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 August 24, 2026
Enterprise AI adoption lags model capability by years
Intelligence as a capability has effectively been solved at the model level, but organizational inertia means enterprises adopt new models far slower than labs release them, concentrating market power with labs and incumbents who control distribution.
- Application-layer power accrues to incumbents and labs, not pure-play SaaS.
- When cheap agents flood enterprises, new legal and governance frameworks will follow.
- Unexpected user prompts reveal what products should become, not what they are.
Altman admits economy adapts to AI far slower than he expected · @haridigresses AI models advance faster than enterprises can adopt them · @haridigresses Cheap AI agents may soon trigger new corporate legal frameworks · @Steve_Yegge Unexpected AI prompts reveal the product users actually want · @hnshah
SaaS companies will wall off APIs to protect agent revenue
Software vendors face a structural choice between becoming open API primitives or locking users into proprietary agents, and financial incentives will push most toward the closed, high-margin agent model, creating openings for rivals who stay open.
- Open API players could capture customers locked out by proprietary agent walls.
- Step-level evals, not just final-output evals, are needed to audit agent behavior.
- Describing an app's UX to an agent surfaces bugs invisible to human testers.
Venture conviction is rare and status crowds out merit
As funds reach top-tier status, the reputational cost of a contrarian bet rises sharply, so real conviction disappears and unusual founders get passed over, while VC rankings reinforce existing hierarchies rather than tracking actual performance.
- VC rankings are reverse-engineered from predetermined status outcomes.
- Tier-one fund status makes looking foolish more costly than missing a winner.
- A win rate above 35 percent signals pricing is too conservative.
Crypto-trained investors are built for narrative-only markets
A market environment with no fundamental anchors and pure speculative momentum mirrors the conditions crypto investors trained in for years, giving that cohort a durable edge as more of the broader economy shifts toward narrative-driven, exponential dynamics.
- Zero-fundamental market intuition is now a transferable and scarce skill.
- Angel investing offers asymmetric career upside even when financial returns are uncertain.