A new signal emerged from this week's earnings reports that deserves more attention than it's getting: Twilio is becoming the picks-and-shovels layer for AI agents. That's not just a revenue beat. That's infrastructure emerging beneath our feet.
The Solo Operator Economy
The trend is already visible. Stripe Capital backed $85M for AI agent startups last quarter. Microsoft added agent functionality across their M365 suite. Atlassian's AI features are driving enterprise adoption earlier than expected. But the real story isn't at the top of the stack — it's at the foundation.
Twilio processed real-time communication infrastructure for AI agents. As agent-to-agent and agent-to-human communication scales, so does Twilio's take-rate. Jason Lemkin called it right: this is the "picks and shovels" play, except these picks are APIs for conversational infrastructure.
The unit economics of a one-person company just changed fundamentally. An individual with AI agents can now handle what previously required a 50-person org.
Medical Diagnosis as Proof of Concept
While we were watching earnings, OpenAI's o1 correctly diagnosed 67% of emergency room patients using only electronic health records and a few sentences from nurses. Human triage doctors hit 50-55%. That's not incremental improvement. That's a threshold crossing.
When an AI model can outperform trained physicians on diagnostic accuracy — with less information — the implications ripple beyond healthcare. Any knowledge work task becomes susceptible to the same pattern: more capability, less input required, faster iteration.
What Changes
The one-person company isn't a future projection. It's already here. We're seeing:
- Solo founders raising seed rounds — No need to hire a full team to prove the concept
- Support functions automated — AI agents handling customer success, bookkeeping, compliance
- Distribution decoupled — Product-led growth through AI-personalized outreach
The constraint is no longer labor cost. It's idea velocity and execution discipline. One person with the right AI stack can now ship what used to require a CTO, a head of product, and three engineers.
That should terrify companies betting on scale without automation. And it should excite anyone who's been waiting for the tools to catch up to the ambition.