RE: https://mastodon.social/@glyph/116696314673841350
I want to clarify my reaction here: context is everything.
I haven’t looked at the code; I don’t even know what the bug is specifically.
If Tridge had done this in a vacuum, if LLMs were a new tool with no history, I would have *absolutely no idea* if the regressions were caused by the LLM or not. What makes this story interesting and worth commenting on is that it’s a part of a pattern that has been recurring everywhere in the industry and the community, particularly with our venerable elders