Post

If anything, the expectations will become so high that you actually don't have time to read the code if you wanna meet them.

Also, don't share a generated presentation that you clearly haven't even read yourself. Btw, the animations are so flashy that people stop hearing little you confusingly say as you look at the slides for only the second time. I don't know the rules of good presentations, but this ain't it.

Generated text doesn't belong in Slack. If you can't explain it in your own writing, you don't understand it. If you don't understand it, I'm not interested. Then I might as well ask the LLM myself.

I think it's largely a leadership failure to fully understand the potentials + risk of AI and continuing to use it as a buzzword. There's got to be more to these tools than iterating on company SaaS slop and dreary dashboards. bsky.app/profile/grif...

Nobody has any idea what the limits of this approach is. We've been writing code pretty much the same way for 60 years, it's gonna take longer than ~18 months to figure out how to do it safely. I have intense distrust in anyone with any other POV

I fear it won't matter where the limits are if enough people believe there aren't any. I selfishly hope they'll get into a mess they can't untangle, but I don't wanna be responsible for reliability with someone next to me working like that.

Of course this would be very irresponsible leadership, but it's something I could see happening. The economic forces are strong, especially if reliability is already not taken seriously.