See full schedule

Most conversations about AI independence focus on the model: who trains it, who owns the weights, and where inference runs. Samar Abbas, co-founder and CEO of Temporal, will argue that's the wrong place to look. The layer that matters is underneath: the execution infrastructure that decides whether a company can actually change course later. Frameworks and agent harnesses are cycling through versions faster than almost anything else in tech. Whatever framework looks essential this year will likely be replaced next year. Samar's argument: building capable AI agents means building on an open foundation that outlasts the model, framework, or harness sitting on top of it, and that no single vendor can take away. Drawing on Temporal's work running durable execution under some of today's most demanding AI workloads, he'll make the case that the real question isn't which model you pick, but whether your infrastructure survives whatever comes next.

Speakers