3 ms·

> GPT is probably absolutely useless for high-level math Terence Tao disagrees with you. https://mathstodon.xyz/@tao/110601051375142142 https://mathstodon.xyz

by gowld 2y ago

> GPT is probably absolutely useless for high-level math Terence Tao disagrees with you. https://mathstodon.xyz/@tao/110601051375142142 https://mathstodon.xyz/@tao/110601051375142142

renewiltord 2y ago
This tool is like all other tools in that expert practitioners extract value from it and novices consider it useless. Fascinating. The advent of search engines was similar with many remarking on how if you searched for something sometimes you'd find websites that were wrong. Things change and yet stay the same. At the time I thought age was the discriminator. Today I know it's not that. It's something else.
- fspeech 2y ago
  But this shows you that the tools (LLMs included) don't understand what they are producing. It takes an expert to extract the value. The tools do enrich the probability distribution greatly, raising the concentration of solutions much above the noise floor. But still they are only amplifiers. Even if they can get over the stage of producing apparent nonsense in the end they still can't get into the user's head to see what is exactly desired, and a precise specification of the intention takes experience and effort. A precise specification of a program can exceed in effort that of actually writing the program.
jerf 2y ago
If it could be solved with a Math Overflow-post level of effort, even from Terence Tao, it isn't what I was talking about as "high level math". I also am not surprised by "Consider a generation function" coming out of an LLM. I am talking about a system that could solve that problem, entirely, as doing high level math. A system that can emit "have you considered using wood?" is not a system that can build a house autonomously. It especially won't seem all that useful next to the generation of AIs I anticipate to be coming which use LLMs as a component to understand the world but are not just big LLMs.

by gowld 2y ago

> GPT is probably absolutely useless for high-level math Terence Tao disagrees with you. https://mathstodon.xyz/@tao/110601051375142142 https://mathstodon.xyz/@tao/110601051375142142

renewiltord 2y ago

This tool is like all other tools in that expert practitioners extract value from it and novices consider it useless. Fascinating. The advent of search engines was similar with many remarking on how if you searched for something sometimes you'd find websites that were wrong. Things change and yet stay the same. At the time I thought age was the discriminator. Today I know it's not that. It's something else.

fspeech 2y ago
But this shows you that the tools (LLMs included) don't understand what they are producing. It takes an expert to extract the value. The tools do enrich the probability distribution greatly, raising the concentration of solutions much above the noise floor. But still they are only amplifiers. Even if they can get over the stage of producing apparent nonsense in the end they still can't get into the user's head to see what is exactly desired, and a precise specification of the intention takes experience and effort. A precise specification of a program can exceed in effort that of actually writing the program.

jerf 2y ago

If it could be solved with a Math Overflow-post level of effort, even from Terence Tao, it isn't what I was talking about as "high level math". I also am not surprised by "Consider a generation function" coming out of an LLM. I am talking about a system that could solve that problem, entirely, as doing high level math. A system that can emit "have you considered using wood?" is not a system that can build a house autonomously. It especially won't seem all that useful next to the generation of AIs I anticipate to be coming which use LLMs as a component to understand the world but are not just big LLMs.