Software maintenance is changing. A few months ago, I had to tell Codex exactly where to look for dead code. Now it can find candidates in Slack, check them against production telemetry, and draft cleanup PRs. Here’s how we do this at OpenAI.
https://t.co/PdQl43nLeQ
@coldboothq Fixed thresholds are too rigid, I let the model choose depending on the evidence it sees. Usually I just don't prompt a threshold at all, but something like this also works: "Ensure dead code is not called for at least 2 weeks, more if you're not certain it's dead."
@BruteSmorch@pickover You can use a symmetric principal value, integral from -R to R where R goes to infinity, still an integration and that’s 0 but we’d need to say it’s a symmetric sum not just a sum