Writing
All writing
Essays and notes on frontier AI, the economics of automation, and the labs racing to build it.
- April 2, 2026·AI
Why Reinforcement Learning Will Eat White-Collar Work & Frontier Labs Are the ONLY Ones at the Table
With Dario's 50% estimate of white-collar job loss by 2027 in mind, things start to get real. The TAM of all white-collar salaries is so astronomically high that the CapEx to automate it is completely justified — and once a frontier lab locks in a company like Goldman Sachs, model feedback, data and results form an increasingly relevant moat.
- January 25, 2026·Tech
Elon Musk, xAI, SpaceX, Macrohard & Tesla
xAI's strategy: prioritizing latency, scale, and real-world deployment over polished abstractions, even if that means letting competitors lead on narrow benchmarks like coding. If Macrohard is real, it reflects a belief that brute-force computer use plus massive compute can outpace carefully scaffolded systems.
- January 30, 2026·Tech
Meta AI Growth Opportunity
Notes from Meta's earnings call and where Meta goes next — GEM, Reality Labs, the Neural Band, Genie 3, and a thesis on the future of advertising and content consumption.
- January 28, 2026·Paper-EconPDF
Against Broad Ex Ante AI Regulation: Dynamic Governance, Innovation, and Geopolitical Competition
A paper arguing against broad, pre-emptive AI regulation in favor of dynamic governance — weighing innovation and geopolitical competition. Written with OpenAI’s Prism tooling for the LaTeX; all ideas are my own.