That does sound impressive, but I'll be really impressed when the AI model actually can tell me what the hell "solve all of physics" means
Also,
No matter how "good" these models get, Douglas Adams has been too popular on the internet for this to ever succeed. All AIs based on data scraping will be forced to respond '42' to all inquiries of that sort.the day is approaching when we can ask an ai model to solve all of physics and it can actually create raving nonsensical rants claiming to be a grand unified theory of everything while denouncing the academic establishment for ignoring its genius, thereby automating the thankless task of giving physics grad students someone to punch down on
For centuries, humanity has worked towards building an AI that will understand all of physics. Now, upon completion, we have learned that the key lies in the idea of "four simultaneous 24-hour days".
This is my new screenplay idea.
it's clearly the future
buy buy buy buy our bespoke political t-shirts and plaster your exact species of brainworms on your torso for all the world to see!
It pisses me off how biggest proponents of "AI" don't seem to understand how these algorithms even work. No, LLMs literally cannot "solve physics", whatever that means
Altman is just doing marketting when he talks about the dangers of AI taking over.
This is just the infinite corridor of monkeys and typewriters bit but played straight, lol
my friend is doing gig work to make chatgpt better at just boring old textbook physics problems, and it's complete dogshit at it. so uh, sure man. nice
you got there.Didn't they manage to make it somewhat good at solving certain math competition problems? Regardless it's a pretty big jump from that to making a breakthrough in physics.
maybe certain ones, but it's generally bad about numbers and mathematical reasoning. he also gets paid to make it fail at math, and it's arguably worse at basic math than physics.
Yeah deepmind had good results with IMO problems, but only geometry problems. They scored almost at the level of gold medalist. That's only a fraction of IMO problems, though. They did it by combining a formal verification system with a LLM to propose solution paths, and then doing some tree search I think.
This is one way to improve large AI systems and will probably be incorporated in some way in the future, for example by integrating with a language like lean (for math proofs).
They will also be improved by combining with tool use like calculators, code interpreters, web search, calendars, etc. This is already starting to happen to some extent.
LLMs by themselves, at least with current architectures using transformers, are not great at reasoning (counting, arithmetic, symbolic reasoning)
Unified field theory, or the theory of everything: https://en.m.wikipedia.org/wiki/Theory_of_everything basically uniting general relativity and quantum mechanics.
I wish these nerds stuck to solving the velocity of a ball rolling down the hill instead of trying to be enlightened philosophers
in fact, speaking of a ball rolling down a hill, i have a fantastic problem for these AI bros to work on. see, there's a big hill and there's a big rock that's at the bottom of that hill which id really prefer to be at the top instead. you think they'll volunteer?
Unless these morons are hiding an honest to god AGI in their back pocket, they're fully insane, and it's boring at this point.
Even if they do, it's probably a mechanical turk somehow. I don't believe these dunces capable of making an actual gestalt organism
Guys, AI just needs a lot more money and it'll work.
- Write down the problem.
- Think real hard.
- Write down the solution.
A lotta yall still don’t get it
AI holders can solve physics with a single AI
After humans solve the problem and post it online somewhere to be hoovered by a LLM data scraper, an AI will be able to solve the problems already solved by humans.
Me when I train a LLM with things that humans have made and done, doing nothing but remixing the training data: "create something novel?" "No."
People legit think that this is the entirety of human creation: only remixing the past
