OpenAI has launched DeployCo, a $4 billion joint venture that embeds forward-deployed engineers directly inside enterprise clients to drive AI transformation. Both OpenAI and Anthropic issued warnings this week about secondary-market tokens tied to private stock, flagging them as potentially void and legally hazardous for buyers.
Thinking Machines is the story worth reading in full. Their new interaction models process continuous audio-visual micro-turns, meaning the AI responds to real-time visual and audio cues without waiting for a discrete prompt. The companion background reasoning model enables proactive responses, not reactive ones. Real-time translation and anticipatory visual cues are the immediate applications, but the underlying architecture signals a structural shift in how human-AI interaction is designed.
The secondary-market token warnings from two major labs in the same week is not coincidence. That coordination, combined by a $4 billion deployment bet and a new interaction paradigm from Thinking Machines, makes this a week where infrastructure, legal risk, and interface design all moved at once. The full episode context on each thread is worth your time.
[WATCH ON YOUTUBE →]