Measuring agent autonomy in practice

AI agents are here, and already they’re being deployed across contexts that vary widely in consequence. But “autonomy” is still a slippery concept.

Studying agents in the wild

This is placeholder content to validate typography and structure. We will replace it with real posts later.

A math snippet

Inline: $E=mc^2$
Display:

$$\int_{-\infty}^{\infty} e^{-x^2} dx = \sqrt{\pi}$$

Code (card-style)

export function hello(name: string) {
  return `Hello, ${name}`;
}

Results and discussion

If we treat autonomy as a set of observable behaviors—longer unassisted runs, better recovery from errors, fewer confirmations—we can evaluate systems more consistently.

Closing thoughts

This is a sample post used to build the platform.