Short notes and long-form technical write-ups, including deep dives and architecture.
MiniMax M2.5: a 230B open-weights model that matches Opus on coding, leads on multilingual SWE, and costs 1/20th—plus Forge RL, MoE, and why it might be the most important AI release of 2026 so far.
Seven evidence-based takeaways on how the fastest engineering teams are compounding advantage with AI—spec clarity, verification, alignment, and why the engineer's job restructured rather than disappeared.
An engineering guide exploring OpenClaw's architecture—lane-based concurrency, hybrid memory search, WebSocket protocol, and the agent loop that powers autonomous AI.
Today I launched my personal TIL blog built with Next.js 15 and deployed on Vercel.