sp spashii
journal about
2026
  • we needed a SaaS engineer, not an AI engineer

    January 30, 2026 •
    #hiring #startups #saas #strategy
  • pii redaction can't be a progressive enhancement

    January 29, 2026 •
    #privacy #architecture #echo #transcription
  • rescue culture trap

    January 29, 2026 •
    #management #hiring #leadership #startups

    hiring post-mortem

  • printable quick reference guides for ECHO

    January 28, 2026 •
    #product #ux #echo #real-world
  • rendering 1000+ conversation nodes without melting the browser

    January 26, 2026 •
    #react #performance #visualization #echo
  • sprint planning with Claude Code + Linear MCP

    January 26, 2026 •
    #claude-code #linear #sprint-planning #productivity
  • the field that didn't exist

    January 23, 2026 •
    #directus #debugging #schema-design #gitops
  • llm router instead of just retrying

    January 22, 2026 •
    #llm #architecture #echo #rate-limiting

    rate limits kept killing chat. built a router.

  • webhooks scoped to conversations, not projects

    January 22, 2026 •
    #architecture #webhooks #event-driven #echo
  • why we built a multi-model LLM router

    January 22, 2026 •
    #llm #litellm #vertex-ai #infrastructure #reliability

    Google's DSQ started throttling us mid-event

Page 1 of 5 Next →
spashii © 2026
GitHub Discord YouTube Twitter LinkedIn Email RSS