Do you mind?
Research on artificial consciousness1, built as a brain.
- Problem
- AI agents forget what happened, lose the thread, and grade their own work.
- Decision
- I built it the way a mind is organized: a hippocampus consolidates memory, a prefrontal cortex decides, and an intent-validator grades every task against its acceptance criteria — PASS or FAIL.
- Impact
- 360+ sessions archived · 300+ active memories · 43 schema versions · 4 core modules built, 8 more mapped.
PRACTICE — Claude Code skills, agents & hooks · Agentic workflows · Orchestration · Context engineering · AI-driven research
STACK — Claude Agent SDK · Claude API · own MCP tools · TypeScript · Node · SQLite + FTS5 search · 43 schema migrations · Three.js knowledge graph · Gemini Live voice · Playwright E2E