Hello World!

George Hotz has written something I wish were less rare in conversations about AI: an enthusiastic, unsentimental take. He loves what LLMs can do, uses them, and still refuses the breathless promises and fear-driven marketing around them.

I recognise a lot of my own position in it. Models are useful, learning to work with them is a real skill, and vibe-coded slop is still slop. You don’t have to choose between worshipping the machines and pretending they’re useless.

Short, sharp, and well worth five minutes: I love LLMs, I hate hype.

Hi! This is just a sample note to test this feature.

I don’t like social networks, but I find the idea of microblogging interesting. So I decided to build a simple, home-made version within my main blog. If, for some strange reason, you’re interested in my random thoughts about technology, the universe, and everything else, feel free to subscribe to the RSS feed.