incredible performance on many tasks, so self-reference might be the key to what's missing (or not). For example, an LLM AI just solved Navier Stokes, but could not explain the LEAN proof, while a human could.
Second, Hofstadter had more than one idea about intelligence and the mind (see the OP topic of this HN discussion!), and LLMs are quite on-point regarding analogy-forming.
So it may be well be that self-reference and analogy are both part of intelligence, and self-reference is missing and that is leading to major weaknesses.
Third, Aaronson links to a (paywalled) Hofstadter essay form 2023, which was eons ago in AI, and from the intro it seems to be about the sadness of AI replacing humans, not a disparagement of AI ability.
"an LLM AI just solved Navier Stokes"
I assume you mean:
"an LLM AI [company] just [claimed that a team of mathematicians they hired, using their AI] [may have] solved [part of] Navier Stokes[, definitely prompted by (and possibly by looking at) the work of human mathematicians."
> But the idea that you’d need explicit self-referentiality before you could get convincing and world-changing conversational intelligence? Let it be buried in a Westminster Abbey or Arlington National Cemetery for the most important wrong ideas in human history
I am not as confident as you that an LLM cannot explain the lean proof of Navier-Stokes. Rather, I would expect human mathematicians to try and understand the proof without assistance, so as to obtain community understanding in a lossless way.
Hofstadter generally seems depressed about the possibility that human cognition is not so special or complicated, and that AI/LLMs may have replicated or even surpassed it. Here’s another piece from 2023 of his: https://www.lesswrong.com/posts/kAmgdEjq2eYQkB5PP/douglas-ho...
Q: How have LLMs, large language models, impacted your view of how human thought and creativity works? D H: Of course, it reinforces the idea that human creativity and so forth come from the brain's hardware. There is nothing else than the brain's hardware, which is neural nets. But one thing that has completely surprised me is that these LLMs and other systems like them are all feed-forward. It's like the firing of the neurons is going only in one direction. And I would never have thought that deep thinking could come out of a network that only goes in one direction, out of firing neurons in only one direction. And that doesn't make sense to me, but that just shows that I'm naive.
It also makes me feel that maybe the human mind is not so mysterious and complex and impenetrably complex as I imagined it was when I was writing Gödel, Escher, Bach and writing I Am a Strange Loop. I felt at those times, quite a number of years ago, that as I say, we were very far away from reaching anything computational that could possibly rival us. It was getting more fluid, but I didn't think it was going to happen, you know, within a very short time.
And so it makes me feel diminished. It makes me feel, in some sense, like a very imperfect, flawed structure compared with these computational systems that have, you know, a million times or a billion times more knowledge than I have and are a billion times faster. It makes me feel extremely inferior. And I don't want to say deserving of being eclipsed, but it almost feels that way, as if we, all we humans, unbeknownst to us, are soon going to be eclipsed, and rightly so, because we're so imperfect and so fallible. We forget things all the time, we confuse things all the time, we contradict ourselves all the time. You know, it may very well be that that just shows how limited we are.
Heh, this fits in with my theory that most people won't find LLMs intelligent, instead we'll discover people aren't.
I don't see why Conway's game of life should not be considered self-referential though... I mean it's isn't it Turing complete? I don't see how any definition of self-referentiality should require throwing out systems which are minimally turing complete... If Turing complete is not enough, doesn't that imply that computable artificial intelligence is impossible in the first place?
All this being an empirically unproven theory/hypothesis. But an extremely strong one (if you ask me).
> The big, old ideas about intelligence that ended up basically vindicated were the ideas about how intelligence is about prediction, and prediction is about compression, and compression is about finding better and better upper bounds on Kolmogorov complexity.
Well, sure - prediction works as a baseline, if you abstract out everything else about what counts as intelligence and subsume it under this framework. Any protocol for intelligence can be entirely reduced down to this without trying to understand anything about the structure of intelligence. Solomonoff induction is "vacuous" in this sense too - it doesn't try to understand any intension about the turing machines it finds simple, it just brute forces over all of them. So you're making a claim about intension, whether you want to or not.
It's really no different than say, Darwinians, saying, "what matters is victory at the end". I think the statement has value in the context of some discussions; if the discussion has veered too far in one direction, push as a reminder. That's good. But trying to make grand universal statements like this makes it vacuous.
It's one of these classic unfalsifiable too general statements. The way the end of the article is framed is also icky - the way I'm reading it, it needs to prove all the old curmudgeon evil theories wrong. It reminds me of internet debates around what "the scientific method" is or whatever. It has this kind of zeal that needs to pit itself against the "enemy" and assert itself as the sole right viewpoint, even when say, naive "let's just be empirical" is wrong (e.g. recently had a discussion here about Mach and Boltzmann about this and had a similar interaction).
---
Basically, "shut up and calculate" type theories are never correct, and furthermore, you yourself don't shut up and calculate, or think that way at all, and a higher level intelligence won't conceive itself as operating on that anyways. So what are we doing here? It seems like a way to get mad and feel like an intellectual victim.
Weirdly, Aaronson doesn't even seem to be conflating the two:
> Consciousness and subjective experience of course remain extremely mysterious.
I think he's just misreading Hofstadter as stating that cognition depends on self-reference?
Anyway, ML was already onto analogies with word2vec, which famously could answer questions like "man is to woman as king is to ____" mathematically. This stuff seems quaint now.
Hebbian, analogies, agent monitoring and self pruning. Look at it or don’t, but I made it and I’m proud of it and I’m tired of marketing tactic-ese for self promotion.
For how it works, https://gist.github.com/adromero/e6a776e6a6100d53116f4053677...
A good place to start for anyone that is interested is his book Metaphor's We Live By.
I will never not find it amusing how a simple thing like money is obscured or "spun" by a thin veil of pseudointellectual bullshit.
And it's relevant today given how people like to anthropomorphize LLMs, and compare biology to digital machines we make. Of course there are similarities. But the point about metaphorical thinking is we are mislead by treating metaphors as literally true.
Anyway if the stronger claims in the book are at all true, it might impact how an alien species thinks differently than us (given a lot of our metaphors are biologically based). And could present significant difficulties for decoding an alien signal.
Whether it is the evolved shape of a flying creature or an abstract depiction of a meteorite (or even a comet!), there is a kind of intuition about the sharp end and the flaring of air or fluid around it, with some sort of trailing tail from whence it came...
One can imagine many observers who could find our space probes would also have some of these experiences?
Of course, all of this assumes some kind of visual perception and cognition to even recognize the plaque as having markings made with an intent to communicate some abstract ideas...
The 2006 date of this talk precedes his "Surfaces and Essences: Analogy as the Fuel and Fire of Thinking" book where he really makes the claim of analogies being the basis of cognition. I haven't read "Essences", but the major criticism of it is that it was also mostly just endless examples of analogies, not delivering on the claim of the title.
Hofstader's most famous book by far was GEB, which in retrospect seems it achieved it's cult status more for it's sprawling scope and intellectual appeal rather than for any coherent exposition or explanation of its ambitious subject matter. It was a book that felt deep and clever, but really was never more than a bunch of interesting talking points, only loosely related.
Still, having read GEB I had high hopes for his "Fluid Concepts and Creative Analogies" which came 15 years later in 1995, which one might have expected to be equally ambitious in scope and breadth of ideas, but it turned out to be his least popular book, and was more of a narrow discussion on some of his GOFAI programs such as "Copycat" trying to at least demonstrate the use of analogy as a computational mechanism, if not one related to the brain.
I'm reminded now why I never bothered to read his final book "Surfaces and Essences", and reading the reviews of that confirms what anyone who had read "Fluid Concepts", or watched this video, might have suspected - that he was basically out of ideas.
For anyone interested in analogies and their role in language and thought, a book that has better stood the test of time is George Lakoff's "Metaphors We Live By".
But I think that LLMs are bad at something that has to do with taking seemingly disparate concepts and assimilating one into the other to convey a novel idea.
Like these articles...
<https://spectrum.ieee.org/jimi-hendrix-systems-engineer>
<https://zed.dev/blog/agentic-xanadu>
...are anything but convincing once you read past the gravitas that the LLM lends the prose.
Metaphors are at the core of LLM cognition too via vector embeddings. The distance between embeddings is an inverse measure of their metaphorical attachment. That lets AIs scrape meaning from natural language, which demonstrates that natural language is an encoded sequence of metaphors.
And after some thought I'm curious whether this means that LLMs ought to be adept at forming new metaphors as opposed to handling pre-existing ones.
People with more knowledge, especially practical working knowledge over many fields, tend to have much more freedom in finding solutions.
Now, an interesting question is how good at LLMs are at analogy, especially deeper transferable concepts?
Edit: it actually is very possible GP wasn't being disparaging, in which case truly sorry! I'm not trying to be too snarky.