Put a mark on the calendar, a line in the sand. LLM assist on programming is going to become a critical statement about a person's normalcy; if they can't jive with "AI", they will be shunned for non-compliance.
OK, just had to say it in the extreme so I can move on. Getting MIDI-level lock-in vibes from the tooling changes. "This is so awesome, transformative…" OK, but if it's not, am I extremely advanced and too good, or am I a nincompoop and too bad?
I find it does the simplest things en masse best, but those are the things that are hardest to double-check. The low-value coding. So I need robust testing? But who's gonna write all those tests for that massive code I wrote in minutes? 😉
@igmp_join I was kinda begging that point. I've written tests with an LLM, and they passed. Now did I write code that I meant to write? I have a mark in the sand for moving forward, knowing if my tests break I changed something.