I'm Richard, building Kaimon Labs — an AI services + superagent platform for knowledge work that gets continuously better the more you use it.

By day I am a decade-plus into the investment side, in VC and PE in US and Asia. Openclaw started it, engineering background supported it, and now have wired a bunch of agents.

- A news-agent that scores and writes a daily multi-language brief tuned to actual reading attention, not engagement bait

- A deep-research agent (DreamSearcher) that produces IC-memo-grade output with full source provenance — not summaries-of-summaries

- A content-injestion-agent that ingests youtubes, EPUBs and writes structured notes I'll reread and also send out daily summaries.

- An identity-agent that maintains a 22-year personal archive (Gmail + work email + calendar + contacts) for grounding the others — and, eventually, my kids and future AI  

The thesis: agentic work isn't a single chatbot. Its a personal substrate that compounds. Memory, voice grounding, frontmatter-gated confidentiality layers, idempotent pipelines, leak audits. Less Twitter-thread, more building-an-OS-for-one-person.

Based in Tokyo and California. Open to talking with builders working on agent infrastructure, deep-research workflows, or AI + investing. 😁