subscript

The passage around this formula

Language models became substantially more capable by thinking for longer at inference time. That option is bounded on a robot, and the bound is not a matter of taste. A closed-loop controller running at frequency f must complete sensing, computation and actuation within one cycle: $t_{\mathrm{sense}} + t_{\mathrm{compute}} + t_{\mathrm{actuate}} < \frac{1}{f}$. The pi-zero model is documented as producing actions at up to 50 Hz for dexterous tasks [ 3 ] , which leaves a 20-millisecond budget for everything. No amount of chain-of-thought fits in 20 milliseconds. Worse, missing the deadline is not merely slow: a controller that arrives late in a contact-rich interaction is acting on a stale world model, and in a stiff contact the resulting force error…

Read this part in the article →