Why it matters
A concrete example of an agent loop running against local, zero-marginal-cost inference instead of a metered API — directly relevant to teams evaluating local vs. hosted routing.
The tokenmaxxing angle
pi.dev feeding specs to a fully local LLM to autonomously code, test, and iterate is a working case study in eliminating per-token API spend by routing agent work on-device.
From the organizers
Speaker Pratik Patel presents 6:45-7:15pm at Westside Motor Lounge, covering pi.dev's configuration, prompt engineering, and custom extensions for building a local "software factory."