This declaration was made September 10th. It has over 5000 signatures now.
Over the last few months, the mathematical capabilities of LLMs have improved dramatically, to the point that they can solve major outstanding problems in many fields of mathematics. However, the push by AI companies to solve mathematical problems as a benchmark is detrimental to the science of mathematics, and to the mathematical community. The goals of the AI companies and the goals of the mathematical community are severely misaligned. We see these as part of broader alignment issues impacting other scientific and creative professions, as well as the whole of society. more
I believe the cases of ”math problems settled by AI” will slow down. Anthropic probably want some high profile math problem at the same level as Navier-Stokes, but after that I don’t think there will be as much progress.
They got the publicity stunt they wanted. What more is there to gain? Navier-Stokes costed many millions to settle, and probably was close to being resolved by human mathematicians either way. Who else has anything to gain by spending ridiculous amounts of compute on true/false answers?
My guess is that the AI companies will focus on finding publicity stunts in other fields. Possibly medicine.
This is the same argument as why you shouldn’t use an LLM to do your homework. The goal of doing the problems isn’t to solve the problems (we can see they are already solved, the universe exists and calculates all kinds of exotic math all the time), it is for humans to understand the math
In recent months, the success of AI in solving major mathematical problems has made headlines even outside mathematical circles. But solving problems is only a tool and proxy for achieving the primary goal of conceptual understanding and insight. Forgetting this in the world of AI may turn the tool against the primary goal. Indeed, the mass production at faster and faster pace of “true/false” statements could destroy fertile ground instead of breathing life into new ideas.
You shouldn’t use an LLM to do your homework but you can certainly use it to learn and clarify things you don’t understand. I teach math and physics and tested local LLMs for learning and I can say they really help, at least in these subjects. The best option is to learn from books and making your homework, but when you’re stuck with something that you don’t understand, ask a good LLM to clarify things, give examples and context. I tested myself while learning about differential geometry and relativity, it really speeds up the learning process and understanding. But don’t use them as a primary source, nothing beats a good book and practice on that. I wish there was something like that when I was in college, I wouldn’t have carried some misunderstandings so many years. Being just a tool it’s up to you if you use it responsibly or not, as any other tool.
The goals of Big Tech in general - not just AI companies - is to make money through surveillance, and collaboration with fascist governments.
I’d say they’re severely misaligned with human beings, democracy and basic decency. Which is totally unsurprising considering that all tech bros are psychopaths.
This is not a problem exclusive to big tech. Absolute power corrupts, absolutely. The oligarchs who own energy and military companies have been committing crimes against humanity the entire time. The vast majority of billion dollar corporations display criminally sociopathic/psychopathic behaviors to both consumers and workers.
Extreme wealth inequality should not exist. Billionaires should not exist. No kings. No oligarchs. No plutocrats. No parasites. Nobody too big to fail. Nobody above the law.
Well, I was specifically addressing AI and the fuckheads who forcibly steamroll it over our lives.
But yeah, this is a systemic problem: the rich and powerful control societies (and not just capitalist societies, to be clear), and it is my personal belief that kind men don’t become rich or powerful, because becoming rich or powerful implies making someone else poorer or taking away their agency.
So if you follow the logic to its simple conclusion, the ultra-rich and ultra-powerful only get there not by being merely unkind, but by being truly evil. And this has nothing to do with the particular field they happen to be involved in.
I feel this has nothing to do with mathematics in specific but generalizes to almost everything. Same story with computer programming, jobs where you’d need some knowledge or skill, school homework, art… As long as there’s humans involved who could grow or learn while doing something, AI is misaligned. Also once there’s more to the result than just turning something, (anything,) in and calling it a day.
_Source: