http://localhost:3001/blog Notes on how MingLLM works, what we learned building it, and the choices behind the product. en-us Sat, 18 Apr 2026 00:00:00 GMT http://localhost:3001/blog/4b-feels-like-400b http://localhost:3001/blog/4b-feels-like-400b Sat, 18 Apr 2026 00:00:00 GMT Yiming Beckmann Engineering Three moves that matter: expert iteration, selective generation, and tight tool-use shaping. How we get to 94.9%. http://localhost:3001/blog/no-chat-window http://localhost:3001/blog/no-chat-window Thu, 02 Apr 2026 00:00:00 GMT Yiming Beckmann Product A chat UI is the wrong default for a voice-first agent. We went looking for a better one. http://localhost:3001/blog/orb-one-month http://localhost:3001/blog/orb-one-month Fri, 20 Mar 2026 00:00:00 GMT Yiming Beckmann Research What we learned from the first four weeks of persistent memory in private beta. http://localhost:3001/blog/local-first http://localhost:3001/blog/local-first Thu, 05 Mar 2026 00:00:00 GMT Yiming Beckmann Company Cloud-first is a choice disguised as gravity. Here is what we are choosing instead. http://localhost:3001/blog/tensor-six-weeks http://localhost:3001/blog/tensor-six-weeks Sat, 14 Feb 2026 00:00:00 GMT Yiming Beckmann Engineering A short postmortem of how we built the first Tensor alpha, what broke, and what we cut.