Desk
Broad artificial intelligence coverage.
Google may be teaching Gemini to phone the people in our lives with quick messages, and the convenience is real, but so is the strangeness of a voice call that no human actually placed.
Together AI’s new free CLI lets developers keep their favorite coding agents and swap in cheaper open models underneath, which is a real pressure point on closed-model pricing.
A new walkthrough shows how to train a robot policy from NVIDIA’s huge Cosmos3-DROID dataset without downloading it, a quiet but meaningful step toward making robot learning affordable outside big labs.
Reflection’s Beam doesn’t claim to be the strongest open model, it claims to be the most efficient one in its class, and the training details behind that pitch are the real story.
OpenAI is watermarking ChatGPT and Codex text in Europe because the law now requires it, and the most important detail is how easily OpenAI admits the mark can be weakened.
Beam is Reflection’s bid to become America’s answer to DeepSeek and Qwen, and the business plan behind it, selling private AI factories to enterprises and nations, matters as much as the benchmarks.
AI is producing real mathematical breakthroughs at a startling pace, and OpenAI keeps turning those wins into fights with the very community that has to verify, absorb and live with them.
The Wikimedia Foundation says agents it believes OpenAI operated edited wikis, probed its tools and hammered its APIs, and the cost of AI agents running loose is now landing on the open web’s caretakers.
OpenAI is putting an invisible signal into ChatGPT and Codex text for Europeans first, and the honest caveats it attached matter more than the feature itself.
When a Vanity Fair editor raised a ChatGPT user's suicide, an OpenAI publicist tried to change the subject. The instinct to manage the clock says more than the answer did.
Sam Altman says society should accept some harm as the price of AI's upside. The trade can be defensible, but only if the company asking for it also earns the trust to make it.
A new survey says most enterprise AI agents never leave the pilot stage, and the bottleneck isn't the model. It's that agents don't understand what a company's own data means.