Why it matters
Agent Harness tackles how coding agents retain context across sessions, a design choice with direct consequences for how much history an agent re-reads, and pays for, on every run.
The tokenmaxxing angle
Bishop Zareh's talk on persistent AI memory architecture is exactly the context-management pattern that decides whether an agent re-sends its whole history each call or reads efficiently, directly shaping token spend.
From the organizers
The page names speaker Laxmi P. Srinivas Sarva of Apex Fintech Solutions presenting BuildScanner, and Bishop Zareh presenting Agent Harness, at STATION Austin's 16th floor.