Platform lock-in isn't going away — it's going invisible. How headless systems and AI agents are changing the dependency relationship, and what to do about it.
Are we asking the wrong question about AI and work? Three concepts from Indian philosophy — what holds, what dissolves, and what you actually do — reframe everything.
Sequoia argues hierarchy is just information routing. They're right — but hierarchy carries five other things that AI can't replace. Here's what the winners will do differently.
Token prices have collapsed 265x in three years. The lock-in risk in AI isn't the model — it's everything you build around it.
A single generalist agent is a single point of failure for a different reason than reliability. It’s a failure of fit. The answer: not one agent, but a mesh.
The people who understand AI best are all quietly building their own personal versions. Not using the products. Building. There's something worth paying attention to in that.
Get new essays by email.