LLM Daydreaming
gwern.net
gwern.net
First time I got good code out of a model, I told my friends and coworkers about it. Not anymore. The way I see it, the model is a service I (or my employer) pays for. Everyone knows it’s a tool that I can use, and nobody expects me to apportion credit for whether specific ideas came from the model or me. I tell people I code with LLMs, but I don’t commit a comment saying “wow, this clever bit came from the model!”
If people are getting actual bombshell breakthroughs from LLMs, maybe they are rationally deciding to use those ideas without mentioning the LLM came up with it first.
Anyway, I still think Gwern’s suggestion of a generic idea-lab trying to churn out insights is neat. Given the resources needed to fund such an effort, I could imagine that a trading shop would be a possible place to develop such a system. Instead of looking for insights generally, you’d be looking for profitable trades. Also, I think you’d do a lot better if you have relevant experts to evaluate the promising ideas, which means that more focused efforts would be more manageable. Not comparing everything to everything, but comparing everything to stuff in the expert’s domain.
If a system like that already exists at Jane Street or something, I doubt they are going to tell us about it.
We haven't successfully resolved famous unsolved research problems through language models yet but one can imagine that they will solve increasingly challenging problems over time. And if it happens in the hands of a researcher rather than model's lab, one can also imagine that the researcher will take credit, so you will still have the same question.
My general sense is that for research-level mathematical tasks at least, current models fluctuate between "genuinely useful with only broad guidance from user" and "only useful after substantial detailed user guidance", with the most powerful models having a greater proportion of answers in the former category. They seem to work particularly well for questions that are so standard that their answers can basically be found in existing sources such as Wikipedia or StackOverflow; but as one moves into increasingly obscure types of questions, the success rate tapers off (though in a somewhat gradual fashion), and the more user guidance (or higher compute resources) one needs to get the LLM output to a usable form. (2/2)Show me the code. Show me your finished product.
After all, the LLM currently has no cognizance, it is unable to understand what it is saying in a meaningful way. At its best it is a P-Noid Zombie machine, right?
In my opinion anything amazing that comes from an LLM only becomes amazing when someone who was capable of recognizing the amazingness perceives it, like a rewrite of a zen koan, "If an LLM generates a new work of William Shakespeare, and nobody ever reads it, was anything of value lost?"
The other day, Claude Code started adding a small signature to the commit messages it was preparing for me. It said something like “This commit was co-written with Claude Code” and a little robot emoji
I wonder if that just happened by accident or if Anthropic is trying to do something like Apple with the “sent from my iPhone”
Google already reported several breakthroughs as a direct result of AI, using processes that almost certainly include LLMs, including a new solution in math, improved chip designs, etc. DeepMind has AI that predicted millions of protein folds which are already being used in drugs among many other things they do, though yes, not an LLM per se. There is certainly the probability that companies won’t announce things given that the direct LLM output isn’t copyrightable/patentable, so a human-in-the-loop solves the issue by claiming the human made said breakthrough with AI/LLM assistance. There isn’t much benefit to announcing how much AI helped with a breakthrough unless you’re engaged in basically selling AI.
As for “why aren’t LLMs creating breakthroughs by themselves regularly”, that answer is pretty obvious… they just don’t really have that capacity in a meaningful way based on how they work. The closest example is Google’s algorithmic breakthrough absolutely was created by a coding LLM, which was effectively achieved through brute force in a well established domain, but that doesn’t mean it wasn’t a breakthrough. That alone casts doubt on the underlying premise of the post.
The same is true of humanity in aggregate. We attribute discoveries to an individual or group of researchers but to claim humans are efficient at novel research is a form of survivorship bias. We ignore the numerous researchers who failed to achieve the same discoveries.
Take yourself outside of that, and imagine you invented earth, added an ecosystem, and some humans. Wheels were invented ~6k years ago, and “humans” have existed for ~40-300k years. We can do the same for other technologies. As a group, we are incredibly inefficient, and an outside observer would see our efforts at building societies and failing to be “brute force”
As individuals, without mentors, we would each die off very quickly. Even if we were fed and whatever until we were physically able to take care of ourselves, we wouldn't be able to keep ourselves out of trouble if we had to learn everything ourselves.
Contrast this with the octopus which develops from an egg without any mentorship, and within a year or so has a fantastically knowledgable and creative mind over its respective environment. And they thrive in every oceanic environment in the wet salty world, including coast lines, under permanent Arctic ice, to the deep sea.
To whatever degree they are "intelligent", it's an amazingly accelerated, fully independent, self-taught intelligence. Our species just can't compare on that dimension.
Fortunately, octopus only live a couple years and in an environment where technology is difficult (very hard to isolate and control conditions of all kinds in the ocean). Otherwise, the land octopus would have eaten all of us long ago.
If it doesn’t (“brute” as opposite of “smart”, just dumb iteration to exhaustion) then you’re right, of course.
But if it does, then I’m not sure it’s patently absurd - novel ideas could be merely a matter of chance of having all the precursors together at the right time, a stochastic process. And it scales well, bearing at least some resemblance to brute force approaches - although the term is not entirely great (something around “stochastic”, “trial-and-error”, and “heuristic” is probably a better term).
Brute force and typical scientific research are such dramatically different things, I have to wonder if bots are getting into HN I almost can’t believe someone would try and argue that’s weird to see the difference.
“Oh you’re using your limbs to move through water? How are you not a dolphin?” lol
But sure, well done, you really got me! Beep boop! I must be a bot because you don’t agree. ”lol”!
Just as a heap of sand stops being a heap when it's small enough, the difference between "science" (not just modern but everything from Newton and Galileo onwards) and "brute force" is the available evidence before whatever hypothesis we're testing.
Scientific research these days requires a lot of prior information, things humanity collectively has learned, as a foundation. We have a lot of weight on our Bayesian priors for whatever hypothesis we're testing.
Sorites even applies to your own attempt to mock it, as the difference between humans and dolphins is "just" a series of genetic changes. Absolutely they're different and it's obvious why you chose the example, but even then it's a series of distinct small changes that are each so small it's easy to blur them together than treat them as a continuum, like we do with water even though that's also discrete molecules.
Humanity massively predates the modern scientific method, it took millennia of mistakes to go from the Greeks being wrong about four elements to finding a bit less than the 91 natural elements, and from there to finding the nucleus (1911) and that it was made of protons and neutrons; and only then did we get to logical positivism (late 1920s), and it was only around WW2 (just before, Karl Popper 1934) that we switched to falsifiability.
Each grain on the heap. We know the fields of work, we know the space of possibilities within the paradigm, the shape of the research can be to constrain that space without finding the answer directly — a divide-and-conquer approach to reducing the space that needs to be then brute forced.
These days we can even automate much of the more obviously brute-force parts, which is why e.g. CERN throws away so much data from the detectors before it even reaches their "real" data processing system. And why SETI automatically processes out any signals that seem to be from in-system before the rest of the work.
It's not a specific number; it's "a collection of items becomes a heap when some indeterminate number of items are obscured by other items on top of them, thus making the total number of items uncountable without disturbing the heap".
Therefore it depends on factors beyond just the number of grains of sand; if you have 1000 grains spread out on a surface so they are can all be distinctly counted, that's not a heap. But if you have 1000 grains gathered together, some on top of each other, then it becomes a heap.
In my opinion, a breakthrough is not the production of new knowledge, it is rather its adoption by the public (beginning with industry).
As such, the rate at which breakthroughs can emerge is bounded by factors external to the producers of breakthroughs. And these outside factors are possibly already limiting.
Another point I would make is that what constitutes a breakthrough is not conditioned by how significant it is, only that it is adopted as a change of processes or mental model. As such, more powerful tools can lead to larger leaps between breakthroughs, but not so much higher rate of breakthrough.
As tools become powerful enough to produce yesterday's year's worth of breakthroughs in a month, then the general public and industry will still wait a year before adopting new technology, only it will see larger progress from the previous iteration. This is in fact the case with LLMs. Even on an avant-garde forum as HN, a very common opinion is "I'm waiting out stagnation before I adopt".
As an over simplification, consider only breakthroughs those that come to have widespread commercial application. If we had an oracle for breakthroughs that could produce arbitrarily many today's-breakthroughs as fast as desired, we'd still be limited by our ability to put them in practice. Work must be allocated, carried out over time, and each new breakthrough requires changing processes and the people involved learning new things, which takes time and energy.
I think this human resistance to change is fundamentally what determines the achievable rate of breakthroughs. As the name implies, a breakthrough is a rupture. It is highly inefficient to be upending one's methods every month. It can even be outright impossible to keep up with all the theoretical advancements, before they have crystallized and been digested into accessible vulgarization, if that is not one's profession (i.e. all time devoted to it).
In my applied sciences field, industry is lagging behind some 20 years. And we ourselves are perhaps a century late to some theoretical advances (I can think of one off the top of my head). At the lowest level, there is resistance to change in that ideas take much longer to be carried to a working prototype, than it takes to have them. Hence, someone who constantly hops to new ideas is guaranteed not to make any progress. By necessity, some stubbornness is selected for. Once things are fleshed out (a multi year endeavour), you still have to convince the broader community (same sub field but not direct collaborators) that your idea has merits surpassing theirs, which is a problem best solved one retirement, and one past mentee hire, at a time. And ultimately convince industrial actors that they should dump millions industrializing these novel methods, when none of their competitors have been doing it (hence it is urgent to wait), the viability (robustness, scalability) of the idea remains to be seen, and the benefits weighed against the risk their practitioner user base won't be able to understand the full scope of the progress and see the need to invest time in learning new things and devising new processes (all of which takes time, money, and makes you dependent on this pioneering supplier). And, lastly, there are three other approaches claiming to be better alternatives.
I don't see a way around this pipeline, and more powerful tools can indeed accelerate some of the stages, but there will remain incompressible delays. Ideas need time to be diffused and understood, all the more if they were advancing at a rapid pace enabled by powerful AIs.
We have a thin conception of genius that can be challenged by Edison's "1% inspiration, 99% perspiration" or the process of getting a PhD were you might spend 7 years getting to the point where you can start adding new knowledge and then take another 7 years to really hit your stride.
I have a friend who is 50-something and disabled with some mental illness, he thinks he has ADHD. We had a conversation recently where he repeatedly expressed his fantasy that he could show up somewhere with his unique perspective and sprinkle some pixie dust on their problems and be rewarded for it. I found it exhausting. When I would hear his ideas, or if I hear any idea, I immediately think "how would we turn this into a product and sell it?" or "write a paper about it?" or "convince people of it?" and he would have no part of it and think that operationalizing or advocating for that was uninteresting and that somebody else would do all that work and my answer is -- they might, but not without the advocacy.
And it comes down to that.
If an LLM were to come up with a groundbreaking idea and be recognized as having a groundbreaking idea it would have to do a sustained amount of work, say at least 2 person × years equivalent to win people over. And they aren't anywhere near equipped to do that, nobody is going to pay the power bill to do that, and if you were paying the power bill you'd probably have to pay the power bill for a million of them to go off in the wrong direction.
I don't see why this is remotely surprising. Despite all the hoopla, LLMs are not AGI or artifical brains - they are predict-next-word language models. By design they are not built for creativity, but rather quite the opposite, they are designed to continue the input in the way best suggested by the training data - they are essentially built for recall, not creativity.
For an AI to be creative it needs to have innate human/brain-like features such as novelty (prediction failure) driven curiosity, boredom, as well as ability to learn continuously. IOW if you want the AI to be creative it needs to be able to learn for itself, not just regurgitate the output of others, and have these innate mechanisms that will cause it to pursue discovery.
So there is quite a lot missing.
Without moving beyond LLMs to a more brain-like cognitive architecture, all you can do is squeeze the juice out of the training data, by using RL/etc to bias the generative process (according to reasoning data, good taste or whatever), but you can't move beyond the training data to be truly creative.
It's clear these models can actually reason on unseen problems and if you don't believe that you aren't actually following the field.
For example, LLM might learn a rule that sentences that are similar to "A is given. From A follows B.", are followed by statement "Therefore, B". This is modus ponens. LLM can apply this rule to wide variety of A and B, producing novel statements. Yet, these statements are still the statistically probable ones.
I think the problem is, when people say "AI should produce something novel" (or "are producing", depending whether they advocate or dismiss), they are not very clear what the "novel" actually means. Mathematically, it's very easy to produce a never-before-seen theorem; but is it interesting? Probably not.
I don't think LLMs are AGI, but in most senses I don't think people give enough credit to their capabilities.
It's just ironic how human-like the flaws of the system are. (Hallucinations that are asserting untrue facts, just because they are plausible from a pattern matching POV)
OK now we're at an impasse until someone can measure this
That said, I think most everyday situations are similar enough to things we've experienced before that shallow pattern matching is all it takes. The curve in the road we're driving on may not be 100% the same as any curve we've experienced before, but turning the car wheel to the left the way we've learnt do do it will let us successfully navigate it all the same.
Most everyday situations/problems we're faced with are familiar enough that shallow "reactive" behavior is good enough - we rarely have to stop to develop a plan, figure things out, or reason in any complex kind of a way, and very rarely face situations so challenging that any real creativity is needed.
I think most human mistakes are different - not applying a lot of complex logic to come to an incorrect deduction/guess (= LLM hallucination), but rather just shallow recall/guess. e.g. An LLM would guess/hallucinate a capital city by using rules it had learnt about other capital cities - must be famous, large, perhaps have an airport, etc, etc; a human might just use "famous" to guess, or maybe just throw out the name of the only city they can associate to some country/state.
The human would often be aware that they are just guessing, maybe based on not remembering where/how they had learnt this "fact", but to the LLM it's all just statistics and it has no episodic memory (or even coherent training data - it's all sliced and diced into shortish context-length samples) to ground what it knows or does not know.
The entropy of communication also makes it so that we mostly won't ever know when a person is guessing or if they think they're telling the truth. In that sense it makes less difference to the receiver of the information what the intent was- even if it came from a human, that human's guessing/ BS level is still unknown to you, the recipient.
This difference will continue to get smaller and more imperceptible. When it will stop changing or at what rate it will change is anyone's guess.
I was just referring to when LLMs fail, which can be in non-human ways, not only the way in which they hallucinate, but also when they generate output that has the "shape" of something in the training set, but is nonsense.
So what. 90% (or more) of humans aren't making any sort of breakthrough in any discipline, either. 99.9999999999% of human speech/writing isn't producing "breakthroughs" either, it's just a way to communicate.
>It's just ironic how human-like the flaws of the system are. (Hallucinations that are asserting untrue facts, just because they are plausible from a pattern matching POV)
The LLM is not "hallucinating". It's just operating as it was designed to do, which often produces results that do not make any sense. I have actually hallucinated, and some of those experiences were profoundly insightful, quite the opposite of what an LLM does when it "hallucinates".
You can call anything a "breakthrough" if you aren't aware of prior art. And LLMs are "trained" on nothing but prior art. If an LLM does make a "breakthrough", then it's because the "breakthrough" was already in the training data. I have no doubt many of these "breakthroughs" will be followed years later by someone finding the actual human-based research that the LLM consumed in its training data, rendering the "breakthrough" not quite as exciting.
All human ideas are a combination of previously seen ideas. If you disagree, come up with a truly new conception which is not. -- Badly quoted David hume
For example, say you appeared out of the future into the stone age and tried to explain to them general relativity or modern semiconductor fabrication .. there are just too many levels of abstraction and discovery from the ideas they would have had for it to be meaningful to say that these are just building upon their ideas, and certainly not just combinations of their ideas.
This is just a completely base level of understanding of LLMs. How do you predict the next token with superhuman accuracy? Really think about how that is possible. If you think it's just stochastic parroting you are ngmi.
>large language models have yet to produce a genuine breakthrough. The puzzle is why. I think you should really update on the fact that world class researchers are surprised by this. They understand something you don't and that is that it's clear these models build robust world models and that text prompts act as probes into those world models. The surprising part is that despite these sophisticated world models we can't seem to get unique insights out which almost surely already exist in those models. Even if all the model is capable of is memorizing text then just the sheer volume it has memorized should yield unique insights, no human can ever hope to hold this much text in their memory and then make connections between it.
It's possible we just lack the prompt creativity to get these insights out but nevertheless there is something strange happening here.
Yes, thank-you, I do understand how LLMs work. They learn a lot of generative rules from the training data, and will apply them in flexible fashion according to the context patterns they have learnt. You said stochastic parroting, not me.
However, we're not discussing whether LLMs can be superhuman at tasks where they had the necessary training - we're discussing whether they are capable of creativity (and presumably not just the trivially obvious case of being able to apply their generative rules in any order - deductive closure, not stochastic parroting in the dumbest sense of that expression).
"Even if all the model is capable of is memorizing text then just the sheer volume it has memorized should yield unique insights, no human can ever hope to hold this much text in their memory and then make connections between it."
Unless you think humans have magic meat then all we are really doing with "creativity" is connecting previously unconnected facts.
In the case of discovery and invention the "facts" being connected may not be things that were known before. An LLM is bound by it's training set. A human is not limited by what is currently known - they can explore (in directed or undirected fashion), learn, build hierarchies of new knowledge and understanding, etc.
Yes, potentially, but the model has no curiosity or drive to do this (or anything else) by itself. All an LLM is built to do is predict. The only way to control the output and goad it into using the vast amount of knowledge that it has is by highly specific prompting.
Basically it's only going to connect the dots if you tell what dots to connect, in which case it's the human being inventive, not the model. The model is trying to predict, so essentially if you want it to do something outside of the training set you're going to have to prompt it to do that.
A human has curiosity (e.g. "what happens if I connect these dots .."), based on prediction failure and associated focus/etc - the innate desire to explore the unknown and therefore potentially learn. The model has none of that - it can't learn and has no curiosity. If the model's predictions are bad it will just hallucinate and generate garbage, perhaps "backtrack" and try again, and likely lead to context rot.
The term "deductive closure" has been used to describe what LLMs are capable of, and therefore what they are not capable of. They can generate novelty (e.g. new poem) by applying the rules they have learnt in novel ways, but are ultimately restricted by their fixed weights and what was present in the training data, as well as being biased to predict rather than learn (which they anyways can't!) and explore.
An LLM may do a superhuman job of applying what it "knows" to create solutions to novel goals (be that a math olympiad problem, or some type of "creative" output that has been requested, such as a poem), but is unlikely to create a whole new field of math that wasn't hinted at in the training data because it is biased to predict, and anyways doesn't have the ability to learn that would allow it to build a new theory from the ground up one step at a time. Note (for anyone who might claim otherwise) that "in-context learning" is really a misnomer - it's not about learning but rather about using data that is only present in-context rather than having been in the training set.
Tobias Rees had some interesting thoughts https://www.noemamag.com/why-ai-is-a-philosophical-rupture/ where he poses this idea that AI and humans together can think new types of thoughts that humans alone cannot think.
> and a critic model filters the results for genuinely valuable ideas.
In fact, people have tryied this idea. And if you use a LLM or anything similar as the critic, the performance of the model actually degrades in this process. As the LLM tries too hard to satisfy the critic, and the critic itself is far from a good reasoner.
So the reason that we don't hear too much about this idea is not that nobody tried it. But that they tried, and it didn't work, and people are reluctant to publish about something which does not work.
This not only affects a potential critic model, but the entire concept of a "reasoning" model is based on the same flawed idea—that the model can generate intermediate context to improve its final output. If that self-generated context contains hallucinations, baseless assumptions or doubt, the final output can only be an amalgamation of that. I've seen the "thinking" output arrive at a correct solution in the first few steps, but then talk itself out of it later. Or go into logical loops, without actually arriving at anything.
The reason why "reasoning" models tend to perform better is simply due to larger scale and better training data. There's nothing inherently better about them. There's nothing intelligent either, but that's a separate discussion.
That the model still makes mistakes doesn't mean it's not an improvement: the non-reasoning base model makes even more mistakes when it tries to skip straight to the answer.
I'm only speaking from personal usage experience, and don't trust benchmarks since they are often gamed, but if this process produces objectively better results that aren't achieved by scaling up alone, then that's a good thing.
Except that we can try the exact same pre-trained model with reasoning enabled vs. disabled and empirically observe that reasoning produces better, more accurate results.
I don't trust most benchmarks, but if this can be easily confirmed by an apples-to-apples comparison, then I would be inclined to believe it.
Research/benchmarks aside, try giving a somewhat hard programming task to Opus 4 with reasoning off vs. on. Similarly, try the same with o3 vs. o3-pro (o3-pro reasons for much longer).
I'm not going to dig through my history for specific examples, but I do these kinds of comparisons occasionally when coding, and it's not unusual to have e.g. a bug that o3 can't figure out, but o3-pro can. I think this is widely accepted by engineers using LLMs to help them code; it's not controversial.
I wouldn't trust comparing two different models, even from the same provider and family, since there could be many reasons for the performance to be different. Their system prompts, training data, context size, or runtime parameters could be different. Even the same model with the same prompt could have varying performance. So it's difficult to get a clear indication that the reasoning steps are the only changing variable.
But toggling it on the same model would be a more reliable way to test this, so I'll try that, thanks.
With code, for example, if a single shot without reasoning would have hallucinating a package or not conformed to the rest of the project style. Then you ask the llm check. Then ask it to revise itself to fix the issue. If the base model can do that - then turning on reasoning, basically allows it to self check for the self-correctable features.
When generating content, you can ask it to consider or produce intermediate deliverables like summaries of input documents that it then synthesizes into the whole. With reasoning on, it can do the intermediate steps and then use that.
The main advantage is that the system is autonomously figuring out a bunch of intermediate steps and working through it. Again no better than it probably could do with some guidance on multiple interactions - but that itself is a big productivity benefit. The second gen (or really 1.5 gen) reasoning models also seem to have been trained on enough reasoning traces that they are starting to know about additional factors to consider so the reasoning loop is tighter.
> There is no such external validator for new theorems
There are formal logic languages that will allow you to do this.
On the other hand any novel science usually requires deep and wide exploratory research, often involving hard or flawed experimentation or observation. One can train LLM on a PhD curriculum in astrophysics, then provide that LLM with API to some new observatory and instruct it to "go prove cosmological constant". And it will do so, but the result will be generated garbage because there is no formal way to prove such results. There is no formal way to prove why pharaohs decided to stop building pyramids, despite there being some decent theories. This is science too, you know. You can't formally prove that some gene sequence is responsible for trait X etc.
I would say a majority of science is not formally provable.
And lastly, you dismiss books/texts, but that is a huge chunk of intellectual and creative work of humans. Say you are an engineer and you have a CAD model with a list of parts and parameters for rocket for example. Now you need to write a guide for it. LLM can do that, it can generate guide-looking output. The issue is that there is no way to automatically proof it or find issues in it. And there are lots of items like that.
Maybe not formally in some kind of mathematical sense. But you certainly could have simulation models of protein synthesis, and maybe even higher order simulation of tissues and organs. You could also let the ai scientist verify the experimental hypothesis by giving access to robotic lab processes. In fact it seems we are going down both fronts right now.
Basically if:
A) Scientist has an idea > triggers LLM program to sift through a ton of data > LLM print out correlation results > scientist read them and proves/disproves an idea. In this case, while LLM did a bulk of work here, it did not arrive at a breakthrough on its own.
B) LLM is idling > then LLM triggers some API to get some specific set of data > LLM correlates results > LLM prints out a complete hypothesis with proof (or disproves it). In this case we can say that LLM did a breakthrough.
The LLM doesn't have to know about the critic though. It can just output things and the critic is a second process that filters the output for the end user.
Something about the whole approach is bugged.
My pet peeve: "Unix System Resources" as explanation for the /usr directory is a term that did not exist until the turn of the millenium (rumor is that a c't journalist made it up in 1999), but AI will retcon it into the FHS (5 years earlier) or into Ritchie/Thompson/Kernigham (27 years earlier).
The bug is that LLMs are fundamentally designed for natural language processing and prediction, not logic or reasoning.
We may get to actual AI eventually, but an LLM architecture either won't be involved at all or it will act as a part of the system mimicking the language center of a brain.
The models are currently trained on a static set of human “knowledge” — even if they “know” what novelty is, they aren’t necessarily incentivized to identify it.
In my experience, LLMs currently struggle with new ideas, doubly true for the reasoning models with search.
What makes novelty difficult, is that the ideas should be nonobvious (see: the patent system). For example, hallucinating a simpler API spec may be “novel” for a single convoluted codebase, but it isn’t novel in the scope of humanity’s information bubble.
I’m curious if we’ll have to train future models on novelty deltas from our own history, essentially creating synthetic time capsules, or if we’ll just have enough human novelty between training runs over the next few years for the model to develop an internal fitness function for future novelty identification.
My best guess? This may just come for free in a yet-to-be-discovered continually evolving model architecture.
In either case, a single discovery by a single model still needs consensus.
Peer review?
So I strongly agree that, especially when are talking about the bulk of human discovery and invention, the incrementalism will be increasingly in striking distance of human/AI collaboration. Attribution of the novelty in these cases is going to be unclear, when the task is, simplified something like, "search for combinations of things, in this problem domain, that do the task better than some benchmark" be that drug discovery, maths, ai itself or whatever.
In examining the possibility of genuinely creative computing, Gelernter discovers and defends a model of cognition that explains so much about the human experience of creativity, including daydreaming, dreaming, everyday “aha” moments, and the evolution of human approaches to spirituality.
https://uranos.ch/research/references/Gelernter_1994/Muse%20...
- are capable of evaluating the LLM's output to the degree that they can identify truly unique insights
- are prompting the LLM in such a way that it could produce truly unique insights
I've prompted an LLM upwards of 1,000 times in the last month, but I doubt more than 10 of my prompts were sophisticated enough to even allow for a unique insight. (I spend a lot of time prompting it to improve React code.) And of those 10 prompts, even if all of the outputs were unique, I don't think I could have identified a single one.
I very much do like the idea of the day-dreaming loop, though! I actually feel like I've had the exact same idea at some point (ironic) - that a lot of great insight is really just combining two ideas that no one has ever thought to combine before.
I noticed one behaviour in myself. I heard about a particular topic, because it was a dominant opinion in the infosphere. Then LLMs confirmed that dominant opinion (because it was heavily represented in the training) and I stopped my search for alternative viewpoints. So in a sense, LLMs are turning out to be another reflective mirror which reinforces existing opinion.
Infact, they're trained to please us and so in general aren't very good at pushing back. It's incredibly easy to 'beat' an LLM in an argument since they often just follow your line of reasoning (it's in the models context after all).
So it's a bit of an anti-climactic solution to the puzzle but: maybe the naysayers were right and they're not thinking at all, or doing any of the other anthropomorphic words being marketed to users, and we've simply all been dragged along by a narrative that's very seductive to tech types (the computer gods will rise!).
It'd be a boring outcome, after the countless gallons of digital ink spilled on the topic the last years, but maybe they'll come to be accepted as "normal software", and not god-like, in the end. A medium to large improvement in some areas, and anywhere from minimal to pointless to harmful in others. And all for the very high cost of all the funding and training and data-hoovering that goes in to them, not to mention the opportunity cost of all the things we humans could have been putting money into and didn't.
Setting up the map-elites dimensions may still be problem-specific but this could be learnt unsupervisedly, at least partially.
The way I see LLMs is as a search-spqce within tokens that manipulate broad concepts within a complex and not so smooth manifold. These concepts can be refined within other spaces (pixel -space, physical spaces, ...)
> Concept 1: {Chunk A} > Concept 2: {Chunk B}
In addition to the other criticisms mentioned by posters ITT, a problem I see is: What concepts do you feed it?
Obviously there's a problem with GIGO. If you don't pick the right concepts to begin with, you're not going to get a meaningful result. But, beyond that, human discovery (in mechanical engineering, at least,) tends to be massively interdisciplinary and serendipitous, so that many concepts are often involved, and many of those are necessarily non-obvious.
I guess you could come up with a biomimetics bot, but, besides that, I'm not so sure how well this concept would work as laid out above.
There's another issue in that LLMs tend to be extremely gullible, and swallow the scientific literature and University press releases verbatim and uncritically.
(See original argument: https://nitter.net/dwarkesh_sp/status/1727004083113128327 )
We're looking at our reflection and asking ourselves why it isn't moving when we don't
Of course random new things are typically bad. The article is essentially proposing to generate lots of them anyway and try to filter for only the best ones.
Given access to unlimited data, LLMs likely could spot novel trends that we cant but still cant judge the value of creating something unique that it has never encountered before.
Gwern isn't doing that here. They say: "[LLMs] lack some fundamental aspects of human thought", and then investigates that.
This mirrors something I have thought of too. I have read multiple theories of emerging consciousness, which touch on things from proprioception to the inner monologue (which not everyone has.)
My own theory is that -- avoiding the need for an awareness of a monologue -- a LLM loop that constantly takes input and lets it run, saving key summarised parts to memory that are then pulled back in when relevant, would be a very interesting system to speak to.
It would need two loops: the constant ongoing one, and then for interaction, one accessing memories from the first. The ongoing one would be aware of the conversation. I think it would be interesting to see what, via the memory system, would happen in terms of the conversation emitting elements from the loop.
My theory is that if we're likely to see emergent consciousness, it will come through ongoing awareness and memory.
Intuitively, it doesn't feel like scaling up to "all things in all fields" is going to produce substantial breakthroughs, if the current best-in-class implementation of the technique by the worlds leading experts returned modest results.
> Among Prime Intellect's four thousand six hundred and twelve interlocking programs was one Lawrence called the RANDOM_IMAGINATION_ENGINE. Its sole purpose was to prowl for new associations that might fit somewhere in an empty area of the GAT. Most of these were rejected because they were useless, unworkable, had a low priority, or just didn't make sense. But now the RANDOM_IMAGINATION_ENGINE made a critical connection, one which Lawrence had been expecting it to make [...]
> Deep within one of the billions of copies of Prime Intellect, one copy of the Random_Imagination_Engine connected two thoughts and found the result good. That thought found its way to conscious awareness, and because the thought was so good it was passed through a network of Prime Intellects, copy after copy, until it reached the copy which had arbitrarily been assigned the duty of making major decisions -- the copy which reported directly to Lawrence. [...]
> "I've had an idea for rearranging my software, and I'd like to know what you think."
> At that Lawrence felt his blood run cold. He hardly understood how things were working as it was; the last thing he needed was more changes. "Yes?"
Eventually LLM output degrades when most of the context is its own output. So should there also be an input stream of experience? The proverbial "staring out the window", fed into the model to keep it grounded and give hooks to go off?
But most promising would be to use the Dessalles theories.
Here is 4.1o expanding this: https://chatgpt.com/s/t_6877de9faa40819194f95184979b5b44
By the way - this could be a classic example of this day dreaming - you take two texts: one by Gwern and some article by Dessalles (I read "Why we talk" - a great book! - but maybe there is some more concise article?) and ask LLM to generate ideas connecting these two. In this particular case it was my intuition that connected them - but I imagine that there could be an algorithm that could find this connection in a reasonable time - some kind of semantic search maybe.
There's a podcast I listened to ~1.5 years ago, where a team used GPT2, further trained on a bunch of related papers, and used snippets + perplexity to highlight potential errors. I remember them having some good accuracy when analysed by humans. Perhaps this could work at a larger scale? (a sort of "surprise" factor)
i.e. AGI is a philosophical problem, not a scaling problem.
Though we understand them little, we know the default mode network and sleep play key roles. That is likely because they aid some universal property of AGI. Concepts we don't understand like motivation, curiosity, and qualia are likely part of the picture too. Evolution is far too efficient for these to be mere side effects.
(And of course LLMs have none of these properties.)
When a human solves a problem, their search space is not random - just like a chess grandmaster's search space of moves is not random.
How our brains are so efficient when problem solving while also able to generate novelty is a mystery.
for example, euclidean geometry organizes the perceptions of "space" around us, or newtonian mechanics organizes the perceptions of "motion" or turing machine organizes computation.
LLMs cannot create theories yet since they are missing the capability to create symbols to represent things and link them via rules.
I am trying to answer that for myself. Since every logic is expressible in untyped lambda calculus (as any computation is), you could have a system that just somehow generates terms and beta-reduces them. In even so much simpler logic, what are the "interesting" terms?
I have several answers, but my point is, you should simplify the problem and this question has not been answered even under such simple scenario.
What the LLM companies are currently selling as "reasoning" is mostly RL-based pre-training whereby the model is encouraged to predict tokens (generate reasoning steps) according to similar "goals" seen in the RL training data. This isn't general case reasoning, but rather just "long horizon" prediction based on the training data. It helps exploit the training data, but isn't going to generate novelty outside of the deductive closure of the training data.
So how do you pick the logic in which to do reasoning? There are "good reasons" to use one logic over another.
LLMs probably learn some combination of logic rules (deduction rules in commonly used logics), but cannot guarantee they will be used consistently (i.e. choose a logic for the problem and stick to it). How do you accomplish that?
And even then reasoning is more than search. If you can reason, you should also be able to reason about more effective reasoning (for example better heuristics to cutting the search tree).
I was talking about the process/mechanism of reasoning - how do our brains appear to implement the capability that we refer to as "reasoning", and by extension how could an AI do the same by implementing the same mechanisms.
If we accept prediction (i.e use of past experience) as the mechanistic basis of reasoning, then choice of logic doesn't really come into it - it's more just a matter of your past experience and what you have learnt. What predictive rules/patterns have you learnt, both in terms of a corpus of "knowledge" you can bring to bear, but also in terms of experience with the particular problem domain - what have you learnt (i.e. what solution steps can you predict) about trying to reason about any given domain/goal ?
In terms of consistent use of logic, and sticking to it, one of the areas where LLMs are lacking is in not having any working memory other than their own re-consumed output, as well as an inability to learn beyond pre-training. With both of these capabilities an AI could maintain a focus (working memory) on the problem at hand (vs suffer from "context rot") and learn consistent, or phased/whatever, logic that has been successful in the past at solving similar problems (i.e predicting actions that will lead to solution).
Yet, the consensus seems to be we don't quite have AGI; so what gives? Clearly just making good predictions is not enough. (I would say current models are empiricist to the extreme; but there is also rationalist position, which emphasizes logical consistency over prediction accuracy.)
So, in my original comment, I lament that we don't really know what we want (what is the objective). The post doesn't clarify much either. And I claim this issue occurs with much simpler systems, such as lambda calculus, than reality-connected LLMs.
Prediction doesn't have goals - it just has inputs (past and present) and outputs (expected inputs). Something that is on your mind (perhaps a "goal") is just a predictive input that will cause you to predict what happens next.
> And I would even say that this problem (giving predictions) has been solved by RL.
Making predictions is of limited use if you don't have the feedback loop of when your predictions are right or wrong (so update prediction for next time), and having the feedback (as our brain does) of when your prediction is wrong is the basis of curiosity - causing us to explore new things and learn about them.
> Yet, the consensus seems to be we don't quite have AGI; so what gives? Clearly just making good predictions is not enough.
Prediction is important, but there are lots of things missing from LLMs such as ability to learn, working memory, innate drives (curiosity, boredom), etc.
The breakthrough isn't in their datasets.
The feedback loop on novel/genuine breakthroughs is too long and the training data is too small.
Another reason is that there's plenty of incentive to go after the majority of the economy which relies on routine knowledge and maybe judgement, a narrow slice actually requires novel/genuine breakthroughs.
I want to remember I heard about it in several podcasts
There is a breakthrough happening. in real time.
The OP's proposed solution is a constant "daydreaming loop" in which an LLM is does the following on its own, "unconsciously," as a background task, without human intervention:
1) The LLM retrieves random facts.
2) The LLM "thinks" (runs a chain-of-thought) on those retrieved facts to see if they are any interesting connections between them.
3) If the LLM finds interesting connections, it promotes them to "consciousness" (a permanent store) and possibly adds them to a dataset used for ongoing incremental training.
It could work.