Ligang Yan颜力刚

Tag

claude-code

2 posts

  1. CurioSeed V4: stack choices and how a small team works with AIThe key decisions behind CurioSeed V4, laid out with their costs: Supabase or Neon, React SPA or Next.js, client-side fetching or SSR, when to split into a monorepo, and a working method of people plus AI plus rule guardrails. Every choice comes with the pitfalls we hit.
  2. A Claude-centric software architectureFor thirty years, software has meant humans translating changing requirements into fixed code. Once a runtime like Claude Code is strong enough, the more efficient approach is to separate the two: settle deterministic logic into a data layer, a function layer and Skill methodology, and hand the uncertain requirements to the runtime to understand and render in real time. With quant-researcher as a worked example.