This crossover episode of No Priors and Latent Space features an interview with Microsoft CEO Satya Nadella at Build 2024. Nadella discusses Microsoft's ecosystem strategy for AI, emphasizing that true platform value comes from enabling others to create more value than the platform itself captures. He shares insights on the MAI models, the importance of private evals and proprietary data for company IP, and the future of agentic workflows and hybrid human-AI collaboration.
Summarized by Podsumo
Nadella reframes Microsoft's AI strategy as an 'ecosystem play,' not just about models or a single platform, but enabling every company to have 'frontier intelligence' using their own private evals and harnesses.
He reveals a key insight from internal teams: the Azure networking team redefined their work to 'build the agentic system that does Azure networking' (named 'Miles'), transforming their job from manual operations to a meta-level AI management role.
Nadella argues that the most valuable IP for companies in the AI era will be their private evaluation datasets (private evals) and the unique trace data from their own agentic workflows, not just the models themselves.
He discusses the importance of community benefits from data center buildouts, stating that unless the industry demonstrates tangible local value (jobs, lower energy costs, water replenishment), public permission will not be granted.
Nadella predicts a major shift in education towards new universities or pedagogies that directly tie AI-augmented learning to economic opportunity, calling this a potentially high-value startup area.
"The most critical thing is we need to tell the story and make it real that everybody has a real shot to participate as a first-class participant in this new economy."
"We need to earn that permission in a path dependent way. We can't wait. The world is going to be way skeptical of tech companies that say 'trust us, we've got it. The future is going to be glorious.'"
"The thing that is missing a little bit in all of our organizations is what is that new conceptual model of what can we build, what was impossible and what can we build."