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 17, 2026
The AI stack has no durable moat
Every layer of the AI stack, from frontier models to wrappers and inference providers, appears equally undefended against commoditization. When AI makes labor effectively free, value may migrate to resources with genuinely fixed supply.
- Data-driven VC pattern-matching favors safe bets, penalizes breakout startups
- Most AI market debates reduce to competing bets on market structure
Models, IDEs, agents, and app builders all lack durable moats · @nikunj Data-driven VCs back familiar startups but miss breakout successes · @credistick Six contested questions shaping the AI industry debate · @vasuman AI product categories are scaling faster than moats can form · '@nikunj' Data-driven VCs win in known categories but miss breakout startups · @credistick When agents make labor free, only truly scarce resources stay valuable · maxbittker AI debates sound ideological but are really market-structure bets · @vasuman
AI widens the gap between agency and passivity
AI amplifies existing differences in drive and judgment rather than equalizing them, a structural effect that policy interventions are unlikely to offset. Consumer-facing personal AI tools compound this by solving the scheduling and logistics complexity of the affluent, not the typical user.
- No regulatory framework is expected to flatten the resulting agency divide
Instinct and its peers are solving Silicon Valley problems · @VinIyengar AI will widen the gap between high-agency and passive people · @antoniogm AI amplifies gaps in drive and judgment rather than closing them · @antoniogm AI amplifies individual agency gaps that policy cannot close · @antoniogm Personal AI assistants are built for the time-poor rich, not most people · '@VinIyengar'
Agentic pipelines are moving from demos to revenue
Specialized bots coordinating on large tasks, social data pipelines bypassing login walls, and automated research-to-outreach workflows are generating measurable returns in production. Simple multi-step agent chains are proving disproportionate ROI relative to their technical complexity.
- Multi-bot coordination lets specialists self-organize on complex projects without human input
- Social access skills unlock agent workflows previously blocked by login walls
- Warm outbound from AI-qualified followers outperforms cold prospecting significantly
Multi-bot group chats tackle game development with specialized roles · @tonbistudio Hermes Desktop launches Bot Mode for multi-agent workflows · @NousResearch X follower research bot built in 20 minutes drives $3M in revenue · @mattepstein Workflows for using Grok Bot in go-to-market operations · @kristaletz How an automated AI content engine drove 4.6M Google impressions · @Trace_Cohen New tool lets AI agents scrape social media behind login walls · @coreyhainesco AI tools make customer clustering accessible with little effort · @danshipper The interesting use of chat-native AI tools is not occasional prompting · @kristaletz Grok turns the social timeline into a native GTM workflow · @kristaletz AI-qualified followers make for warmer outbound than cold prospecting · '@mattepstein' The useful version of AI SEO is not mass-producing generic pages · @Trace_Cohen AI makes it cheap to cluster and understand your customers · @danshipper The interesting part of AI SEO is not mass content generation · @Trace_Cohen