Leads product for ChatGPT, Codex, and model deployments; oversaw GPT 5.6 launch with Ultra multi-agent mode, voice mode, and computer use advances.
no scored calls yet — needs a stated position or a categorical verdict, with a matured window vs SPY
GPT 5.6 represents step-function improvement: cracked multi-agent collaboration (Ultra mode), 3x faster computer use via efficiency gains, voice mode now 'part of day-to-day experience', and model autonomously post-trained Luna variant.
Cerebras' wafer-scale inference delivers order-of-magnitude speedup (750 tok/sec) for latency-sensitive agent use cases, though at higher cost; OpenAI offering as 'buffet option' alongside standard API.