AGI Is Old News
Nvidia’s Blackwell architecture, announced in March 2024, only reached consumer‑ready models by late 2025, a roughly twenty‑month lag from keynote to usable intelligence. Early Blackwell‑powered models like Anthropic’s Claude Mythos have already sparked emergency meetings among central banks due to perceived safety risks, while OpenAI’s upcoming "Spud" model finished pre‑training on the same hardware. Nvidia is now unveiling Vera Rubin, a chip promising five‑fold inference cost savings, and previewing Feynman, a die‑stacked design slated for 2028. The author argues that hardware advances, combined with rapid software optimization, create a super‑exponential growth curve for AI capability.
3 Trillion-Dollar Tech Stack Titans
The post argues that the next wave of AI dominance will be decided by companies that own the entire technology stack, not just the models. Nvidia, SpaceX, and Anthropic are highlighted as exemplars of vertical integration, each covering hardware, software,...