Comment by ComplexSystems
1 day ago
It's also "out-brute forcing them." It just never gets tired. If a mathematician picks a research direction and spends a whole week on it and it doesn't pan out, they will likely be annoyed, need a break for a while, etc. This thing just does not ever get tired or discouraged or care; it's just onto the next thing until something ends up working.
AI has sitzfleisch
https://en.wikipedia.org/wiki/J._Robert_Oppenheimer#:~:text=...
My browser doesn't do the fancy URL text selection...
sitzfleisch: the ability to endure or carry on with an activity
Something Oppenheimer did not have, apparently.
loaned from German, where it's originally a way to say buttocks, literally "sitting flesh". If you have more Sitzfleisch you can sit for longer. Both in the literal sense (a bigger butt makes sitting more comfortable) and in the figurative sense (having the mental ability to sit for longer, get more desk work done)
2 replies →
Don’t worry, the link provided by the GP did nothing to explain what was meant anyway.
This is fantastic. Now I have a sophisticated-sounding german word for my attention deficit.
There's also a less flattering reading of the word, where Sitzfleisch means having a "flat ass" (from sitting too much, e. g. Sitzfleischparade describing a group of flat-arsed people, or something like Sitzfleischmaxxer, and so on).
2 replies →
in Gujarati its "Gand vagar no loto". ( Utensil with Round Ass ). I guess every language as its equivalent.
Brazilian portuguese has cu de ferro, which literally translates to iron ass, something one must of course have in order to apply themselves to scholarly activities.
This was such an interesting thing to learn
We Americans call it grit.
The connotation of grit to me feels different. Grit is pushing through adversity, rather than just tedium.
2 replies →
That's just life without a bidet.
Wow, what a great comparison. LLMs are great at reasoning but absolute dogshit at simple arithmetic. If there's a raw calculation involved I always tell it to use python to add it all up.
The abacus seems to date back around 4500 years.
If your goal is to implement an absurd comment you can use this one next time: instruct your LLM to implement Conway’s Game of Life to implement an abacus.
Because, you know, humans are notoriously poor at implementing the x86-64 instruction set in their minds. This is why God had to create Guido van Rossum.
Arguably Mathematica would be a better fit, but there are those who frequent this corner of the Internet who rather not have to read anything that might cause them to think about Stephen Wolfram.
So I won’t mention it.
Ever heard of string theory.
People go whole lives without being able to make it pan out.
String theory is a great example of a dead end kept alive by ego and sunk cost fallacy. An AI would have declared it dead and moved on 10 years earlier.
When people make these comments about string theory, are they at the forefront of quantum physics theory and have spent years on modern string theory? Or did they just watch a YouTube video and then keep parroting this?
19 replies →
> It's also "out-brute forcing them." It just never gets tired. If a mathematician picks a research direction and spends a whole week on it and it doesn't pan out, they will likely be annoyed, need a break for a while, etc. This thing just does not ever get tired or discouraged or care; it's just onto the next thing until something ends up working.
> String theory is a great example of a dead end kept alive by ego and sunk cost fallacy. An AI would have declared it dead and moved on 10 years earlier.
Not only are LLMs perfect machines with all the intelligence of humanity without any of our problems, they are also everything else. I wait to get my hands on one of those LLMs people on hn seem to be using. I want to believe too. Let me into the religion of the perfect thinking machine gods.
Ok Sabina, everyone is string theory is all just a bunch of egotistical morons.
24 replies →
even if its not particularly relevant to the physical world around us, it still explored cool math
Also out-speeding them, and that was before high speed inference.
Out-ralphing them, you might say!
https://ghuntley.com/ralph/
AGI ≈ artificial stupidity × infinite persistence
> It's also "out-brute forcing them."
That is also approximately what people have always done to succeed.
Mathematicians routinely spend years on a problem without getting anywhere.
It helps though if you have many problems to work on
You don't, at least traditionally in math research.
This will almost surely change with AI.
The key here is that it’s depending on the human inability to connect the sum of relevant knowledge, but said knowledge comes from humans.
Theres going to be this field day of low-hanging fruit that ML can round up, but after that I suspect it will be in fits and starts as a “connection maker” rather than some proof producer.
I think we're underestimating just how much low hanging fruit there is. I've been trying to apply this LLM research process to physics (QM and solid state) and there is so much missing in Physlib and the rest of the Lean ecosystem that most of my work has been trying to formalize the theories and validating them against the specification problem (and mostly failing badly).
It's not only going to be "connection maker". If and when robotics advance to a point where the LLMs are embodied, they can run experiments in the physical world and find new knowledge.
Being embodied is not the important barrier to running experiments. It's having access to a body of resources (i.e. funding and infrastructure).
7 replies →
Robots in labs already exist, but mercifully they're not hooked up to anything as unpredictable as an LLM. Robots tend to work best as specialists doing high-throughput, extremely repetitive tasks which nonetheless require a degree of precision. Giving a robot a "human" body makes very little sense if we're talking about the needs and productivity of a non-human; humanoid robots are marketing for humans.
19 replies →
Counterpoint: something has happened in frontier models, and yes they now get discouraged and will sometimes prefer to not continue working on a problem unless you tell them to anyway.
I don't know how or why this would be trained on behavior, but no, it isn't true anymore that models don't say things like, "Ugh," or "this is going to take hours and maybe we should stop here."
I think it's to combat runaway token usage.
We're just going to slowly deconstruct every element that could be a factor of intelligence.
It's not out-thinking, it's just out-remembering
It's not out-thinking, it's just out-working
It's not out-thinking, it's just able to consider more things simultaneously
It's not creative, it's just randomly generating things and then selecting viable ones
A new technology being able to do something better than humans does not mean it’s intelligent though. A calculation program is not intelligent just because it can remember more digits than me, work more than me
That's insightful. AI is teaching us things about our own intelligence by simply evolving under our eyes.
If you make that list comprehensive, there's probably a Nobel prize in it for you.
But the difference really does matter and is not just a case of "whittling down" what intelligence really is.
We have known for a very long time that computers and machines are much faster than humans, more accurate, are scalable in certain ways that humans aren't, and they don't tire. I think most people who are not in the "AI cult" would agree that LLMs and modern generative AI are really just an extension of those faster/more accurate/more scalable and never tiring traits. But there does seem to be (and I'm sure folks much smarter than I have quantified this or described it better than I can) a fundamental difference in how humans think, especially as it applies to what true "understanding" really entails, and for the ability to think up truly novel and unique things that are not just a rejiggering/recombination of training data. I believe those skills really are at the heart of human cognition, and as impressive as LLMs are in replicating what this looks like, there are plenty of "LLM failure modes" where it's clear that LLMs lack a true understanding of concepts or the ability to generate useful, completely novel ideas.
There's definitely a lot missing from the current state of the art in machine learning that all brains manage to beat, and we can observe this just because an animal that needs as many examples as an AI to learn motor functions would starve to death before learning to eat.
However I can only guess that this is important, I'm not absolutely certain. They're at risk of being an economic disruptor just by being extremely stupid (by how much they need to study) faster than us to the same ratio we jog faster than continental drift.
Nobody knows what true "understanding" really entails, or what are "truly novel and unique things that are not just a rejiggering/recombination of training data". For the latter, you'd at least have to find an example in history of someone who came up with some idea that has been widely considered "truly novel" by experts, who didn't have any education or training, so no "rejiggering/recombination".
It's almost as if we're building something that... mimics intelligence.
Can you mimic intelligence?
2 replies →
>that could be a factor of intelligence
Could is carrying a lot of weight here.
Because, what's really happening is we're saying "Oh these things are what defines intelligence" then implementing them and /discovering/ "oh wait, there's more to this than we knew".
We've known, for decades, for example that an IQ test is not a measure of Intelligence, even though people still refer to it as though it is. A computer passing an IQ test, therefore, would have been thought of as possessing intelligence way back when, but would not now.
Oh, on the point of "creativity" - is a RNG "creative"? It creates a value unbounded by human intervention (in theory, yes Pseudo RNGs have limitations) - therefore it must be creative... right?
> If a mathematician picks a research direction and spends a whole week on it and it doesn't pan out, they will likely be annoyed, need a break for a while, etc.
Your timelines are a bit unambitious. There's nobody expecting to make significant progress with a week of work.
> There's nobody expecting to make significant progress with a week of work.
You underestimate my ADHD.
Source: I am mathematician.
Hell, I underestimate my own ADHD.
Source: the post-it notes, ALL OF THEM.
They were just illustrating their point, I wouldn't take that literally.
In other words: Thousand monkeys with a thousand typewriters...
https://news.ycombinator.com/item?id=48231974
Take something like
If a thousand monkeys typed at a character per second, on a keyboard with the 23 relevant characters, it would take roughly 10^136 years for them to come up with this counterexample. Though, to be fair to monkey scenario, there's a large family of them known now, so it's not quite this bad: suppose there are a trillion permutations and similar examples that fit in this string. Then we are down to 10^124 years.
If LLMs are monkeys, somehow trained LLM weights allow them to model and prune massive numbers of universes in parallel.
Just FTR - 1 character per second is glacially slow - it's 12 wpm - fine for (slow) transcription, but the monkey typing exercise doesn't require them to know what they are typing out
2 replies →