Login
You're viewing the discuss.systems public feed.
  • Edwin Bradyedwinb@types.pl
    Oct 5, 2026, 10:27 PM

    RE: mastodon.social/@DamnInteresti

    I don't know about anyone else but I don't think I care what AI can or can't prove in most cases. What I care about is what humans understand in the process.

    Nothing in my daily life changes if the Collatz conjecture is proved or disproved by machine, for example. Many things - I can't possibly know what - could be changed by the human understanding that arises in the process.

    Believe in yourselves people, for goodness sake. Humans are amazing. It's time you understood that.

    (There's plenty of nuance to this as always but this is the internet so I assume people I don't know will assume the worst. You can all go away. Life's too short.)

    💬 1🔄 16⭐ 42

Replies

  • Ethan Bodzioneybodzioney@mathstodon.xyz
    Oct 5, 2026, 10:36 PM

    @edwinb I agree the majority of the interest comes from how the theorems are proved. But isn’t there some case to be made for interpreting the machine generated proofs? Do we benefit that much more if someone works out the proof by hand first?

    💬 3🔄 1⭐ 1
  • cassieporglezomp@mastodon.social
    Oct 5, 2026, 10:41 PM

    @bodzioney @edwinb there were a lot of other articles written about this around the navier stokes scam that justify it more thoroughly but yes—the structure of the things that were attempted and failed are very important to the understanding, and those are all gone if you just publish a proof.

    💬 0🔄 0⭐ 2
  • Edwin Bradyedwinb@types.pl
    Oct 5, 2026, 10:51 PM

    @bodzioney Indeed. There are many things to learn from many sources, and therein lies the nuance. I'm just here to remind people that Thinking remains an option.

    💬 0🔄 1⭐ 3
  • 💬 0🔄 0⭐ 0