Subscribe
Field Notes

Agents Aren't Ready, but the Tooling Underneath Is

Six weeks of agent failures and the substrate that emerged underneath.

↳ Build the rails, not the train. The boring middleware is where 2026's defensible companies will live.

I spent six weeks in late 2025 trying to ship an autonomous agent into a real workflow. It did not work. The interesting finding was not the failure — it was the layer that emerged underneath the agent, which is the layer worth building right now.

Build the rails, not the train. The boring middleware is where 2026’s defensible companies will live.

Full notebook coming soon.

I write to think.
You can read along.

I'm Rijul. I write essays, host a podcast, and build small things on the web — all of it in service of one question: how do we leverage AI in the next decade without giving away what mattered in the last? New work lands here when it's ready. Subscribe and I'll send it once.