Towards AI That Can Actually Interact
Summary
This episode covers OpenAI’s launch of DeployCo, a $4 billion consulting joint venture, and Anthropic’s crackdown on unauthorized secondary stock markets. It also details the White House walking back FDA-like AI regulations and Trump’s tech envoy to China. The main segment analyzes Thinking Machines’ new “interaction models,” which enable real-time, continuous human-AI collaboration rather than traditional turn-based chat.
Topics discussed
Intro, sponsors, and growth engineer hiring announcement
OpenAI launches DeployCo consulting joint venture
Anthropic and OpenAI crack down on unauthorized stock trading
White House walks back FDA-like AI regulation proposal
Trump's tech envoy to China and Jensen Huang's absence
Sponsor reads: KPMG, Granola, Scrunch, and Section
Thinking Machines introduces 'Interaction Models' concept
Architecture: Real-time interaction vs background reasoning
Demo capabilities: Visual proactivity and multitasking
Community reaction and the 'GUI moment' for AI
Unlock index, new benchmarks, and future implications
Listen ad-free on Castria