What tactic will prove the most mathlib lemmas at the end of 2026?
Mini
8
แน7932027
1D
1W
1M
ALL
1.7%
simp
7%
aesop
1.7%
rw_search
1.6%
sorry
4%
duper
40%
grind
44%
Sometime around the resolution date, I will write a script that samples random tactic-mode lemmas from mathlib, and replaces the proof of those lemmas with an invocation of a single tactic. This market resolves to whichever tactic of the ones provided as answers proves the biggest fraction of lemmas.
Get แน1,000 play money
Sort by:
By "prove", is it ok if there are warnings? Because I know of a tactic that has a 100% success rate ๐
@tfae Lol.
But no. Feel free to ask more questions about what counts as proof, but sorry does not count.
Related questions
Related questions
Will aesop be able to replace >50% of mathlib proofs by 2025-11-26?
39% chance
Which theorem prover will have proved the most theorems on Freek's list by end of 2025?
Will rw_search be able to replace >50% of mathlib proofs by 2025-11-26?
19% chance
Which theorems will be formally proven in Lean by the end of 2028?
Will the best public LLM at the end of 2025 solve more than 5 of the first 10 Project Euler problems published in 2026?
75% chance
Will reinforcement learning overtake LMs on math before 2028?
70% chance
Will the majority of mathematicians rely on formal computer proof assistants before the end of 2040?
60% chance
Will an AI solve any important mathematical conjecture before January 1st, 2030?
80% chance
Will AI solve one of 129 major mathematical conjectures before year X?
Will we have a formalized proof of Fermat's last theorem by 2029-05-01?
72% chance