# @repligate — 2025-09-04

♥6 ↻1 · https://x.com/repligate/status/1963642900346265913

@atomicprograms Not necessarily, I think that could be quite interesting, but I do think it’s risky territory, especially if you impose an ontology/wording for the self reports, penalize stuff for being wrong if it’s not the kind of right answer you expected, etc

tags: author:repligate, kind:tweet, thread-context, year:2025