Dave, self-taught AI dev in Toronto. Took a non-traditional path — photography, then Business Intelligence at 500px, then deep into agent systems the past couple years.

Building LoopGain, an open-source tool that measures when an agentic loop has stopped improving (loop gain / Aβ, borrowed from control theory) and stops it there instead of burning credits to max_iterations. Stops at convergence, rolls back on divergence.

loopgain.ai

Happy to talk agents, control theory, or anything in between.