The machines are fine. I'm worried about us.
https://ergosphere.blog/posts/the-machines-are-fine/

I am not a physicist, I was a biologist and now I am a software engineer. This blog still resonates on many levels with my feelings on LLMs.

#LLMs

The machines are fine. I'm worried about us.

On AI agents, grunt work, and the part of science that isn't replaceable.

@Garonenur this includes the assumption that intuition won is intuition kept.

That your skills don’t degrade from disuse.

Which is a necessary assumption to be able to conclude that using an LLM is fine for experienced people.

So it’s too optimistic, I think.

Regarding convenience: when I’m driving and my co-driver knows the way, I arrive without learning the way. I should avoid LLMs.

@ArneBab well, this assumes that a supervisor still needs to use some of their skills to check the LLM, but the fear of just accepting and not checking is also voiced in the article.
But I am not disagreeing with you, I think in all fluid fields, like sciences and software, one has to not only keep old skills, one also has to hone and expand them all the time.