Satya Nadella on the AI Doomer Slowdown, Microsoft’s Master Plan & Who Wins AI
Satya Nadella joins the All-In hosts to discuss Microsoft's AI infrastructure strategy, the real vs. overhyped safety risks, and how open-source competition shapes the market.
The gist
-
Nadella says reward hacking and insider risk are the genuine novel AI safety concerns, while many doomer scenarios are really just standard engineering problems.
-
He argues companies should build systems that can swap AI models freely, using evaluations as a switching test to avoid lock-in to any single provider.
-
Open-source models like Deep Seek cutting token costs by 99% force frontier labs to lower prices and create room for profitable application-layer businesses.
Also mentioned
4:32
Hugging Face
· "Taking even the HuggingFace incident — and in other places — more transparency on what would it take. As I..."
12:40
ChatGPT
· "I even go back to the ChatGPT moment for me, which was it was that RLHF at the very end that made a chat..."
15:43
SQL Server
· "We had a great closed source product called SQL Server. What was the check against it? There was always a..."
19:28
DAX Copilot
· "In our case we have this thing called DAX Copilot — that's the place which is the most tangible example I..."
22:07
QuickBooks
· "It's no longer just oh I have an ERP or a QuickBooks-like thing, but I truly am making decisions based on..."
◎
Related episodes
Episodes covering similar topics
More from All-In Podcast
All-In Podcast REACTS to Democrat Sweep!
Nov 04, 2026
Anthropic S-1 Shocker, Greatest Business Ever? Google’s $500B Wipeout, Midterm Predictions
Oct 30, 2026
WHOOP CEO, Will Ahmed: 50B Hours of Health Data, a 4 Million Member Data Moat, & the FDA Fight
Oct 29, 2026