The death of the GPT wrapper has been greatly exaggerated. Modern harnesses are just GPT wrappers rebranded. Before you go looking for your pitchfork, consider this. Claude Code, Codex, and Cursor are all GPT wrappers in the most meaningful sense: programs that decide what to put into a model and what to do with the output. These are also among the most useful AI products I’ve used. Those…
This is a reflection on the makings of Project Basilisk, a short-ish game-ish experience about creating safe AGI in a world where being first is all that matters.
Moment recently held a wonderfully fun programming challenge: You write a program in a custom assembly-like (we call it ant-ssembly) instruction set that controls 200 ants. Each ant can sense nearby cells (food, pheromones, home, other ants) but has no global view. The only coordination mechanism is pheromone trails, which ants can emit and sense them, but that’s it. Your program runs identically…
Welcome to my notes on memory, where I make an effort to understand memory, its effects on our person and our society, and implications for the development of artificial intelligence.
Summary: The Postgres query planner can produce query plans with much longer runtimes than optimal due to poor choice of join methods. One reason the planner chooses non-optimal join methods is by underestimating the results of previous joins, an error that compounds in more complex queries with multiple joins.
Decision-making can be hard. Even seemingly innocuous choices come with a barrage of tradeoffs. At a fundamental level, every choice you make has an opportunity cost: all of the other choices you could have made that are incompatible with your current choice. Even the time and energy spent making a decision is a cost.