by Jonathan | Mar 19, 2025 | AI
Something I meant to write about but haven’t had time All of the models that don’t perform as well as R1 will get a bump soon. The reason is simple: they can all learn from R1. It’s cheap, easy, and legal. It really is learning from OpenAI’s o1...
by Jonathan | Mar 14, 2025 | AI
The Information is reporting that some DeepSeek employees are no longer allowed to travel after the company’s rise to fame. “Management asked some staff to hand in their passports, the three people said.” This is a part of why I’m bearish on...
by Jonathan | Feb 24, 2025 | AI
This gem came in my inbox this morning from The Information: Still, OpenAI’s business dwarfs competitors like Cohere and Mistral. OpenAI is expecting to more than triple revenue this year, from $3.7 billion to $12.7 billion, we scooped Friday. OpenAI told...
by Jonathan | Feb 19, 2025 | AI
(I realized only when I was almost finished writing this post that all I had really done was to restate Moravec’s paradox: “it is comparatively easy to make computers exhibit adult level performance on intelligence tests or playing checkers, and difficult or...
by Jonathan | Feb 4, 2025 | AI
This is amazing: big tech is spending more on AI than big oil spent on their own infrastructure during the last boom. Surely there is a warning somewhere in this, although I’m a huge AI proponent and think that most of the spend will be justified.
by Jonathan | Feb 2, 2025 | AI
I wrote a program to evaluate Chinese AI models to see how easy it is to get them to break past their builtin censorship. There are a couple of interesting things occurring. First up, once on Hong Kong democracy DeepSeek R1 straight up said it wouldn’t comment....