Something I thought was interesting: You could take his rant about C++[1] from 20 years ago and with minimal edits make it an extremely coherent complaint about LLMs and vibe coding.
Definitely very interesting to watch this unfold. I didn't know that Linus expressed a public opinion on AI use for coding until recently, but I think most people would've pinned him as someone who would not accept it. I wonder if he has been following the development of the capabilities of agentic coding, or if he has just been exposed to the more recent and more capable iterations, making it an easier pill to swallow.
It's not that he changed his mind. He said there: "The kind of expectations I have is that people do their best and I expect people to understand what they are doing. But that understanding has very little to do with what tools they use."
He's always been entirely pragmatic. This is unsurprising of him.
A pragmatist could have tried early LLMs and written them off. People are generally hyping them as much now as they have been for a long time, so it is possible to consider the people still promoting them as delusional, as many honestly were about early capabilities. That is the point I was making.
It's about time that someone put the gatekeepers in their place.