Positive users express amazement at the rapid progress and scale of using GPT-5.6 to formalize an Erdos counterexample in Lean, while negative users criticize remaining flaws and the impractical 1M-line size.
Based on 6 visible X reactions from 7 accounts; directional sample.
Ask a question below.
Published answers will appear here.
@SebastienBubeck Pretty cool that this can be applied to non-math substrates, formalizing easily provable theorems that can be plugged into business processes or distributed systems (e.g., defi). Very important future area.
@SebastienBubeck ngl 1M lines of formal LEAN from one person sounds absurd the field is moving way faster than most people realize
@SebastienBubeck that is absolutely wild, formalizing that much logic solo is insane.
@SebastienBubeck one million lines and there's still a sorry sitting in the theorem statement
@SebastienBubeck 1M lines of proof is still too much for this
@SebastienBubeck @boazbaraktcs ridiculous
Btw this was done with gpt-5.6. One million lines of LEAN to formalize the unit distance solution. Pretty cool that this can be done by a single person in a short time (as opposed to a team over year(s))! https://twitter.com/AlexKontorovich/status/2070561716778254769
Positive users express amazement at the rapid progress and scale of using GPT-5.6 to formalize an Erdos counterexample in Lean, while negative users criticize remaining flaws and the impractical 1M-line size.
Based on 6 visible X reactions from 7 accounts; directional sample.
Ask a question below.
Published answers will appear here.
@SebastienBubeck @boazbaraktcs ridiculous