There's no evidence of any of that, it's pure speculation. I bet people were saying the same thing about Intellisense, and Google, and documentation-on-hover, and live compiler errors, etc. etc.
However, in this case there is empirical evidence copilot makes code quality worse: https://visualstudiomagazine.com/articles/2024/01/25/copilot...
I work in such an industry and it's certainly not "self-evident". Clearly it's also not self-evident to many other people commenting on this post.
> However, in this case there is empirical evidence copilot makes code quality worse: https://visualstudiomagazine.com/articles/2024/01/25/copilot...
One study does not prove anything, although at least it is _some_ evidence. But other studies linked in that very article show opposite results.
To be clear I don't know if LLMs are better or not for developer productivity. I think it's too early to tell. But given that people, in general, are prone to conservatism, combined with the fact that past developer tooling enhancements have not caused the negative outcomes that many people at the time predicted they would cause, leads me to believe LLMs will probably be a positive change.
Tools like the replit bot evoke similar emotions. I have no idea why they generate so much negativity from most HN comments. This cohort must be on average younger than myself. I'm pushing 50 and for the first time in decades I fee like we are actually about to witness actual productivity boost in programming craft because let me tell ya, server side js or spa or nosql certainly weren't it.
It may suck, but facts are facts. Dealing with them is better than pretending they don't exist.
Technology can be disruptive, but "just deal with the disruption," is not a viable response. It breads resentment. It favors populism. It says, "sorry, but your problems are not worth fixing."
When that equation changes en mass, and we respond with "things change, too bad," it is a social failure.
Hillary Clinton? Is this bait?