he doesn’t “support” LLMs yet, he tolerates some uses
he’s also getting older, and if the past decades have shown anything is that brain rot can happen to anyone - and just because someone was based in the past, doesn’t mean they are going to have a solid moral compass forever
Linux is being sponsored by AI providers. Consider this as de facto corruption.
IMHO, he, and some other “vibe code bad, moderated LLM use good” developers are falling for the industry’s definition of anti-AI “extremism” (simple refusal of AI usage, refusal of engaging with genAI output), and being a moderate with AI (reading the output, etc.).
Yea I have to agree I hate this “moderated LLM use good” stance. I have used LLMs (through private browser windows without login) to understand documentation before, or to get a usage example for a specific function, and I already feel dirty for it. I will absolutely not use that snake oil to generate code for me or even consider installing anything on my hardware.
Even worse is, it still makes up stuff when used in “moderation”, so you’ll get the downsides of vibe coding without the benefits, because you didn’t want to read the docs.
When you are equating something that Linus Torvalds supports with fascism, I think you might have lost the plot.
Yea I have to agree I hate this “moderated LLM use good” stance. I have used LLMs (through private browser windows without login) to understand documentation before, or to get a usage example for a specific function, and I already feel dirty for it. I will absolutely not use that snake oil to generate code for me or even consider installing anything on my hardware.
Even worse is, it still makes up stuff when used in “moderation”, so you’ll get the downsides of vibe coding without the benefits, because you didn’t want to read the docs.