Rich Sutton on AI creativity and discovery

https://twitter.com/RichardSSutton/status/2061216087744946656

by yimby • 1 month ago

212 125 1 month ago

twitter.com

https://www.youtube.com/watch?v=K5LAFEjTlBA

Legend2440• 1 month ago

[flagged]

dang4• 1 month ago

"Please don't post shallow dismissals, especially of other people's work. A good critical comment teaches us something."

https://news.ycombinator.com/newsguidelines.html

habitue• 1 month ago

[flagged]

dang4• 1 month ago

"Please don't post shallow dismissals, especially of other people's work. A good critical comment teaches us something."

https://news.ycombinator.com/newsguidelines.html

lioeters2• 1 month ago

That reminds me of a quote by Freeman Dyson on Stephen Wolfram's work.

"There's a tradition of scientists approaching senility to come up with grand, improbable theories. Wolfram is unusual in that he's doing this in his forties."

I always felt it was an unfair dismissal of someone's life's work. Maybe it was true but it didn't enrich the discussion or our understanding. I suppose it means even a respected thinker can be guilty of shallow dismissals and saying hurtful things in public about others.

It's similar to a "thought-terminating cliché", in that it just reinforces an existing opinion without adding anything, making us think deeper, or furthering the conversation.

edot• 1 month ago

I don't quite follow his point. Is it: a) that we need a new foundational algorithm that integrates a goal (one with "taste") directly into the training step, or b) that we need to point trained models towards goals as they iterate?

If it's a), he doesn't propose such an algorithm, and I don't know how you'd do it at such a low level because how do you quantify abstract goals? Did he suggest such an algorithm and I misread? If it's b), that already exists, see AlphaEvolve or any number of things he said. Or, to be a bit of a smart-ass, just type /goal and let it rip ...

I also think he's just categorically wrong that LLMs cannot do good and novel things. And if it can, then you could just say "well that's not novel, that's derivative". A simple example, if I make up a programming language with an LLM and it works well for my purposes, then is that not novel and good? I mean, is any language other than FORTRAN not novel?

Everything is derivative and you can put an LLM in a loop to evaluate LLMs trying things. I must be misunderstanding because he's too smart to be this wrong.

oliveiracwb• 1 month ago

LLMs possess the map but are unable to discern fertile from barren ground. For instance: how does Anthropic's new model generate promising 'medications'? Because, beyond the knowledge embedded within the model, it has assimilated AlphaFold's reasoning paradigm. By itself, Claude would be incapable of engineering a protein analysis method

whattheheckheck• 1 month ago

Idk one of his yt video presentations was saying we're entering a "designer" age of the universe

https://youtu.be/ThFq87Rp21s?si=SrKj72_X8bjnB6ED

Around 35min mark

nateroling• 1 month ago

No, I think I he’s saying that we have that, and we should use it more.

AlphaGo uses discovery when it evaluates potential moves and iterates.

Claude Code uses discovery when it generates a script and the evaluates whether it works or not.

He’s saying we need to allow ai systems to do the evaluation and iteration themselves for science and engineering the same way we do for code.

Basically, harness engineering for engineering.

anthonypasq3• 1 month ago

but this has to happen during training no?

Papazsazsa• 1 month ago

Creativity = variation + evaluation + selection. It's not bad, though every example he gives has a built-in scoring function haha.

Best thing about nerds is watching them try and build frameworks and formulas for the creative act. Like a metronome trying to compose a symphony.

dwd2• 1 month ago

"We have many AI systems which can give us more. ... and Claude-Code, which have brought true advances in science, mathematics, and programming."

That contradiction kind of says he doesn't know what he's talking about.

phyzix5761• 1 month ago

Yes, the guy with a PhD in Machine Intelligence, co-author of Reinforcement Learning: An Introduction, which is universally considered the bible of the field, recipient of the AAAI fellowship award and the Turing Award, and the inventor of Temporal Difference Learning doesn't know what he's talking about.

E-Reverance2• 1 month ago

I don't completely disagree but its worth noting how new a lot of the empirical evidence in favour of LLMs are, so its not impossible to be a tad ignorant of the present

lowbloodsugar• 1 month ago

“When a distinguished but elderly scientist states that something is possible, he is almost certainly right. When he states that something is impossible, he is very probably wrong.”

https://en.wikipedia.org/wiki/Clarke%27s_three_laws

dwd2• 1 month ago

Sure, but does that mean he's right all the time about all things, including everything in his own field?

He is saying no generative AI is going to produce output that is both good and novel because it is always derivative. And then adds a generative AI (Claude Code) into his list of AI that have produced output that he feels is good and novel, invalidating what he is arguing.

"...no matter how many instances of white swans we may have observed, this does not justify the conclusion that all swans are white."

aureate• 1 month ago

Surprisingly enough, Turing Award winner and father of reinforcement learning Richard Sutton knows perfectly well what he's talking about. The whole talk is about the need to have the ability to test novel outputs against reality and iterate to find ones that are good. This is exactly what Claude Code, the agent framework, adds to Claude, the LLM, to allow it to find novel coding solutions that actually work.

AndrewKemendo2• 1 month ago

You know it’s occurred to me recently that there really is no value in this age to any kind of professionalism or seniority

If you don’t agree with somebody, nothing else matters

It’s like people (you as an example) have taken the concept of experts and fucked it up so bad that simultaneously everybody thinks they’re an expert while also dismissing everybody else who claims they are an expert

It’s like the whole concept is entirely poisoned. Worse everyone is smugly pointing at the Wikipedia for “appeal to authority.”

Nothing new I suppose, Socrates after all was driven to suicide by the madness of his society accusing him of impiety.

Lerc• 1 month ago

I think the variation, evaluation, and selection idea is a good, if not the only, way do do creative work.

I don't think I would attribute anything in that process that I would consider an AI to be incapable of.

The characterisation of variation like this would seem to rest on the same 'random but directed' crutch that some free will arguments rest upon.

There is no random but directed of course, there is random and there is caused, and there are things that use both as components, but the random remains wholly random, and the caused remains entirely deterministic.

I think there is a good case to say that, in many fields, AI is better than humans at evaluation.

To find avenues to consider, I'm not entirely convinced that human innovation is more than a heuristic that appears more chaotic by virtue of a inconsistent and opaque formulation.

Many aspects of ideas com from noting how some two things are different and then considering that axis of difference when applied to another thing.

The possibilities thrown up by this extremely simple method are vast enough to require multiple layers of evaluation, most could be dismissed out of hand by a quick 'This is nonsense' check that I suspect people do so often and at a rate that it wouldn't even rise to the level of consciousness.

erickhill• 1 month ago

(For others like me who need to look it up.)

Stochastic: describes a process, model, or system that involves inherent randomness, meaning its future states or outcomes cannot be predicted with absolute certainty but can be described using probability distributions.

rembicilious• 1 month ago

"So that is my call to arms. If we want the full power of AI scientists, then we should share the goals with them so they can create, evaluate, discover, and in these ways fully participate in achieving the goals. Let’s be bold! Let’s fully automate Creativity and Discovery!"

Should we automate exercise and play as well? How about learning?

The machine didn't have a soul, so we donated ours.

Eureka! My AI found it!

musebox353• 1 month ago

I understand the skepticism. I am worried about the implications of AI as well. The deeper issue at stake is that the depth of scientific knowledge has been increasing for a very long time. Now you get to have a PhD in esoteric subproblems and that slows down research especially if the discoveries require depth in multiple subdomains. Socially and economically training people in every combinatorial combination of subfields at the required depth may not be possible. I am especially interested in two problems to be resolved and do not care if an AI scientist performed the discovery. It will be humbling, but totally worth it:

- Fusion (a clean sustainable form): Without this I think we are heading in a very wrong direction, whether it is conflict or climate change does not matter. Everyone is aware of this and instinctively afraid of the implied loss of quality+quantity of life.

- Cure for Cancer: It is a world wonder even in Civ. I and for good reason. As a father of a teenager, every time I hear a story of someone losing a parent/child I cringe. We have to accept this as a reality of life until a proper/generic cure is found that eliminates the most common offenders.

I am skeptical that we will have AGI anytime soon and I think the social aspects will help balance the technical developments even it becomes a reality (Three laws, A Butlerian uprising, you name it).

Chess bots can beat grandmasters, but I have a friend who takes his son to tournaments. Humans are still playing chess, kids in the same tournament with grand masters. We have to have faith in the humanity, or all else will not matter.

And I will definitely keep playing Factorio even if AGI comes to pass ;-)

magicalist• 1 month ago

Eh, it seems increasingly clear that no matter the fusion breakthroughs to come, it's going to be complicated and big and therefore expensive compared to renewables. By the time we figure it out it'll be a nice breakthrough but will end up on the shelf.

chadgpt33• 1 month ago

Note that all fusion energy is unsustainable as it irreversibly turns water into helium.

antonvs2• 1 month ago

Commercial fusion energy is almost certainly a pipe dream. We already have good alternatives. Short of some unpredictable major breakthrough in physics on the level of something like cold fusion, the current activity in the fusion space will basically just serve to demonstrate definitively that it's not viable.

scotty79• 1 month ago

Should we automate exercise and play as well? How about learning?

Obviously. Exercise, play and learning is in large part what makes us smart. If we want AI to be smart we should definitely automate those.

But that's not what you had in mind. To address that I can only say that you are allowed to still do the automated things manually.

tines• 1 month ago

It seems human beings spend a significant portion of their lives trying to figure out how to live as little as possible.

newsclues• 1 month ago

It seems like human societies are run by elites who convince the masses to live as little as possible so that the elites can fully enjoy themselves.

Like how during Covid the beaches were closed while the billionaires partied on yachts.

amarant• 1 month ago

They do say "live a little", and it strikes me now that perhaps I've completely misinterpreted that phrase. For some reason I've always thought it to be a call to live a lot, or at least more.

But now I see that my interpretation is almost opposite of what the words mean... Huh...

pixl972• 1 month ago

Most discovery takes a ton of iteration and repetition. That's a lot less fun than the hedonism of sitting on a beach drinking a mahi thai.

Same with learning, humans historically where generalists without that deep of knowledge when compared to one another. Now we study a quarter of our short lives just to get to the point where we can specialize for the rest of our lives. This situation doesn't seem exactly tenable as complexity increases in the future.

I don't know what the future looks like, but I can tell you that entropy and complexity will only increase.

E-Reverance2• 1 month ago

I think its worth emphasizing that his argument isn't completely against generative ai, but rather its environment. Although I don't see why it would be impossible for something like an LLM to learn some sort of self-play within its context window

highfrequency5• 1 month ago

Unless I'm missing something, this argument seems to apply only to the original pretraining era (eg GPT 1-4). The post-training and reinforcement learning paradigms are clearly doing variation, evaluation and selective retention no?

kibibu• 1 month ago

The transcript does seem to overlook post-training steps like Reinforcement Learning with Verifiable Rewards (RLVR) (but I'll certainly won't claim that Rich Sutton is unaware of such things; RLVR has a very narrow set of evaluation approaches).

I wonder if this is a precursor to Keen Tech leaning into David Silver's Ineffable Intelligence approach.

LarsDu88• 1 month ago

This was exactly what I was thinking of. RLVR is the secret sauce behind o3 and its many successors.

Its the secret sauce behind why the current models are so great at coding and soon to be unbeatable at math.

LLMs can pose many questions and if they are easily verifiable, fine tune very heavily. A lot of the world models discussion will inevitable lean into simulations as verification.

porridgeraisin6• 1 month ago

RLVR still does not expand beyond the base distribution though, it only mode-seeks within it.

i.e, evaluation, retention yes. variation or "planning" no.

That is not to say you cannot use LLMs. Alpha evolve does exactly that. It uses an external simple evolutionary planner though. The overarching point he's making is that our planner is still "dumb" and we need to work on it.

When you iteratively guide an LLM in claude code, you are the external planner. That also works.

highfrequency5• 1 month ago

RLVR still does not expand beyond the base distribution though, it only mode-seeks within it.

Seems clearly false. Pretraining finds the mean/mode of the data distribution. RL can easily generate many samples around that mode, evaluate them on an external source of truth (eg compile the code and run it) and then selectively train on the good samples. This clearly can go beyond the initial data distribution.

balazstorok2• 1 month ago

There seems to be a problem with how he poses the problems alphaGo and these GAI models face.

AlphaGO is given a hard evaluation externally. It did not itself come up with it.

When GAI models are given an external hard evaluation, they can also succeed in many different domains (that is one of the remarkable features, succeeding in many domains) ranging from simple programming tasks to frontier mathematics (disproving conjectures recently) to writing more optimized kernel code than before.

And there is plenty of RL especially in these fields where the solution may be extremely complex but eval is rather less complex. And even the discovery and the "evolution-like" trace-selection is also happening.

For this reason it seems strange to compare it to AlphaGO as alphago is given a hard eval independent of itself, from an external source (humans) in a narrow domain. If GAI is given such, it can also show some remarkable results.

But what I find more strange is that innovation and moving forward in many many many cases does not require truly novel ideas but instead a high-quality execution of layering different methods, tactics, ideas on top of each other. Because in many domains our collective knowledge is incredibly sparse and complex, something being able to recombine tools, models, ideas in a high quality way (as he mentions being selective) I think is extraordinarily powerful. And in such cases, with a finite exploration horizon (time, resource available) with 1% "good choices" vs 3% "good choices" are worlds apart, incomparable.

Most importantly: none of the above is about intelligence, it's barren solution-farming to important, valuable problems we have. Most of the AGI and intelligence-related debate seems to miss out on this simple fact. (Insert the usual stuff like a plane being unable to fly like a bird or a submarine not swimming is totally irrelevant to it being useful).

And then a final point: do we really think this thing is incapable of doing better on average on problems we average people face in our lifetime? What should we think, how should we define human intelligence when we give out degrees in science or medicine for 60-70% exam results on problems considered to be generic in the field?

YeGoblynQueenne• 1 month ago

> (Insert the usual stuff like a plane being unable to fly like a bird or a submarine not swimming is totally irrelevant to it being useful).

Just a brief reminder that planes have wings with airfoils just like birds and submarines have air tanks just like fish have swimming bladders.

Some birds fly without flapping their wings much, too, e.g. albatrosses.

balazstorok2• 1 month ago

You are right in that. Propulsion is different in both cases. The point is, until you have not recreated the function, difficult to say which part is important in other implementations (~Marr levels).

Leonardo tried to create flapping wings.

thedreammachine• 1 month ago

Humm maybe. But a plain model sampling outputs obviously isn't doing discovery in the AlphaGo sense. But once you put the model in a loop with tests, feedback, tools or even a human picking the good result, it starts to get much closer to the process he's describing.

armchairhacker• 1 month ago

I don’t think ML can’t be creative or make discoveries. I think creativity and discovery are, ultimately, simultaneously thinking about the right seemingly-disparate concepts (whereas algorithmic thinking is more obviously related concepts). If not an LLM, some other model can generate random ideas, rank them, then output the best.

But I think humans are better at it, while ML is better at algorithmic thinking. “Better” being more efficient and something we more enjoy doing; we can also more accurately rank what subjectively appeals to humans (i.e. taste), especially ourselves.

I think ML should be optimized for tasks that require more generalization than programming, but are still mostly logic. Like software development, translation, and tools for art and discovery.

simianwords5• 1 month ago

I'm trying to keep an open mind and understand what the author is trying to say because he is credentialed.

His main point is that discoveries involve

1. Variation,

2. Evaluation, and

3. Selective retention.

He makes a jump saying AI is only capable of 1) and humans are capable of 1) 2) and 3). I don't know what makes humans special enough that they can do 2) and 3)?

In fact, the more you think of this it is kind of strange - in science humans can only do "evaluation" because they have access to the real world. They can evaluate a new drug because they can do it on people so it is not some inherent limitation of AI but rather access to physical realm.

Finally I want to ask a specific thing: how do you mathematically falsify what this person is saying? How can you formally prove that - no AI can not "evaluate"? I ask because I make AI evaluate a lot of people's claims and it works for me.

zeroonetwothree3• 1 month ago

He actually says the areas in which AI has had the novel successes are those which can be evaluated (like coding or Go). Not that it can’t happen at all.

simianwords5• 1 month ago

That’s my point, he says ai does well where evaluation is neurosymbolically closed.

But so do humans? How do humans make discoveries without having formal ways to evaluate? In my pharma drug example, humans could evaluate only because they had access to the physical realm.

I can’t think of an example of humans evaluating a discovery in a way that LLMs can’t. can you?

skybrian• 1 month ago

He's saying that pre-training an LLM alone can't do it, but if you run an LLM in a loop with tools (like any coding agent) then it can. Also, the technique his group came up with should be used more:

This is the weakness of deep learning that is alleviated with a new algorithm that my group presented in Nature a couple of years ago. Our “continual backpropagation” made one small change: every so often a less-used neuron would be re-initialized to small random weights. This allows the variation to continue and plasticity to be retained.

Here's the paper: https://www.nature.com/articles/s41586-024-07711-7

It has a fair number of citations, but I haven't looked into how much it's used.

simianwords5• 1 month ago

Sorry this makes no sense — humans also use tools to evaluate their discoveries.

Kant said something like this: knowledge can’t be obtained by pure thinking, it needs interaction with the world.

This is obvious to me so why is the author making a claim that LLMs can make knowledge without access to environment but purely through thinking in aether

cortesoft• 1 month ago

When we ask for a fiction or novelty, the AI can give it to us because its processing is in part stochastic. Every decision can go multiple ways and will go different ways and produce a different trajectory every time. The trajectory can be random—and thus novel—or it can be based on the training data—and thus “good” because the training data is good, sourced from people or reality. Thus, the trajectory is either novel or good—based on randomness or based on data—but never both at the same time.

This doesn't seem true? You can be both random and based on training data.

zeroonetwothree3• 1 month ago

Only to a limited extent.

maxbond• 1 month ago

I think they meant more "it can be extrapolated or interpolated" or "it can be high variance and 'creative' or it can be low variance and 'reliable/correct/likely'". If you want to see something new, the model will need to step off the manifold. But the manifold is where you've learned the "correct" solutions live.

Shitty-kitty2• 1 month ago

One has to be very specific when throwing around words like "creative" when talking about A.I

Can A.I create art. Well it can create something that's pleasing to our senses but art is ultimately about conveying human feelings and emotions. Even as humans, understanding art is not universal. "feelings and emotions" and therefore art, can be deeply tied to a particular groups shared beliefs and experiences.

Can it be creative in non-subjective fields such as math or sciences. Einstein derived GR from his creative thought experiments. If an A.I poped out GR's field equations simply by testing different mathematical frameworks that resolve the issues discovered by experiments, is that creative? Perhaps but certainly not in the same way.

orbital-decay• 1 month ago

His definition of creativity is closed-loop, i.e. the produced concept has to be novel for the system itself, not for the side observer.

brador• 1 month ago

Contemporary humans downvote art because it makes them uncomfortable.

They just want dopamine. No thinking because that hurts.

artistonn2• 1 month ago

art is ultimately about conveying human feelings and emotions

you made a small error, art is mostly about generating an emotion in the viewer/listener/.... not about transmitting an emotion of the creator

the Wikipedia page on art starts with:

Art is a diverse range of cultural activity centered around works utilizing creative or imaginative talents, which are expected to evoke a worthwhile experience

https://en.wikipedia.org/wiki/Art

so AIs can do art, because they are only required to generate an emotional response in the receiver

jejeej• 1 month ago

No art is a means of expression - the creator uses an object as a vehicle to capture and distribute an expression.

Also lol @ using Wikipedia as some source of absolute truth.

Shitty-kitty2• 1 month ago

Using your provided definition, you come to a conclusion completely ignoring the "utilizing creative or imaginative talents" part.

TheOtherHobbes2• 1 month ago

Current models are trained on image pastiche and style remixing. But there's no reason you couldn't add an Artistic Director layer which has been trained on emotional and cultural signifiers and to direct the pastiche and remixing.

The practical problem is that models have very limited prompt adherence. The level of detail you can specify in scene design is very crude. So you can get the slop effect where there's a lot of in-fill pastiche detail, but you could never create something like this, where all of the incidental objects are specifically included to enforce the message.

https://en.wikipedia.org/wiki/The_Awakening_Conscience

It's basically the professional version of the "Draw me a pelican on a bicycle" problem.

There are situations where you want that level of creative control, and current image generators don't get close to it.

And without it you can't get to the meta-creativity level where you're creating a new aesthetic that's a cultural landmark - which is what the famous artists did, and still do.

liamlaverty• 1 month ago

But there's no reason you couldn't add an Artistic Director layer which has been trained on emotional and cultural signifiers and to direct the pastiche and remixing.

I gave this approach a shot over the first few months of this year[1] (although my director didn't have any custom training). The results were interesting, but I'd not call them "art", since they're low-quality derivative pieces. With reasoning traces enabled, you can see that there's not much intent going on. Though they do attempt to include "incidental objects" to reinforce meaning, like in this jungle scene[2].

[1] https://news.ycombinator.com/item?id=48105385

[2]https://www.liamlaverty.com/paint-by-language-model/inspect/...

ACCount373• 1 month ago

Sounds like a skill issue?

Recent image models are advancing rapidly at prompt adherence specifically, and being able to iterate on the same image is propelling them even further. Images 2.0 being the poster child of this "agentic iterative image composition" approach.

volkercraig• 1 month ago

You can create art with everything from sticks and mud to glass and air. Of course you can make art with AI.

Now if the question is, can a machine make art, well ultimately someone needed to turn the machine on and design the machine to make art, so arguably that person/people are the ones making the art.

Historically, every question of "is x art" ends up having the answer "yes". I don't know why people fall for the same thing over and over.

jaakl• 1 month ago

AI reminds me a lot https://en.wikipedia.org/wiki/TRIZ , it is like a good machine implementation of it.

anabis• 1 month ago

He seems to be saying that Claude Code can make discoveries. Does anyone think novel discoveries can be made from systems created by supervised learning only, and attempting to do so?

Claude-Code, which have brought true advances in ... programming. ... these systems have found things that are both novel and good.

WithinReason• 1 month ago

https://www.forbes.com/sites/anishasircar/2026/04/17/ai-solv...

hashta• 1 month ago

I think a lot of deep learning is compositional generalization. Models learn reusable pieces (abstractions, styles, procedures, constraints, etc) and recombine them in ways that may never have appeared as a whole in the training data. So even if the ingredients come from past data the final composition can still be novel in a meaningful sense

doctoboggan• 1 month ago

That is, I would say that creativity requires that the new things generated be Evaluated. Without evaluation, and retention of the best, there is nothing created. The novelty flickers into existence but, if its value is unrecognized, it flickers away and is lost.

I really like the way he frames this here. I think a lot of people in the twitter comments (and maybe a few here) aren't reading past the introduction. He isn't saying AI systems are incapable of creativity and discovery. He is claiming generative AI without a harness is not capable of creativity and discovery. There needs to be some other system that "recognizes the value" of the novel idea and remembers it. He gives examples of where this value recognition step is automated and thus by his definition achieve creativity and discovery in a fully automated system.

jason_oster3• 1 month ago

I agree in general. But making this claim with a narrowed definition of generative AI that doesn't match what is being used in practice is setting up a strawman.

whatever1• 1 month ago

It’s ok, LLMs are useful as they are today. Even if they can never can come up with the next generation of math, physics etc.

Even for humans the brains who managed a step change in thinking are so rare that we literally know them by name.

whiplash4512• 1 month ago

You might be missing that those rare humans were sitting on tons of failed or somewhat useful discoveries made by more “mediocre” humans that history forgot.

vasco• 1 month ago

It's funny to me that one would rate their own takes as "new and possibly controversial". Whatever comes next is read under that light of an author that thinks this about their own thoughts.

And the core point is not even true. They can definitely output novel things that are good - less so but they can and they do. Plenty of examples.

Thus, the trajectory is either novel or good—based on randomness or based on data—but never both at the same time.

This assumes no possible unexplored path yields good results, or said another way, that none of the random results can be good, which is not true. The whole text seems to try to prove a point decided a-priori rather than make a case based on reality.

acosmism• 1 month ago

This reset my thoughts

skor• 1 month ago

yes, when you drive these systems you can get novel output, depends on the effort you put in

musebox353• 1 month ago

The most successful applications like coding are not the result of pure LLM/generative modeling. They come from closing the loop with an agentic harness. The generate-test-selectively refine loop is the core modality of scientific work. An LLM + RL with Verifiable Rewards + feedback from compiler/terminal runs mimics this process to a great extend.

This is Fisher/Box feedback loop (https://www-sop.inria.fr/members/Ian.Jermyn/philosophy/writi...) implemented on a modern computational system. LLM is just a component. I wish Sutton had commented on this fuller picture of what we have now instead of commenting just on the LLM/Backprop side of things. I am honestly curious of whether such a loop can at least partially automate discovery.

There are more elements to discovery though. It is still not clear where the initial working model/hypothesis comes from or how the updates are selected (unless it is just parameter induction). I recently read about Hanson's Patterns of Discovery which aims in that direction. I have still not read it, but I am curious if it has any mechanistic clues.

flir• 1 month ago

Completely agree on the importance of the harness.

The problem I see is the same problem Evolutionary Algorithms had: you can generate potential solutions until you run out of cash, but you still need to evalulate those solutions. You need a fitness function, and that means you need to at least know the general shape of the solution. If anyone knows of any work towards more open-ended fitness functions, I'd love to read it.

piker• 1 month ago

Seems to a layperson like myself that in Math they're using Lean and in programming contexts they're using compilers, such that the models themselves tend towards embedding that determinism "intuitively".

ffwd• 1 month ago

Just some speculation, but, I think humans have on the one hand a lot of degrees of freedom in behaviors and thoughts they can do, but at the same time all that freedom is reigned in by our biological needs, like preserving the integrity of our body, but also preserve the integrity of our minds. But this extends further to preserving our surroundings (for our safety, a changing environment brings uncertainty), but also of people we care about and even entire societies that we have. And preserving our future selves through prediction of future environments.

So all that is to say, I'm not sure it is even theoretically possible to create a single algorithm to do open ended search and evaluation. Biology has billions of years of evolution and accumulation, whereas a simple algorithm in a computer, even if smart and connected to the real world, has no such accumulation.

I think humans hit the perfect sweet spot where we have the simplicity of the self preservation instinct, but we have the complexity of the cortex and lots of degrees of freedom because of it, plus on top of that we have a lot of accumulated degrees of freedom in the society and technology and knowledge that have we, which has been built up for thousands of years, all of which we can't just create an algorithm to encapsulate without going through the actual evolution.

And just to make it explicit - a large percentage of what humans think derives from an instinct to preserve the self, the mind, the future and the environment, even if it is very abstract at times. Not absolutely all, but I think a good chunk. And the complexity and degrees of freedom comes from that we have so many neurons in the brain, and a complex body with hands and whatever else that allows a lot of behaviors, as well as a complex environment that is constantly challenging us.

whiplash4512• 1 month ago

The TRM architecture models both the problem and the solution at the same time. You might find it an interesting read.

https://arxiv.org/abs/2510.04871

visarga3• 1 month ago

If anyone knows of any work towards more open-ended fitness functions, I'd love to read it.

There is research in open-ended learning, see "Why Greatness Cannot Be Planned" by Kenneth O. Stanley. The core idea is that in open-ended scenarios you don't know what action was good except in hindsight because your path is deceptive. So the idea is to replace fitness with novelty search which provides more stepping stones towards the goal.

singularity2001• 1 month ago

Most importantly, the reinforcement loop is used during training. I don't agree with Sutton's original hypothesis, but it holds even less after reinforcement learning.

porridgeraisin6• 1 month ago

RLVR still does not expand beyond the base distribution though, it only mode-seeks within it.

i.e, evaluation, retention yes. variation or "planning" no.

That is not to say you cannot use LLMs. Alpha evolve does exactly that. It uses an external simple evolutionary planner. The overarching point he's making is that our planner is still "dumb" and we need to work on it.

When you iteratively guide an LLM in claude code, you are the external planner. That also works.

visarga3• 1 month ago

There are more elements to discovery though. It is still not clear where the initial working model/hypothesis comes from or how the updates are selected

That is a problem in RL, so we usually do supervised training first, teach it to imitate some trajectories, then do RL to refine the model. RL alone has a huge problem because it might be hard to reach a reward, hence hard to learn the task by pure reinforcement. Humans also combine supervision (learn from books) with search (solving problems) to break the discovery problem. For example, a human with no initial instruction in math would not produce great results no matter how smart they are. The bootstrap was exploration paid for in the past.

anthonypasq3• 1 month ago

the role of evolution is always a confounding factor as well and all the various analogies to how it maps onto AI research are always not quite satisfactory.

musebox353• 1 month ago

SFT + RL connection to model/hypothesis search is insightful. Brute force / scalable search is where Sutton's Bitter Lesson also points to. Once your search domain is small compared to your search budget, that makes a lot of sense.

If I get your meaning right, SFT creates the right inductive bias so that the RL search + reward guidance does the trick.

For novel discovery, the question might then be whether the inductive bias builds a strong enough prison so no new discovery is possible by RL or if the search can escape the boundaries set by SFT given enough randomization and the right reward function.

I know that RL is usually not performed at inference time, but in-context learning mechanisms might be developed by RL to discover at test time. Edit: I would love to hear if that actually happens or not, like new induction heads (https://transformer-circuits.pub/2022/in-context-learning-an...) forming during RL. I really have no idea.

anthonypasq3• 1 month ago

Yes it seems most anti-LLM researchers take issue with LLMs on fundamental math/architecture based properties, but seem to miss all the engineering going on around the model to make it useful.

Those mathematical shortcomings very well might mean they arent a path to true AGI, but that honestly seems fairly irrelevant at this point tbh.

CGMthrowaway• 1 month ago

The generate-test-selectively refine loop is the core modality of scientific work.

Can you expand on this? Because when I think of it, it calls to mind p-hacking and selective publishing more than anything.

Like, when you try prompts over and over and over until you get code that finally works. At which point you stop and pronounce the dubious claim, "AI is amazing, look what it does!"

est• 1 month ago

Richard talks about AI been either novel or good

Then it shifts to discovery.

These seems related but not exactly the same thing.

sph• 1 month ago

The world will not be satisfied until we have read, and discussed every half-famous person’s opinion on AI.

Still about ten million discussions to go.

yanis_t• 1 month ago

I mean even I know who Sutton is [1]. He is one of the reinforcement learning pioneers.

[1] https://en.wikipedia.org/wiki/Richard_S._Sutton

AndrewKemendo2• 1 month ago

He is a Turing award winner

1dom• 1 month ago

I enjoyed reading this at the start, the language is very... inspiring. By the end, I was disappointed. I don't disagree with what they're saying, but the opening style and statements made me expect some more specific or groundbreaking conclusions.

The point seems to be that generative AI just generates stuff, and that real discovery requires variation, evaluation and selective retention.

The call to arms seems based on the assumption that people only every talk about generative AI as discovery machines themselves. I think it's pretty widely accepted that's not the case by everyone apart from cliche out-of-touch CEOs.

But the talk makes me realise that generative AI are incredible tools to do the discovery cycle with, and this is what I imagine professionally successful AI users are doing: variation, evaluation and selective retention of their inputs and outputs to generative AI.

oulipo2• 1 month ago

I think that creativity / intelligence / humour is indissociable from "making errors".

A joke is just an "error" - your brain predicted something, and the butt of the joke goes in another direction, and it's the mismatch that makes it funny.

The same goes with creativity, and intelligence.

The problem is that, by design, while trying to make machines "reliable", we make it impossible for them to be intelligent and creative

lioeters2• 1 month ago

In addition to the importance of making "mistakes", I would say "surprise" is a big element in creativity and humor. Perhaps these are related concepts, because a mistake is surprising. It's an unexpected departure from the normal routine and habit of behavior or thinking.

The term "surprisal" is used in information theory:

For a given probability space, the measurement of rarer events are intuitively more "surprising", and yield more information content than more "common" events.

Can a machine surprise us? Given enough complexity, I think so. They can produce unpredictable results, even novelty, something we've never seen before. But does that mean a machine can be creative? Or funny? Maybe there's a threshold of acceptance, where eventually its output will become surprising enough that we might as well call it creative.

oedemis• 1 month ago

why no one talks about claude shannon deep understanding of information theory and the consequences for compression and intelligence

jason_oster3• 1 month ago

Can't tell if you are spending time in the wrong circles or just shitposting. Here's one: https://youtu.be/l6DKRf-fAAM?si=6u54mry0TLgS868a

rando77• 1 month ago

I think there is a rich seam of work for computers to help humans spot when there are anomalies in the hypotheses humans hold that can help humans form the corrections to hypothesis that fit our aesthetic sensibilities

Qhemlomo• 1 month ago

Creativity is personal taste and the taste which captures the most is mainstream.

I tell jokes and the group of friends get them, for family they don't get them anymore.

I do not like 'basic jokes' and despite that, german television is full of it.

Most things in our world are more of a challange of finding the answers and not 'creating' the answers.

Math: the answer is already there, you only need to find it an fast space of posibilities. This is perfect for LLMs.

Creativity: a LLM can iterate over things a lot faster than a person. So we can iterate over this space too. We can also get feedback from people, tiktok, instagram.

data_maan• 1 month ago

Another famous dude dumping his thoughts on HN who is gulping it up like an addict.

Add this to the long list of names like Terence Tao, and others who seem to be intellectually incontinent lately in the sense that one cannot navigate this space anymore without encountering their thoughts

premieroncall• 1 month ago

Hmm, so the guy who said more data and compute will outperform any adhoc heuristics has shared a three step adhoc heuristic?

visarga3• 1 month ago

The ad hoc heuristics are the domain knowledge baked into the model by human experts, like features, architecture and loss function.

"Evaluation" means environments or datasets, the model is supposed to discover its representations from scaled up experience. That was the bitter lesson - more data and compute beat heuristics.

dang4• 1 month ago

Can you please make your substantive points thoughtfully and without snark? I'm sure there is one here, but it's hard to make out what it is, and in any case the poison does more harm than the information does good.

I'd link to the HN guidelines here but I'm on my phone!

fragebogen• 1 month ago

Assuming OP meant "the bitter lesson" https://www.cs.utexas.edu/~eunsol/courses/data/bitter_lesson...

dredmorbius• 1 month ago

Alt link: <https://xcancel.com/RichardSSutton/status/206121608774494665...>

(Currently returning 502 "Bad Gateway" for me, but should be restored at some point.)

Rich Sutton on AI creativity and discovery

Related Stories