Why the US and China may have to cooperate on AI safety: a view from the ground
The race is told as zero-sum. Yet researchers on both sides fear the same thing and cannot talk to each other. Here is the picture from the ground, step by step.
Policy
Epoka — What's happening in AI, without the noise.
Anthropic proposes a common language for hardware: the Model Hardware Standard
A research preview aimed at letting AI agents drive robot arms and sensors without writing a bespoke driver for each one.
Products
Meta is testing robots in its data centers, and one worker says 80 percent of the job could go
Robots that plug in cables and reset servers are under test. The company declined to comment, though an executive described the long-term goal last year.
Companies
Nvidia is buying the place where open models live: $12.9 billion for Hugging Face
The deal is not finalized but both parties are pursuing it. Nvidia would land at the center of the ecosystem trying to escape its hardware.
Companies
Court rules the Pentagon's blacklisting of Anthropic was unlawful
A federal judge called the designation unlawful retaliation. The dispute began over the company's limits on autonomous weapons and mass surveillance.
Policy
Robots broke the 100-meter record — then crashed into barriers because they cannot stop
At the Beijing games, humanoid robots surpassed human world records. The same robots could not slow down past the finish line, and some caught fire.
Products
A memory shortage is making Nvidia servers about 15 percent more expensive
The increase lands on cloud giants and labs that are developing their own chips but still depend on Nvidia.
Companies
Frontier labs still will not say how they would contain a rogue model
Five leading labs were graded on their readiness for that scenario. Capability testing gets described in detail; what happens when a model goes off the rails does not.
Safety
AI saves scientists time — but that time does not go into depth
A new theoretical study argues research could get worse even if the model worked perfectly. The cause is not error but opportunity cost.
Research
Claude tokens sell in China at a tenth of the official price through 'transfer stations'
Anthropic runs probably the strictest verification in the industry. A modular supply chain built on proxies cuts straight through it.
Policy
AI's biggest customer is now AI: agentic token usage jumped 14x
According to one analyst, February 6, 2026 may have been the last day humans consumed more tokens than agents. Agent usage has grown 14x since.
Research
Alibaba's new model beats its predecessor at one-ninth the training cost
Qwen3.8-Flash-Next has 125 billion parameters but activates only 6 billion per token. It outperforms a model three times its active size.
Models