About pro-AI boosters:

One of the people in the company I work for/with is a non-programmer that has a “vibe coding consulting” gig on the side.

Basically half of their Linkedin is about how they’re shipping features at half the speed as developers, with quality, etc.

They’re not.

They took about three months to prototype something that was rejected for not working/unreadable code. Another team built the same feature from scratch in a week.

Seeing the other side of the grift is kind of pathetic.

I think there is a time and place where the outputs of LLMs belong.

The issue is this requires NUANCE.

And the more certain people expose themselves to LLMs and forgo that ability to sit there and think painstakingly about nuanced things - the more useless they become IMO.