Microsoft's Satya Nadella takes a veiled swipe at Anthropic and other AI model makers
THE SO WHAT
Public pushback on distillation complaints from a major platform player shows that IP norms for models are still contested—and that large distributors may favor more permissive reuse. If you rely on closed models or plan to license your own, revisit your threat model around distillation and negotiate contracts assuming this tension won’t be settled quickly.
READ THE SOURCE
MORE FROM THE WIRE
Applied AIApple Sues OpenAI, Apple’s Real Problem
A trade-secrets lawsuit between Apple and an AI lab is less about one employee and more about how messy talent and IP flows have become at the frontier. If you’re hiring from big tech or labs, your IP hygiene, onboarding, and data access controls need to assume this level of legal scrutiny.
Applied AIDeepSeek AI review
Frontier-class performance at a steep discount with unresolved privacy questions is exactly the trade many teams will be tempted to make under budget pressure. If you’re evaluating DeepSeek, separate experiments from production use and get explicit about what data is allowed to touch which model endpoints.
Applied AIMicrosoft admits its carbon emissions grew 25 percent in 2025, including a rise in fossil fuel use
A 25% emissions jump tied to AI workloads at Microsoft—mirroring Google and Amazon—means AI growth is now visibly colliding with corporate climate commitments. If you’re scaling AI infra, expect board-level questions on power mix and efficiency, and start logging emissions per workload like you log cost.
Anthropic's latest big-name hire: Monzo cofounder Tom Blomfield
A fintech founder and YC partner joining Anthropic’s compute team shows how AI infra is pulling in operators who’ve scaled regulated, high-uptime systems. For founders, it’s a reminder that the competitive frontier isn’t just models—it’s the operational discipline around compute, reliability, and capital allocation.