Torvalds and Hohndel talk AI at the recent Open Source Summit, where the former thinks it's made Nvidia better, the latter thinks it's just autocorrect.
Linus Torvalds, the notoriously grumpy man known for shouting at Windows, swearing at Nvidia, and as the creator of Linux, is surprisingly upbeat about the rising hype around the current state of artificial intelligence. "It's hilarious to watch," he notes in a keynote at the Open Source Summit (via ZDNet) this week. "Maybe I'll be replaced by an AI model!"
Torvalds is talking in a keynote "conversation" with his friend and other noted open source luminary, Dirk Hohndel, and their discussion covers a host of different topics. Though of course it had to touch the current tech obsession de jour: AI.
"Let's wait 10 years and see where it actually goes before we make all these crazy announcements," he says, which is probably the most sensible stance to take in a world where people seem to be vacillating between either AI is the saviour of mankind or it's going to irrevocably doom us all.
Hohndel, however, is less than complimentary about the current state of AI, making the point that he thinks most AI is just "autocorrect on steroids."
Torvalds does at least see some positives, though. "Nvidia has gotten better at talking to Linux kernel developers and working with Linux memory management," he says, referencing the company's increasing reliance on Linux systems running all those large language models efficiently. And anything that gets a positive comment from the generally rather anti-Nvidia Torvalds has got to be something, right?
It was over decade ago that he very publicly issued the immortal phrase: "Nvidia, f*ck you!" while looking down the lens of a camera during a talk in Finland and raising his middle finger. And that was after already calling it "the single worst company we have ever dealt with."
So, maybe he's mellowing out?
What is artificial general intelligence?: We dive into the lingo of AI and what the terms actually mean.
Specifically, Torvalds does also see a value in AI tools for the sort of development that has been the main focus of his professional life. He states that he is "looking forward to the tools actually to find bugs. We have a lot of tools, and we use them religiously, but making the tools smarter is not a bad thing.
"Using smarter tools is just the next inevitable step. We have tools that do kernel rewriting, with very complicated scripts, and pattern recognition. AI can be a huge help here because some of these tools are very hard to use because you have to specify things at a low enough level."
Honestly, this is probably the most even-handed I've heard Torvalds talk about a subject that's traditionally been incredibly divisive, maybe ever. I'm almost shook. Though he does also warn us all to be careful of "AI BS", but I think that's a fair warning most of us can readily heed.
