I’m with Linus on this. The reality is that when used responsibly, LLMs are tools that enable some amazing stuff to be built. Responsibly meaning the code it generated is read, and refined.
This is a separate issue from data centers and corporate greed. You can always just run a local LLM. If you are digging your heels in at this point then you are ignoring the reality the same way someone who was using punch cards would ignore an IDE.
I’m with Linus on this. The reality is that when used responsibly, LLMs are tools that enable some amazing stuff to be built. Responsibly meaning the code it generated is read, and refined.
This is a separate issue from data centers and corporate greed. You can always just run a local LLM. If you are digging your heels in at this point then you are ignoring the reality the same way someone who was using punch cards would ignore an IDE.
Not remotely comparable.
The IDE vs punch cards comparison is ridiculous, there’s a provable performance difference when writing code with an editor (not everyone uses IDEs!).