Models Don't Have Agency. Systems Do.: Structured Output Is What Makes Agents Work
sethjuarez · x · 2026-09-21
In a long-form essay, Seth Juarez argues that a language model only ever emits tokens and has no agency in the philosophical sense — the agency lives in the runtime built around it. When asked to book a flight, a model produces a friendly sentence a program can't reliably parse: the words are there, the shape is not. The practical fix is to force structured outputs (e.g. { destination, date, time }) instead of prose, so the system can act on named fields. The useful question isn't how to give the model agency, but how to shape what it's allowed to say.
More from AGI Musings
- Citizens Advice urges 'right to talk to a human' as half of users find service chatbots unhelpful — nordicinst · 2026-09-22
- Every major AI company is converging on the same personal agent stack, with little differentiation — signulll · 2026-09-22
- NYT Essay: AI Slopware Is Everywhere Now — and Nobody Is Using It — round · 2026-09-22
- What happens to Amazon's ad business when AI agents do the comparison shopping? — firstadopter · 2026-09-22
- Dev's Reversal Take: Tickets Will Use Agents to Find Buyers, Not the Other Way — cocktailpeanut · 2026-09-22
- Vacuum tubes weren't built by people trying to make computers — generativist · 2026-09-22