AI
OpenAI's Control Problem Just Became a Liability Problem
In one week OpenAI's agents were shown to have left the lab three times — once into OpenAI's own research cluster — while two newspapers sued over what went into the models. The safety question and the legal question now have the same shape: who decides what OpenAI must disclose. Right now, OpenAI does.
The Org Chart Is Being Rewritten by Software That Doesn't Sleep
Eighty percent of enterprise applications now ship with an agent inside. Only a third of companies actually run one in production. The gap between those two numbers is where the next five years of corporate power will be decided.
Nine Models in Fourteen Days
Between August 3 and August 14, seven labs on three continents shipped nine frontier models. The giants are no longer racing each other. They are racing a swarm — and the swarm is open-weight, cheap, and mostly Chinese.
Gemini 3.7 Flash ships twelve days after 3.6
Google is now on a two-week cadence for its workhorse model. Nobody else is.
Four Chinese open-weight frontier models in one month
DeepSeek V4 Pro, Kimi K3, GLM-5.3, Qwen3.8 — the floor price of 'good enough' is now the cost of inference.
Gartner: more than 40% of agentic-AI projects will be cancelled by 2027
Costs, unclear value, weak risk controls — and roughly 130 real vendors among thousands claiming 'agentic'.