designing for agents you don't controlsome more learnings from pm-ing an mcpSeptember 10, 2026·4 min read
what the heck is a system of recordwhere i share what i think a system of record is in 2026, and why it's more than a google drive mcpAugust 15, 2026·10 min read
mcp is a very expensive linter for your apihow this new surface area highlights system design flawsJuly 15, 2026·5 min read
ai slop, shallow thinking, and the problem that was always therethoughts on calibrating your sources and exercising judgement while consuming online contentJune 19, 2026·6 min read
skills are a shamhow to stop treating a help center abstraction like a capability, and why the real compounding happens in codeMay 30, 2026·18 min read
the agent empathy canvaswhere i present the agent-equivalent of a user-empathy canvas that helps with understanding the new user personaMay 13, 2026·5 min read
the new user personabuilding product in a world that's soon going to be dominated by agentsMay 11, 2026·12 min read
Notes on Building a BrainWhere I document building Jaswiki, my personal AI-powered knowledge systemApril 18, 2026·18 min read
The Predictability Quadrant A framework I kept drawing on whiteboards, and then accidentally proved to myself on a Sunday afternoon. April 14, 2026·8 min read
LLM CommoditizationReflections from a round table with Anthropic's GTM team on model lock-in, commoditization, and what it really means to build for the future.October 13, 2025·3 min read
Customer Obsession (part I)Early thoughts on customers conversations while building 0-1September 14, 2025·2 min read