Many users reacted negatively to frontier AI models surpassing humans on prestigious math tasks because they fear rapid progress will render mathematicians obsolete and diminish the value of human achievements.
Based on 13 visible X reactions from 96 accounts; directional sample.
Ask a question below.
Published answers will appear here.
The claim prompted jokes about AI winning the Fields Medal.
@littmath my knee-jerk response was to bring up the calculator (a lot of mathematicians took pride in elegant ways to do arithmetic) - but frontier models are fundamentally different. however I still think problem solving as a field will only be enriched + become more rewarding, not less.
@littmath Why do you expect some adaptation is possible, rather than human minds simply becoming obsolete? It seems unlikely to me that humanity will survive the creation of superintelligence, but if we do I don't see much reason to suspect our cognitive abilities would remain relevant
@littmath @subcountability I’m not so sure (not a mathematician though). These math conjectures have no practical applications… they were purely challenges for the human mind and now they’re crumbling. What if George Mallory had dynamited 5000 feet off the top of Mount Everest to summit it first?
@littmath If AIs begin making all the important new mathematical discoveries, and humans cannot understand them but must regard them as almost certainly true unless the AIs are deliberately deceiving us, the human mathematical prestige system probably will not be able to respond.
@littmath that's one way to try and get some funding from frontier labs 🤡
@boazbaraktcs To be fair, maybe we can count from GPT-4's launch date, because to me that's when this all really started. So they have lots of time.
Which model will win the first Fields Medal?
@AndrewCurran_ Does it have to be under 40 in "model years"? 🤣
Many users reacted negatively to frontier AI models surpassing humans on prestigious math tasks because they fear rapid progress will render mathematicians obsolete and diminish the value of human achievements.
Based on 13 visible X reactions from 96 accounts; directional sample.
Ask a question below.
Published answers will appear here.
@boazbaraktcs @AndrewCurran_ GPT-4o disqualified alas