Why it matters
It is a working session, not a talk. You can run both coding agents on the same task and see differences in loops, context handling, testing and approval steps side by side with other developers.
The tokenmaxxing angle
The page lists measuring latency, token usage, cost and completion quality as a topic. Bring a real task, run it on both tools, and compare tokens and dollars per accepted change, not just which tool finishes.
From the organizers
Host Tony Siu, group Coffee & Code Philadelphia (C.O.D.E). Sat Nov 28, 1:00 PM ET at Turkish Brew, 1444 N 7th St. No formal presentation; bring a laptop and a repo. Topics include subagents, sandboxes and debugging failed agent runs.