Max and Sofia unpack Turbo AI’s sprint to 5M users, Google’s potential multi‑tens‑of‑billions cloud deal with Anthropic, and OpenAI’s prompt injection warnings for its Atlas browser. They dive into the data‑center energy crunch fueling aero‑derivative jet‑engine generators, a quick‑fire on national‑scale telecom reliability with Ibikunle Peters, and Mohammad Adnan’s pragmatic AI strategy from cold‑start fixes to mentorship. The duo cover a “brain rot” study showing low‑quality data degrades LLMs, break down multiple linear regression in plain English, and explore OpenInfra’s stack for Confidential Computing with Kata Containers. Plus: nine Indian AI startups to watch, Microsoft CEO pay in an AI‑charged market, the AI bubble debate, Apple’s M5 chip as an on‑device AI booster, and a WearOS quality‑of‑life upgrade. Three takeaways close the show: build augmentation first, prioritize reliability and security, and obsess over data quality.
Sources:
