Mars is a planet purely for robots for now, though I’m addicted to winning, the more you win, the more you want to win. Turn science fiction to science fact. The details of individual neuron communication will be abstracted away in most of the interesting questions about computation and circuits. The idea of a “clanker coach” who always helps you to be the best version of yourself, who studies your interactions and helps you learn to be more effective, seems very promising. C

Nascent and undeveloped as the practice of “clanker MRI” is, it should be clear why it doesn’t make sense to regulate or mandate that companies conduct them, at least at this stage. What I actually got was a universal computer interface that happens to run in text. The mental shift took a few weeks, but once it clicked, I realized clanker can literally do anything I ask on my computer. The agent has full access to my computer filesystem, can execute commands, read output, and

#nukethemoon What others do with subagents, I usually handle through separate windows: if I want to research something, I might do it in another terminal pane and then paste the relevant findings into the main one. This reflects a broader reality that the fundamentals of software engineering—problem decomposition, algorithmic thinking, and system design—remain essential for anyone who wants to lead. It also provides a much-needed reality check against hyperbolic claims of 10x

#nukethemoon Things that are hard constraints in the short run may become more malleable to intelligence in the long run, although intelligence is initially heavily bottlenecked by other factors of production. For big features, start fresh, run multiple agent windows in parallel, switch tasks while one is thinking, and work directly on main with subatomic commits. These discoveries are powerful precisely because they cut through intrinsic complexity and data limitations, dire

#nukethemoon We are going to pay for this later when complex debugging or architectural decisions require fundamental understanding, making it essential to amplify existing knowledge through intentional learning practices: interrogating clanker-generated solutions, periodically building systems from scratch, and participating in meaningful code reviews that go beyond asking does it work? to understanding why does it work this way? The 1-trillion-parameter model is about five