OpenAI Team: "Your Agent Is Slower Than it Should Be" | Steve Coffey, Joe Gershenson, Ahmen Ibrahim
In this episode of MTS, Steve Coffey joins to discuss the Responses API at OpenAI, how WebSocket integrations improve agent efficiency, and why coding models have made him 10X more productive. Then, Joe Gershenson and Ahmed Ibrahim join to discuss token efficiency in Codex, code mode for reducing tool call latency, and designing harnesses that let models express themselves more fully.
--
https://x.com/stevendcoffey
--
MTS is a live news and interview show covering technology, business, politics, and culture as it happens.
Follow MTS:
X | https://twitter.com/MTSlive
YouTube |…
The skinny
The skinny isn't ready yet — notes appear once the transcript is processed.