A French Startup Just Pantsed Nvidia With Free Software
ZML just gave away a free tool that runs your AI models fast on any chip, not just Nvidia’s. It is a small, well-aimed kick at the CUDA lock-in that keeps the whole industry hostage.
ZML just gave away a free tool that runs your AI models fast on any chip, not just Nvidia’s. It is a small, well-aimed kick at the CUDA lock-in that keeps the whole industry hostage.
ETH Zurich built a fingernail-sized quantum chip that stores data as tiny mechanical vibrations, like a guitar string. More memory, longer coherence, smaller footprint, and it ran real algorithms.
Meta is building its own silicon to escape Nvidia, chasing 14 gigawatts of compute. Smart business. Also another few gigawatts of power and a lot more sand nobody wants to talk about.
Startups raised a record $510 billion in six months, more than the whole world managed in all of 2025. Two AI companies took 43% of it. The numbers are genuinely unhinged.
Quantum computing normally requires a PhD and a language that looks like hieroglyphics. MIT and IBM just made a version where you can tell the machine what you want in English.
For a decade the humanoid robot has been arriving “next year.” In 2026 it finally clocked in for a real factory shift. Where things actually stand, minus the Musk-grade hype.
Palantir’s Peter Thiel stood on a stage and accused the actual Pope of working for the Chinese Communists, because the pontiff had the nerve to suggest AI might need some rules.
SambaNova just raised a billion dollars to attack Nvidia where nobody is looking: inference, the unglamorous work of actually running AI models. Here is why that might be smart.
The Federal Reserve just handed a venture capitalist with billions staked on AI a seat studying AI’s effect on jobs. He also happens to be the Fed Chair’s college friend.
ESPN brought back the World Series of Poker with an AI sidekick that reads players’ blink rate and posture to call their bluffs on air. The poker face may be an endangered species.