Post #1342186
2026-02-11 18:09 UTC
Replies (4)
-
@yvg@indieweb.social 2026-02-11 18:11
@janmon@social.lol While claude can help, it's better to assume authors have better understanding of what and why they wrote what they wrote than relying on a pattern recognition machine that is 20% times wrong 😄
-
@gregor@toot.cafe 2026-02-11 18:22
@janmon@social.lol you can ask AI how something works, but in most cases not why it was build a certain way. Even if you ask the team how sth works you'll get a better answer if the people who built it are there. Also, why not both
-
@hl@social.lol 2026-02-11 19:36
@janmon@social.lol Yes, that sounds disappointing. Perhaps people can use it to talk more about the history of the project, and why certain decisions were made, and what people might do differently given what they know now. I would imagine that could lead to some interesting insights from those who've been there longer, and something an AI can't answer. Might also give more insight into some of your colleagues. Any significant codebase is a kind of shared history, and in my experience, contains interesting fossils and associated stories.
-
@erick@social.erick.sh 2026-02-12 02:41
@janmon@social.lol I love those sessions. Team culture is key here. The people who join and drive these calls need to be sure to keep the space safe for everyone to keep sharing and asking without being worried of someone throwing a modern version of RTFM or LMGTFY. My favorite phrase to make people ask me questions is: “the only stupid questions are the ones you are not asking”