Why is being reliant on an LLM for proofs any different than relying on computers for calculation?
retained text
Say you have a sequence of a trillion random coin flips. No mathematician on the planet can memorize that sequence. No mathematician could go to the chalk board and tell me if the 8th billionth flip was heads or tails. No mathematician using a chalk board could tell me if it has an even number of heads. If I ask a mathematician about this problem, they will just swivel chair to a computer and have the computer spit out answers. So epistemologically what is different about having the computer spit out a LEAN proof of the Riemann Hypothesis? Why is that considered more of a crisis than calculators are?
document fetched 2026-10-10 00:31 · canonical https://reddit.com/r/singularity/comments/1x1rvh2/why_is_being_reliant_on_an_llm_for_proofs_any