February Wrap Up & News 2025

Feb 27, 2025

I’ll be mixing up the monthly wrap-up emails, including more new and recent stuff I’m still reading, digesting, and doing…

Big new things

I’ve read so many books in the past year that I decided to make a quick list of the best 16 out of the 100 I’ve read. I keep notes anyhow, so why not share them: Best books of 2024.
I’m adding a new section to my monthly update about posts and news I’m still reading and contemplating.

A preprint on category theory and LLM models. The two authors work at the intersection of a mathematical field called category theory (related to topology, just a lot more abstract) and LLMs. They basically look at how LLMs, in general, behave using advanced math. That does remind me of geometrical deep learning, and that there might still be a lot to discover in terms of higher structures and deeper theoretical foundations for deep learning and LLMs.
Who wins the AI agent battle. It is a piece from every.to I’m still thinking about. In particular, I like their ideas on horizontal and vertical AI selling. I am not sure yet what conclusions should be drawn from the piece, but I like the structure it brings to the AI agent market.
“Slow and extensive usually go hand in hand, not slow and quality” is a quote I got from somewhere. World-class chess players are fast and have high quality in their moves. Indeed, high quality is how you achieve speed. It’s the same in the lean movement, agile, and the basis for it, the Toyota Production System. It’s a lesson I keep on coming back to because it is easy to go for slow and expensive instead of high quality and fast.
Bluesky, Twitter's open protocol spin-off, is approaching 30m users and 1 billion posts fast.

I'm not saying I recommend any of those; I’m sampling, reading, and discarding a lot of books nowadays.

“Shape Up” from the Basecamp folks about their product development/PM philosophy. So far, I think I like some aspects (the ones that get you from ok to good products), and I don’t like the parts that make it hard to go beyond good products to amazing products.
“Fortune Formula” is a book about a personal hero of mine, Ed Thorp (inventor of Card Counting, mathematician, and financial markets wizard), haven’t gotten all the way through it yet.
“The Technological Republic” is a recent book from the CEO of Palantir about America’s tech & defense culture. I haven’t read much yet, I'm not sure it contains that much interesting perspective other than the initial pitch.