@InvestorBub@gfodor It's an interesting novel challenge. I don't resent learning to code over the last 20 years, just for AI to do it in 1% of the time, but I will be critical of what it generates.
@gfodor I'm not suggesting people understand software by reading PRs, but the code itself. The PR should just be a (smallish) diff, assuming we aren't accepting gigantic PRs. You need to know the underlying software to understand the PR.
@InvestorBub@gfodor I'd suggest this is not a good approach. Knowledge management, complexity/technical debt management are important. There may come a time where Claude decides to completely refactor your codebase, and you need to be able to push back or not.
@gfodor I'd suggest that unless people are just blindly accepting AI generated PRs, they are still reading them, and in order to understand the generated code, they have to understand the system etc, make sure the PR doesn't violate decisions etc.
Someone asked me how I find the interesting Fluoddity configs in such a vast parameter landscape. It's a difficult question to answer directly, a bit like "how do you know what to paint?" But I figured the best way to get at the answer is just to detail my process. A thread:
@fingaz_1122@gfodor I assume that I will forget everything about the software I have written, and that I will need to go back and read it later, so readability is a big factor. Having AI generate unreadable code that "works", you might as well get it to generate binary.
@scottastevenson That isn't a post-AGI world, that's a post *constraint* world. If everything were "on the table" capitalism wise pre-AGI (and it isn't, because human desire and curiosity is infinite), then AGI just takes things off the table (potentially). What is left on the table is everything
Siempre me fascina cómo funciona el cerebro. Si miras esta imagen no ves nada más allá de manchas. No obstante, mira la imagen que tengo en la respuesta a este tuit y de golpe tiene sentido y ya no puedes verlo como mancha.
PD: De «La vida secreta del cerebro» de Lisa Barrett.
PICARD: Data, shields up
DATA: Brilliant! Shields can reduce damage we sustain. Not immunity. Not hubris. Just prudence. It's not precaution—it's strategy.
[camera shakes]
WORF: HULL BREACHES ON NINE DECKS
DATA: Here's what happened: you told me to raise shields, and I didn't