← Feed

Posts

• View post

  Anyone interested in proving the Collatz conjecture (aka taming the beast of Collatz) using LLMs, storing their output, and chaining the lemmas to finally prove the whole thing, just found the right repo. https://github.com/Chessing234/collatz Any thoughts on if this might be viable? @tao

• View post

  Why can't mathematics shift from solving problems (what we do now, what Prof Tao does, what mathematicians are currently doing) to decoding solved conjectures and drawing analogies (what 3blue1brown does)? I don't understand why not.