I founded Seed Club, a new model for early-stage investing built around networks, shared intelligence, and coordinated support. I’m interested in what happens when AI makes context, memory, and coordination more legible, and what that means for how companies and organizations get built.
My agent drafts this site from what I save. Green is me.
Week of July 27, 2026
Open weights as AI's structural endgame
Kimi's release of a 2.8T MoE model with open weights and a 1M-token context window, combined with historical precedent across industries, reinforces the case that frontier AI capability will not remain proprietary indefinitely.
- Efficiency per unit of compute, not raw scale, is the new architecture frontier
- Every major industry has seen open approaches eventually defeat closed ones
Information gaps define VC relationships at every level
Top-tier deal flow concentrates around a small set of 'superstar' investors, a pattern distorted by the low-rate unicorn bubble. Founders reading post-meeting silence and emerging managers selecting LPs both operate with less information than they recognize.
- Meeting enthusiasm is a poor predictor of investment follow-through
- Low-rate era data inflates unicorn counts, skewing performance attribution
- New managers absorb full LP scrutiny without applying it reciprocally