This is the way.

Not only is Aaron right here, but I learned this week that people bet on preseason games. The NFL has broken so many minds.

No shade at David here, but worth mentioning that the most important work at Bell Labs was done decades before the antitrust suit (1982). Before 1950 they had invented: the radio (1915) the negative feedback amplifier (1927) the transistor (1947) the LLM (1945)

Crazy question, I know, but what if turning everything into gambling is bad?

This semester, I’m back to live blogging my course lectures. I’m teaching Convex Optimization in the Age of LLMs. argmin.net/p/convex-optim…

The most salient aspect of Convex Optimization in the Age of LLMs is LLMs can solve all the problem sets. argmin.net/p/cheating-at-…

In the first technical lecture of my convex optimization class, I describe a functional programming language for generating convex sets. argmin.net/p/programmable…

Silver has a long passage in his book about how he can feel the difference between 52/48 and 50/50. Apparently, he's not feeling his model.

Regardless of statistical power or whether they are random guessing, election forecasting has added nothing good to the world and has probably had a bunch of negative effects. Like many things, it's a bad thing that humans have a huge appetite for.

Markets can remain irrational longer than you can remain solvent. An evergreen subtweet.

Convex optimization is the class of problems where local search always works. In today's lecture, I argue this is almost a definition. argmin.net/p/think-global…

You know we've crossed the AI rubicon when Lawrence Lessig is writing articles arguing any regulation is good regulation. x.com/lessig/status/…