Microsoft AI chief thinks superintelligence is near, but won't take your job
Summary
Nilay Patel interviews Microsoft AI CEO Mustafa Suleiman about the company’s strategic shift toward independent frontier model development, highlighted by the new MAI Thinking 1 release. They discuss the evolving partnership with OpenAI, Microsoft’s rejection of model distillation in favor of original research, and the path to superintelligence. The conversation also addresses growing consumer skepticism, the need for social permission, and whether current AI products deliver enough value to justify their societal impact.
Topics discussed
Intro and sponsors: ServiceNow, Klaviyo, Thrive
Mustafa Suleiman's new role and Microsoft AI structure
The split from OpenAI and building independent models
Microsoft Build, new models, and agile development cycles
AGI, human-level performance, and measuring AI progress
Sponsors: Shopify, NVIDIA, Rippling
MAI Thinking 1, distillation, and training data ethics
Enterprise focus, consumer anxiety, and healthcare AI
Sponsors: Klaviyo, Geico
Job automation, tasks vs roles, and AI's PR problem
Enterprise adoption, token budgets, and future devices
AI consciousness, suffering, and the singularity
Outro and sponsors: Geico, Fetch Pet Insurance
Listen ad-free on Castria