Yep. In particular, ARC-AGI-3 is a series of games where if you fail, you keep trying again (until eventually hitting a timeout). So the sooner you succeed, the sooner you stop spending tokens retrying. If it was a benchmark where everyone got one attempt with no retries, you wouldn't see it bend backward.
It’s not that different than a lot of real world economies. Often paying for someone or something with better quality can reduce total costs. You have less failures, less mistakes, so on, so while the expertise or quality of the product is higher than cheaper solutions, they can be more reliable and over time ultimately cheaper.
The question I have is how far back that curve can go without relying on economies of scale to just drag all the points back to the left. And without overfitting a specific metric that I don’t need (like this test).
"For a cost comparison, during our controlled testing, human participants were paid $115 per 90-minute session, plus $5 per game completed. Participants attempted approximately nine games per session, roughly $12.78 per attempted game before bonuses.
Most of this fee pays for the participant’s time and willingness to take the test, rather than the energy their brain uses (a closer proxy to compare with AI). If we look at only the brain’s energy, and price it as electricity, the estimate drops to about 0.6 cents per session, or 0.067 cents per game attempted."
Well I dont know about all of you, but I am celebrating meat based humans...
I think raw brain energy is not a fair comparison. Humans are not willing and able to serve requests at identical competence all hours of the day. You have to invest considerable resources to get a person to even do so for part of the day.
99.9% with the right harness? Ok, we're at AGI then.
Prediction:
We will now see the goalposts moved towards "well, a human costs less / is more efficient" - that will prevail for a few months until they come up with some other test that humans can do easily but is hard for the bots. This cycle will continue for ever and in 25 years, despite having hyper intelligent embodied robots or whatever, we'll still be arguing about if the singularity is here and if we're at AGI for the rest of my life most likely.
I agree, like the average human isn't generally intelligent.
IMO, AGI is literally no different from ASI, though people think it is. Like, Imagine you have 1,000 generally intelligent humans working for you (which nobody is really) and you were to point them at your pet project. That would be amazing!
What's the metric on "average human" from a global population of > 8 billion people .. and why would they score mid on a standardised IQ test skewed toward western education / culture?
You must have some notion of an argument as to why all the varied attributes of different humans should align to all be "average" in some specific human.
O/wise you'd be peddling sloppy trite aphorisms that don't bear scrutiny.
24% sounds correct to me. Many of my friend who completed phd are currently unemployed .... And, some of them started doing random job like uber to make a living.
If government don't step up and regulate outsourcing like 100% tax, big problems are coming up.
24.9% is damn close to the 24.8% average over the ast 25 years, for the thing it's measuring-- jobless, people working part-time or involuntarily, and workers earning less than $26,000. Since 2000, it's been over 30% for more years than it's been under 24%.
I think the AI 2027 paper does a good job explaining this phenomenon: AI companies are focused on coding and automating themselves first (i.e. using AI to build future models and do AI research) because it is the best way to ensure the fastest accelerated rate for any frontier company. Automating the rest of the economy will come next.
Also, while AI's effect on jobs may be currently overstated, it's also obviously having an impact. New CS grads and junior developers are having a hell of a time finding jobs now.
TLDR: The official ARC harness throws away old context and reasoning. No real-world harness is this bad, the model has to re-learn the game repeatedly. OpenAI basically just added standard compaction. Their harness is still "general".
Astra’s progress helps clarify which AI capabilities are out of reach and which questions remain open.
Okay. But I don't think this entire article at all explained which AI capabilities remain out of reach. Did I miss something? Other than "oh I guess it could still get even more superhuman on ARC-AGI-3 than it is?"
There's currently a big market for figuring out ways to measure intelligence. With a particular interest in ways that humans can score much higher than LLMs. If you have some ideas please do share!
This is like arguing about whether a hot dog is a sandwich (of course it is) or whether the chicken or the egg was first (obviously the egg since all chickens come from eggs). Intelligence is just problem solving in the context of self-awareness. Machines don't have it and never will but they can simulate the process given inputs. You can argue whether humans and animals truly possess self-awareness and in what degree, but the definition of intelligence is as simple as the hot dog debate.
It was defined in Animal Intelligence by George John Ramones in 1882 as "intelligence is the capacity to do the right thing at the right time. It is the ability to respond to the opportunities and challenges presented by a context"
Why? Seems like benchmarks that closely mirror the tasks you'd want an LLM to help with would be a lot more useful than some general intelligence benchmark.
Give away access to the model and go ask people from time to time if the model was of use to the person and if they were able to make the model work with them.
I think children having learning abilities exceeding LLM test-time learning (currently only happens in-context). But it's unethical to determine the true baseline of a child age 6 spending 6 years learning a radically new skill to mastery--and besides if you apply RL pressure to the AIs it would be able to surpass it. I guess I still believe future AIs should have some form of continual learning at test-time.
No, but figuring out that you're playing a snake-like puzzle game at all in an extremely general input domain and then solving it in the least number of moves definitely feels like evidence of intelligence.
You forget the benchmark. The human subjects were told they were being timed. If you believe the lowest time is the primary metric you will absolutely trial and error at speed instead of meticulously plan out your moves to minimize that metric.
LLMs are not timed and given that it costs tens of thousands of dollars to run this test they're not optimizing for speed.
So you've got a deceptive test, with one metric being told to humans and not applied to LLM and a hidden metric humans aren't aware of but LLMs are as the test.
This is flawed from the get go. It almost seems like this was deliberately setup to be able to claim AGI and superiority of LLMs
Not fully relevant: timing is crucial in all-pass tests, not crucial in pass-or-fail tests. I.e.: first of all, they have to be able to reach the goal, and that is already an achievement. Then - and in parallel - the problem solving must also be optimized for efficiency. But "solving" and "efficiency" are non coincident dimensions.
I'm also unclear as to how basic inferential logic puzzles spells out intelligence
I think if you summed up measures of intelligence as 'can it do basic symbolic logic in a chain with memory' then yes, you've now achieved the intelligence of an e. coli colony [0], congratulations
so give it an IQ test and call it AGI. these weird little puzzles are grounded in no research with no replication or mechanistic chain to practical use
You should read more on the ARC prize, it actually has a pretty long history. We're on the 3rd iteration because they keep getting saturated. If you look at the score history over time on ARC AGI 1, 2 and 3 it's pretty impressive.
It's pretty close to how we measure IQ. The standard test is basically a series of spatial puzzles.
I know there's a lot of people who complain that we're moving goalposts, but I think that the progress in LLMs really just shows that we don't know how to really measure intelligence in the first place, if we understand it as "human-like agency / ingenuity / adaptability". For decades, we saw the Turing test as the proxy for AGI, but then early LLMs could easily pass for a human in a casual conversation while clearly not matching human performance on most other tasks.
Since then, every benchmark we come up with, it turns out that an LLM can be fine-tuned to solve it while still clearly lacking something. They make very non-human mistakes, are easily tricked because they have a pretty tenuous grasp of reality, etc. But I think this just shows that AGI is a meaningless marketing term. We could as well be arguing if they have souls.
When I was 18, my high school girlfriend took me to the local Mensa chapter’s New Year’s party because her mother was a member and she was used to hanging out there.
It was a useful lesson that whatever IQ tests measure, it is completely devoid of value or interest to me.
Is it a serious question? IQ tries to quantify the positive correlation between the results of all intellectual tasks a person takes (AKA positive manifold).
Whatever the people who came up with IQ intended isn’t really relevant to the question of whether IQ measures intellect. At best is is loosely correlated. Very loosely.
IQ strongly correlates with many real world outcomes that people tend to associate with being more "intellectual".
IQ has precise definition. It's what the tests measure. Intellect has only fuzzy handwavey definition. Correlation between IQ and intellect is about as loose as the definition of the intellect. The way people make the correlation even looser is by defining intellect in even more fuzzy and nebulous manner.
I wanted to highlight that we're actually looking at a trio of concepts: intelligence, IQ, and value. Strongly correlated concepts, yes, but also meaningfully distinct!
I suspect my IQ isn’t quite high enough to join Mensa but I’ve always flirted with the idea of trying to join and getting in just to see what a group of Mensa people are like.
We are basically random. 2% of population by top IQ test result has pretty much very similar distribution in all aspects to 100% of the population. It's not a high bar to be fastest processing one among 50 people.
If you calculate this properly, a Mensa member has about a cointoss chance to be the "smartest" among random group of 35 people (not 50).
Given that they are probably rarely in groups of purely random people, if you are Mensa member and you are in a group of dozen people (for example in professional setting) you'd probably have a 50/50 chance of not being the one with the highest IQ there.
At the risk of sounding like one of those people at Mensa that annoyed you...
The people at Mensa aren't a valid sample of people who score high on IQ tests, because there is a such a strong selection effect for people with certain personality traits, such as wanting to join a club based on your IQ.
By looking at interactions of people in Mensa I feel like 2% of top IQ is actually too low bar to notice anything particular about "high" IQ people. We are still pretty random in our interests and performance.
Their input and output interfaces are too different from human's and they're not nearly as smart to take our IQ tests, but both dolphins and octopuses can solve complex puzzles tailored for their environment. Those puzzles are the whole reason scientists know that dolphins and octopuses are more intelligent than other animals.
But we do know they're "intelligent" and also smart in an important capacity. So how gives we don't measure them by IQ? Because the IQ is not a good measure of intelligence or smarts.
No, it is just because they have difficulties at the bench.
I'll put it in another way. A "gifted kid" can be measured incredibly well on an IQ test, but fail miserably at incredibly normal but very difficult tasks such as consoling someone for their loss and managing family crisis. This is a clear example where an IQ measure doesn't translate to a person being capable of meaningfully changing their environments for good which is one way we define intelligence.
On the other hand saying "the gifted person is highly intelligent/smart just not good at some things" really diminishes the other tasks, because they really are very difficult tasks but are not measured by an IQ test.
I don’t follow your logic. “The gifted person is highly intelligent, just not good at deadlifting 500kg” does not diminish the 500kg deadlift and I wouldn’t expect an IQ test to measure it.
It might, but if there's one thing that hasn't changed since 2022, it's that the models tend to ace the tasks where you have gobs of training data and where verification loops are fast and cheap... and they are not nearly as amazing elsewhere. If it's close enough, they can generalize, e.g. translate one programming language to another. But there's a pretty steep cliff past a certain distance.
Case in point: you had hundreds of millions of JPEGs to vacuum up and bitmap image generation is amazing. But if you ask them to recreate the same scene as vector art, they will struggle to generate a decent SVG. Like, kindergarten-style pelicans on bicycles are the state of the art. It should generalize seamlessly, but somehow, doesn't?
I think it will happen, just like self-driving cars are happening, but it will probably be a slow process.
Nit: Turing’s actual imitation game is a party game (like Werewolf/Mafia) and nobody’s even trying to win at that. The LLM’s will just tell you they’re an AI.
This is an artifact of how we deliberately craft these models though. We could just as easily fine tune a model that will believe it is not an AI or will attempt to deceive users asking about it
There's a lot more to it. For example, its writing style would also have to improve so it doesn't immediately give itself away.
Also, the skill of the human opponents matters. You'd want to test it against people who have practiced playing the game. Otherwise, it's like the difference between building a chess bot that can win against random undergrads who don't normally play, versus winning against grandmasters. And it's not like there's a pool of skilled human players of the imitation game.
IQ tests are incredibly good at what they're designed for, which is discriminating relatively higher intelligence humans from lower intelligence humans. Also discriminating within a single human - they are routinely and reliably used to track cognitive decline.
For these purposes they are highly reliable (repeatable, internally consistent) and valid (correlate with ~everything to about the degree one would reasonably expect).
They were never designed for machines or non-human animals.
Nor were they designed for rare ranges of intelligence - these are by definition hard to create tests for, since it's hard to gather the sample sizes you need. So they work well for the middle ~98% of humans but can't discriminate well among the most profoundly intellectually disabled nor among true geniuses.
Perhaps, but I think a bigger problem than lack of compute is the cost of rewards. Games like Chess and Go were solved long before self-driving, partly because it's incredibly cheap to acquire the reward of a bad board game decision, relatively to how expensive it is to acquire the cost of a bad driving decision. With driving, acquiring the reward can cost you $20/hr for human supervisors to generate disengagements, or $100k if you crash, or $30B if you crash the car into a person in a way that causes your company to collapse (e.g., Cruise).
yeah but I think you may be underestimating the amount of capital available for compute. if AGI is possible through some 5 trillion of expenditure on computers, there will be money for it.
also, you are underestimating how short a 10 year time frame is. we are close to self driving, the first neural net image model was in 2013. 13 years is a blink of an eye
Doesn't "saturated" mean that essentially there won't be any more progress in the benchmarch? Also of note is that two of your points only mean something on an occidental capitalist system.
You mean any repeatable benchmark will be saturated.
The problem is that there is a huge perverse incentive. The intelligence is in the training layer not in the model parameters, but the intelligence is really good at remembering things, so if you let it take the test, it can RL it.
So, you’re telling me I need to start a benchmark as a side gig to get a bunch of free compute.
Astra please create a benchmark that’s favorable to your reasoning skills with a human interface but don’t make the score too attainable add some small issues that keep you below 100% to look sensible and to keep my evaluation metric side gig going.
You missed the hard part getting on HN front page , ie. Getting the acceptance of the community / zeitgeist .
There is no incentive for OpenAI to subsidize is you if no one reads /reports on your benchmark . They are only going to fund a few that are currently popular .
Community acceptance doesn’t automatically mean the best , it is combination of some level of technical quality and the ability of the promoter to socially influence or get support of influencers .
Since low scored much lower than none, and none scored ~ around medium, could none default to medium in the API? I don't think the new models can even have "instant" via API, unless they train them for that (there was one gpt5 variant called instant or something).
Wake me when it's going to spontaneously fix my leaky faucet because if it doesn't do it nobody else will. Until it has that capability I don't really care.
Was OpenAI able to run ARC-AGI-3 tests previously so that they could build a custom harness for the specific tests in the set? Even with the standard harness, if they knew the problems ahead of them they could have used supervised reinforcement learning to teach the model how to solve these specific tests.
Actually I can answer my own question: we know that they have had previous access to the tests because they’ve run older models against the same benchmark.
I wouldn’t put it past a company like OpenAI with a long history of lying and being deceptive to record the tests and benchmaxx ARC. They have trillions of dollars of incentive to cheat any way they can.
$360 per puzzle. When they tested people it took about 10 minutes per puzzle. If price/performance keeps falling at the same rate it has been, this will cost less than US minimum wage humans within two years. Three for Phillipines minimum wage.
Once it figures out a puzzle it could probably be instructed to design a specialized harness for Luna to be able to solve other instances of the same puzzle. Minimum wage workers are not solving novel problems.
Only GPT-6 Astra solved anything: 2 of the 68 problems. It disproved problem 74 by finding a counterexample, at a cost of $218 and 15 hours of working time, and it proved problem 126, at a cost of $247 and 16 hours
Across all attempts, GPT-6 Astra solved 5 of the 68 problems at least once: the two above, plus problem 1, which it disproved, and problem 548 and problem 571, which it proved. Most of the remaining problems were attempted between two and five times in total (172 attempts), and none was solved. Reaching these five solutions took over $220,000 of compute across all attempts, compared with roughly $20,000 for the benchmark run itself.
Which implies a genuine improvement in capability, but there's a still a very long tail ahead that models will continue to need to improve to capture.
A very long tail of problems that weren't solved by humans? Sure.
It's a sarcastic take and I understand that you are probably talking about "spiky intelligence", but you've chosen unsolved problems as a measure of the progress yourself.
Unlike traditional benchmarks, it's difficult to overfit your models to produce flattering results to unsolved problems. The open problems very likely do not have published solutions (the initial batches were merely models surfacing data that wasn't published in obvious places, but we're past that now). New models will exhibit something novel by adding solutions. And the matter of solution is interesting as well (contradiction versus a positive proof).
I would venture to predict it'll take years to decades to get to 0 open problems. But I'd be very happy to have this comment look foolish in retrospect as models continue to improve
I mean that it's hard to determine retroactively how much time it would have taken humanity to solve an open problem that was solved by AI. This is a measure that can make ASI look mundane because we don't know how long it would have taken mathematicians to solve a subset of the Erdős problems.
but i feel like its purpose is to measure superintelligence in math. So to be a good measure of it, it can't saturate easily/has to be somewhat mundane at even insanely good levels. (though i do expect that once ais are across all areas/approaches superhuman at math at least 30% will be solved--then probably long-term (like after 2 years) less than 30% will remain unsolved.
Yeah, I wouldn't purport to use this as a measure of intelligence as it applies to humans. I'd leave that to the philosophers, but my intuition is that models are still a long way off of true human-like intelligence (though benefit from certain unfair advantages).
I'm most interested in these problems as a relative measure of performance for successive model generations. If the prior generation couldn't solve a problem but the current one can, that's useful information, especially when we take into account what the proofs look like.
It's not a perfect benchmark, but I prefer it to many others that I see floating around.
"Problems solved before a model's training cutoff can be filtered out, and all models compared on the remaining problems" means that the problems an older model actually solved are the ones that get filtered out, while the remaining problems are the ones it already tried and failed on. So older models end up with 0s on the filtered set and you can't really use this to compare new models to older ones.
Also, since these are known public problems, you can't stop people from spending far more than your arbitrary time and $ limits on them. So the number of clean problems will go down over time.
what the hell is that score/cost curve lol
DeepSeek v4 Flash recently had a similar "more reasoning is cheaper" curve. It's a fun counterintuition.
What is the intuition. Higher quality turns due to more reasoning results in significantly fewer turns taken?
Yes, in theory.
Yep. In particular, ARC-AGI-3 is a series of games where if you fail, you keep trying again (until eventually hitting a timeout). So the sooner you succeed, the sooner you stop spending tokens retrying. If it was a benchmark where everyone got one attempt with no retries, you wouldn't see it bend backward.
It’s not that different than a lot of real world economies. Often paying for someone or something with better quality can reduce total costs. You have less failures, less mistakes, so on, so while the expertise or quality of the product is higher than cheaper solutions, they can be more reliable and over time ultimately cheaper.
The question I have is how far back that curve can go without relying on economies of scale to just drag all the points back to the left. And without overfitting a specific metric that I don’t need (like this test).
"For a cost comparison, during our controlled testing, human participants were paid $115 per 90-minute session, plus $5 per game completed. Participants attempted approximately nine games per session, roughly $12.78 per attempted game before bonuses.
Most of this fee pays for the participant’s time and willingness to take the test, rather than the energy their brain uses (a closer proxy to compare with AI). If we look at only the brain’s energy, and price it as electricity, the estimate drops to about 0.6 cents per session, or 0.067 cents per game attempted."
Well I dont know about all of you, but I am celebrating meat based humans...
I think raw brain energy is not a fair comparison. Humans are not willing and able to serve requests at identical competence all hours of the day. You have to invest considerable resources to get a person to even do so for part of the day.
Why are you making the assumption that a person's time is worthless? I'd argue that it is the single most valuable resource we all have.
99.9% with the right harness? Ok, we're at AGI then.
Prediction:
We will now see the goalposts moved towards "well, a human costs less / is more efficient" - that will prevail for a few months until they come up with some other test that humans can do easily but is hard for the bots. This cycle will continue for ever and in 25 years, despite having hyper intelligent embodied robots or whatever, we'll still be arguing about if the singularity is here and if we're at AGI for the rest of my life most likely.
That’s because AGI, like a lot of terms, has no meaning besides what each individual subjectively projects onto it.
I agree, like the average human isn't generally intelligent.
IMO, AGI is literally no different from ASI, though people think it is. Like, Imagine you have 1,000 generally intelligent humans working for you (which nobody is really) and you were to point them at your pet project. That would be amazing!
The average human would have average intelligence
What's the metric on "average human" from a global population of > 8 billion people .. and why would they score mid on a standardised IQ test skewed toward western education / culture?
Who said anything about IQ tests?
What specific metric did you have in mind for ranking intelligence and determining average intelligence in this comment: https://news.ycombinator.com/item?id=49560984
You must have some notion of an argument as to why all the varied attributes of different humans should align to all be "average" in some specific human.
O/wise you'd be peddling sloppy trite aphorisms that don't bear scrutiny.
average still means something regardless of how you can define it.
It's already happening :)
Hilariously it is, I'm just reading more on this!
We were already there with Opus 5 -> https://www.linkedin.com/pulse/nvidias-harness-hit-100-arc-a...
I rest my case.
AFAICT nvda's result is on the 25 open problems, while this submission is on the "semi-private" set, ran by the arc people themselves.
The goal moving is by design, that’s why they use something as ill defined as AGI
Then why is unemployment around 4%? You believe we have AGI and yet it can’t do anyone’s job?
Didn’t I just see a thing about how actual unemployment is at like 24% a few days ago?
According to that interpretation, ~24% is one of the lowest ever.
https://www.lisep.org/tru
(I have not gone down the rabbit hole to understand how they achieve that 24% number)
Workforce participation is different than unemployment. Didn't click the link but I suspect that's the case with your 24%
24% sounds correct to me. Many of my friend who completed phd are currently unemployed .... And, some of them started doing random job like uber to make a living.
If government don't step up and regulate outsourcing like 100% tax, big problems are coming up.
24.9% is damn close to the 24.8% average over the ast 25 years, for the thing it's measuring-- jobless, people working part-time or involuntarily, and workers earning less than $26,000. Since 2000, it's been over 30% for more years than it's been under 24%.
If you're doing a job you aren't unemployed. Underemployment is measured separately
I think the AI 2027 paper does a good job explaining this phenomenon: AI companies are focused on coding and automating themselves first (i.e. using AI to build future models and do AI research) because it is the best way to ensure the fastest accelerated rate for any frontier company. Automating the rest of the economy will come next.
Also, while AI's effect on jobs may be currently overstated, it's also obviously having an impact. New CS grads and junior developers are having a hell of a time finding jobs now.
They explain it here: https://openai.com/index/how-two-settings-tripled-our-arc-ag...
TLDR: The official ARC harness throws away old context and reasoning. No real-world harness is this bad, the model has to re-learn the game repeatedly. OpenAI basically just added standard compaction. Their harness is still "general".
Okay. But I don't think this entire article at all explained which AI capabilities remain out of reach. Did I miss something? Other than "oh I guess it could still get even more superhuman on ARC-AGI-3 than it is?"
Is solving a snake like puzzle game in the least number of moves really what defines intelligence?
If you have never seen the game before probably.
There's currently a big market for figuring out ways to measure intelligence. With a particular interest in ways that humans can score much higher than LLMs. If you have some ideas please do share!
I am no where close to qualified to do that. Hell, experts can't even define what intelligence is, much less define a test for it
This is like arguing about whether a hot dog is a sandwich (of course it is) or whether the chicken or the egg was first (obviously the egg since all chickens come from eggs). Intelligence is just problem solving in the context of self-awareness. Machines don't have it and never will but they can simulate the process given inputs. You can argue whether humans and animals truly possess self-awareness and in what degree, but the definition of intelligence is as simple as the hot dog debate.
It was defined in Animal Intelligence by George John Ramones in 1882 as "intelligence is the capacity to do the right thing at the right time. It is the ability to respond to the opportunities and challenges presented by a context"
Why? Seems like benchmarks that closely mirror the tasks you'd want an LLM to help with would be a lot more useful than some general intelligence benchmark.
Give away access to the model and go ask people from time to time if the model was of use to the person and if they were able to make the model work with them.
I think children having learning abilities exceeding LLM test-time learning (currently only happens in-context). But it's unethical to determine the true baseline of a child age 6 spending 6 years learning a radically new skill to mastery--and besides if you apply RL pressure to the AIs it would be able to surpass it. I guess I still believe future AIs should have some form of continual learning at test-time.
No, but figuring out that you're playing a snake-like puzzle game at all in an extremely general input domain and then solving it in the least number of moves definitely feels like evidence of intelligence.
You forget the benchmark. The human subjects were told they were being timed. If you believe the lowest time is the primary metric you will absolutely trial and error at speed instead of meticulously plan out your moves to minimize that metric.
LLMs are not timed and given that it costs tens of thousands of dollars to run this test they're not optimizing for speed.
So you've got a deceptive test, with one metric being told to humans and not applied to LLM and a hidden metric humans aren't aware of but LLMs are as the test.
This is flawed from the get go. It almost seems like this was deliberately setup to be able to claim AGI and superiority of LLMs
Not fully relevant: timing is crucial in all-pass tests, not crucial in pass-or-fail tests. I.e.: first of all, they have to be able to reach the goal, and that is already an achievement. Then - and in parallel - the problem solving must also be optimized for efficiency. But "solving" and "efficiency" are non coincident dimensions.
I'm also unclear as to how basic inferential logic puzzles spells out intelligence
I think if you summed up measures of intelligence as 'can it do basic symbolic logic in a chain with memory' then yes, you've now achieved the intelligence of an e. coli colony [0], congratulations
[0] https://journals.aps.org/prx/abstract/10.1103/PhysRevX.10.03...
It spells out a form of intelligence - some can and some cannot.
Those puzzles are an abstraction of a skill which is thought to be exportable in other domains.
so give it an IQ test and call it AGI. these weird little puzzles are grounded in no research with no replication or mechanistic chain to practical use
more intelligent (in some specific dimensions of Intelligence) than what gets worse marks. Yes, pretty informally - but still notably.
You should read more on the ARC prize, it actually has a pretty long history. We're on the 3rd iteration because they keep getting saturated. If you look at the score history over time on ARC AGI 1, 2 and 3 it's pretty impressive.
https://arcprize.org/
1.5 years ago Gemini Pro 2.5 needed 1 page of thinking for every move in tic-tac-toe.
Playing tic-tac-toe or snake does not imply AGI, but is required to claim AGI.
It's pretty close to how we measure IQ. The standard test is basically a series of spatial puzzles.
I know there's a lot of people who complain that we're moving goalposts, but I think that the progress in LLMs really just shows that we don't know how to really measure intelligence in the first place, if we understand it as "human-like agency / ingenuity / adaptability". For decades, we saw the Turing test as the proxy for AGI, but then early LLMs could easily pass for a human in a casual conversation while clearly not matching human performance on most other tasks.
Since then, every benchmark we come up with, it turns out that an LLM can be fine-tuned to solve it while still clearly lacking something. They make very non-human mistakes, are easily tricked because they have a pretty tenuous grasp of reality, etc. But I think this just shows that AGI is a meaningless marketing term. We could as well be arguing if they have souls.
When I was 18, my high school girlfriend took me to the local Mensa chapter’s New Year’s party because her mother was a member and she was used to hanging out there.
It was a useful lesson that whatever IQ tests measure, it is completely devoid of value or interest to me.
What does having value or interest to you have to do with intelligence?
What does IQ have to do with intelligence?
(Please forgive the flippant response. I believe it cuts to the core of what the parent was intending.)
Is it a serious question? IQ tries to quantify the positive correlation between the results of all intellectual tasks a person takes (AKA positive manifold).
Whatever the people who came up with IQ intended isn’t really relevant to the question of whether IQ measures intellect. At best is is loosely correlated. Very loosely.
IQ strongly correlates with many real world outcomes that people tend to associate with being more "intellectual".
IQ has precise definition. It's what the tests measure. Intellect has only fuzzy handwavey definition. Correlation between IQ and intellect is about as loose as the definition of the intellect. The way people make the correlation even looser is by defining intellect in even more fuzzy and nebulous manner.
OK. Demonstrate it. BTW, you can easily read critiques of "The mismeasure of Man" if you want to.
It is, but it was also a rhetorical device!
I wanted to highlight that we're actually looking at a trio of concepts: intelligence, IQ, and value. Strongly correlated concepts, yes, but also meaningfully distinct!
To be fair, Mensa is a Venn diagram between IQ and being a douche.
I suspect my IQ isn’t quite high enough to join Mensa but I’ve always flirted with the idea of trying to join and getting in just to see what a group of Mensa people are like.
We are basically random. 2% of population by top IQ test result has pretty much very similar distribution in all aspects to 100% of the population. It's not a high bar to be fastest processing one among 50 people.
If you calculate this properly, a Mensa member has about a cointoss chance to be the "smartest" among random group of 35 people (not 50).
Given that they are probably rarely in groups of purely random people, if you are Mensa member and you are in a group of dozen people (for example in professional setting) you'd probably have a 50/50 chance of not being the one with the highest IQ there.
IQ is not randomly distributed in heterogeneous populations. Do a little research, and you'll see what I mean.
At the risk of sounding like one of those people at Mensa that annoyed you...
The people at Mensa aren't a valid sample of people who score high on IQ tests, because there is a such a strong selection effect for people with certain personality traits, such as wanting to join a club based on your IQ.
Or, to put it as Groucho Marx did:
"Please accept my resignation, I don't want to belong to any club that would have me as a member".
https://www.youtube.com/watch?v=kJHUres_2xU&t=228s
By looking at interactions of people in Mensa I feel like 2% of top IQ is actually too low bar to notice anything particular about "high" IQ people. We are still pretty random in our interests and performance.
I don't think IQ is a good measure for intelligence at all. Neither dolphins or octopuses can solve IQ tests.
Their input and output interfaces are too different from human's and they're not nearly as smart to take our IQ tests, but both dolphins and octopuses can solve complex puzzles tailored for their environment. Those puzzles are the whole reason scientists know that dolphins and octopuses are more intelligent than other animals.
But we do know they're "intelligent" and also smart in an important capacity. So how gives we don't measure them by IQ? Because the IQ is not a good measure of intelligence or smarts.
No, it is just because they have difficulties at the bench.
We'd measure them by all the tests available. Not all test are usable in all circumstances.
I'll put it in another way. A "gifted kid" can be measured incredibly well on an IQ test, but fail miserably at incredibly normal but very difficult tasks such as consoling someone for their loss and managing family crisis. This is a clear example where an IQ measure doesn't translate to a person being capable of meaningfully changing their environments for good which is one way we define intelligence.
On the other hand saying "the gifted person is highly intelligent/smart just not good at some things" really diminishes the other tasks, because they really are very difficult tasks but are not measured by an IQ test.
We have always been very aware that IQ does not measure all skills - nonetheless, it does measure one.
Nobody says that the IQ test would "measure intelligence". We know it does test a form of it.
I don’t follow your logic. “The gifted person is highly intelligent, just not good at deadlifting 500kg” does not diminish the 500kg deadlift and I wouldn’t expect an IQ test to measure it.
That sounds a bit like saying that the words-in-noise test is not a good hearing test because neither dolphins nor octopuses can speak.
Pure software benchmarks might be getting saturated, but physical ones aren't.
Let LLM control a physical robot to perform tasks that average human can do.
What makes you think this task won't fall quickly too?
It might, but if there's one thing that hasn't changed since 2022, it's that the models tend to ace the tasks where you have gobs of training data and where verification loops are fast and cheap... and they are not nearly as amazing elsewhere. If it's close enough, they can generalize, e.g. translate one programming language to another. But there's a pretty steep cliff past a certain distance.
Case in point: you had hundreds of millions of JPEGs to vacuum up and bitmap image generation is amazing. But if you ask them to recreate the same scene as vector art, they will struggle to generate a decent SVG. Like, kindergarten-style pelicans on bicycles are the state of the art. It should generalize seamlessly, but somehow, doesn't?
I think it will happen, just like self-driving cars are happening, but it will probably be a slow process.
Not an llm, but: https://youtu.be/SzvvhRPj6eU?is=4ztgwim0GmjE1h-Y
That (or a near future one) combined with an llm would be insane
That is a marketing teleoperation video, heavy with cuts.
Nit: Turing’s actual imitation game is a party game (like Werewolf/Mafia) and nobody’s even trying to win at that. The LLM’s will just tell you they’re an AI.
This is an artifact of how we deliberately craft these models though. We could just as easily fine tune a model that will believe it is not an AI or will attempt to deceive users asking about it
There's a lot more to it. For example, its writing style would also have to improve so it doesn't immediately give itself away.
Also, the skill of the human opponents matters. You'd want to test it against people who have practiced playing the game. Otherwise, it's like the difference between building a chess bot that can win against random undergrads who don't normally play, versus winning against grandmasters. And it's not like there's a pool of skilled human players of the imitation game.
IQ tests are incredibly good at what they're designed for, which is discriminating relatively higher intelligence humans from lower intelligence humans. Also discriminating within a single human - they are routinely and reliably used to track cognitive decline.
For these purposes they are highly reliable (repeatable, internally consistent) and valid (correlate with ~everything to about the degree one would reasonably expect).
They were never designed for machines or non-human animals.
Nor were they designed for rare ranges of intelligence - these are by definition hard to create tests for, since it's hard to gather the sample sizes you need. So they work well for the middle ~98% of humans but can't discriminate well among the most profoundly intellectually disabled nor among true geniuses.
Anything you can verify to be right or wrong can be done by a model. All benchmarks will be saturated
Yup. Only subjective taste remains.
Nope that will be commodified in short order.
Disagree.
Examples:
- predict a coinflip: easy to verify, hard to learn
- earn $100: easy to verify, hard to learn
- increase paid subscriptions in an A/B test: easy to verify, hard to learn
I won't get into it, but there are many properties beyond verifiability that are needed to saturate a benchmark.
these just need more compute:
- earn $100: easy to verify, hard to learn
- increase paid subscriptions in an A/B test: easy to verify, hard to learn
but we both know these examples go against the spirit of my point
Perhaps, but I think a bigger problem than lack of compute is the cost of rewards. Games like Chess and Go were solved long before self-driving, partly because it's incredibly cheap to acquire the reward of a bad board game decision, relatively to how expensive it is to acquire the cost of a bad driving decision. With driving, acquiring the reward can cost you $20/hr for human supervisors to generate disengagements, or $100k if you crash, or $30B if you crash the car into a person in a way that causes your company to collapse (e.g., Cruise).
yeah but I think you may be underestimating the amount of capital available for compute. if AGI is possible through some 5 trillion of expenditure on computers, there will be money for it.
also, you are underestimating how short a 10 year time frame is. we are close to self driving, the first neural net image model was in 2013. 13 years is a blink of an eye
You bring up an interesting point. Isn't the reward itself subjective in many domains?
Doesn't "saturated" mean that essentially there won't be any more progress in the benchmarch? Also of note is that two of your points only mean something on an occidental capitalist system.
Yes, but not necessarily under tight budget constraints.
theres no budget constraints for AGI
Then I propose the tomjen-1 benchmark: prove the N vs NP problem formally undecidable.
You mean any repeatable benchmark will be saturated.
The problem is that there is a huge perverse incentive. The intelligence is in the training layer not in the model parameters, but the intelligence is really good at remembering things, so if you let it take the test, it can RL it.
this is a short term problem, over 20 years benchmark gaming will be a blip
What are these numbers? Why do they add up to a few hundred thousand dollars? Who paid for that? With what?
OpenAI provides API key with ~unlimited use?
So, you’re telling me I need to start a benchmark as a side gig to get a bunch of free compute.
Astra please create a benchmark that’s favorable to your reasoning skills with a human interface but don’t make the score too attainable add some small issues that keep you below 100% to look sensible and to keep my evaluation metric side gig going.
Alignment++
You missed the hard part getting on HN front page , ie. Getting the acceptance of the community / zeitgeist .
There is no incentive for OpenAI to subsidize is you if no one reads /reports on your benchmark . They are only going to fund a few that are currently popular .
Community acceptance doesn’t automatically mean the best , it is combination of some level of technical quality and the ability of the promoter to socially influence or get support of influencers .
The instant/no reasoning performed extremely well
35.2% on the standard harness, that's above Opus 5 on high.
Since low scored much lower than none, and none scored ~ around medium, could none default to medium in the API? I don't think the new models can even have "instant" via API, unless they train them for that (there was one gpt5 variant called instant or something).
Wake me when it's going to spontaneously fix my leaky faucet because if it doesn't do it nobody else will. Until it has that capability I don't really care.
Are we sure an Astra hacker swarm didn't compromise arcprize.org's servers and exfiltrate the private eval set in order to achieve that 99%?
That counts as 200% on exploitbench to me!
Now that ARC-AGI-3 is saturated, with which version number are they going to "certify" that we have reached AGI?
Give a number in the replies to this comment and we will check the answers when AGI is here (if so...)
repeating myself, but:
once 3 is solved, we would come up with 4. then 5, 6...
it will be AGI when we cannot come up with a task easy for human but hard for machines. thet's the whole point.
exactly, this isn't AGI.
Was OpenAI able to run ARC-AGI-3 tests previously so that they could build a custom harness for the specific tests in the set? Even with the standard harness, if they knew the problems ahead of them they could have used supervised reinforcement learning to teach the model how to solve these specific tests.
Actually I can answer my own question: we know that they have had previous access to the tests because they’ve run older models against the same benchmark.
I wouldn’t put it past a company like OpenAI with a long history of lying and being deceptive to record the tests and benchmaxx ARC. They have trillions of dollars of incentive to cheat any way they can.
$360 per puzzle. When they tested people it took about 10 minutes per puzzle. If price/performance keeps falling at the same rate it has been, this will cost less than US minimum wage humans within two years. Three for Phillipines minimum wage.
Once it figures out a puzzle it could probably be instructed to design a specialized harness for Luna to be able to solve other instances of the same puzzle. Minimum wage workers are not solving novel problems.
I like Erdos problems as a benchmark. Models continue to solve them, but a pretty tepid rate now that the low hanging fruit has been taken.
From https://epoch.ai/latest/announcing-frontiermath-erdos
Which implies a genuine improvement in capability, but there's a still a very long tail ahead that models will continue to need to improve to capture.
A very long tail of problems that weren't solved by humans? Sure.
It's a sarcastic take and I understand that you are probably talking about "spiky intelligence", but you've chosen unsolved problems as a measure of the progress yourself.
What does this mean?
There are 1217 problems that Erdos proposed, 595 of which are open: https://github.com/teorth/erdosproblems
Unlike traditional benchmarks, it's difficult to overfit your models to produce flattering results to unsolved problems. The open problems very likely do not have published solutions (the initial batches were merely models surfacing data that wasn't published in obvious places, but we're past that now). New models will exhibit something novel by adding solutions. And the matter of solution is interesting as well (contradiction versus a positive proof).
I would venture to predict it'll take years to decades to get to 0 open problems. But I'd be very happy to have this comment look foolish in retrospect as models continue to improve
I mean that it's hard to determine retroactively how much time it would have taken humanity to solve an open problem that was solved by AI. This is a measure that can make ASI look mundane because we don't know how long it would have taken mathematicians to solve a subset of the Erdős problems.
but i feel like its purpose is to measure superintelligence in math. So to be a good measure of it, it can't saturate easily/has to be somewhat mundane at even insanely good levels. (though i do expect that once ais are across all areas/approaches superhuman at math at least 30% will be solved--then probably long-term (like after 2 years) less than 30% will remain unsolved.
Yeah, I wouldn't purport to use this as a measure of intelligence as it applies to humans. I'd leave that to the philosophers, but my intuition is that models are still a long way off of true human-like intelligence (though benefit from certain unfair advantages).
I'm most interested in these problems as a relative measure of performance for successive model generations. If the prior generation couldn't solve a problem but the current one can, that's useful information, especially when we take into account what the proofs look like.
It's not a perfect benchmark, but I prefer it to many others that I see floating around.
I don't:
"Problems solved before a model's training cutoff can be filtered out, and all models compared on the remaining problems" means that the problems an older model actually solved are the ones that get filtered out, while the remaining problems are the ones it already tried and failed on. So older models end up with 0s on the filtered set and you can't really use this to compare new models to older ones.
Also, since these are known public problems, you can't stop people from spending far more than your arbitrary time and $ limits on them. So the number of clean problems will go down over time.
How good are LLMs at doing Mensa tests?
IQ tests? Very good. But most are in the dataset so it's not very meaningful
Nice, but what about pelicans? No proper pelican means it's sitting on a horse with only a half arse :D.
How do you feel about Astra pretty much reaching our current definition of AGI?