@argv_minus_one@mastodon.sdf.org
2026-07-22 19:24 UTC
Replies (2)
-
@argv_minus_one@mastodon.sdf.org 2026-07-22 19:27
@borgbackup@fosstodon.org “You also shouldn't just assume that AI is used irresponsibly if you don't have proof.” I fail to see, when you use a machine that's this bad at programming to do your programming and then claim it's good, how such use could be anything other than irresponsible. “Also, don't assume that just because claude (or other AI) contributed to a commit means that it is purely AI generated and nobody looked at it.” What evidence is there to refute that assumption?
-
@borgbackup@fosstodon.org 2026-07-22 20:41
@argv_minus_one@mastodon.sdf.org Guess it depends on the problem, on the LLM and on how you use it. I had quite good results from claude (opus and fable, for simple tasks also sonnet) as well as from gemini. Of course it is not always perfect, that's why there needs to be a good human developer in the loop. But quite frequently it exceeded my expectations by a lot! Of course I review the output, also we have tests that make sure the code works.