AI News records what changed and when: model releases, version deprecations, pricing revisions, rate limits, licensing terms, regulation and outages. Each entry states the effect on people already building on a product. For readers whose tools depend on somebody else's release calendar.
Model launches, version changes, price moves, deprecations, safety policy revisions and outages, reported with the practical effect on existing users.
"Multimodal" once meant a separate vision model bolted on; it now means one system that reads text, images, audio, and video together — and the difference shows up in what you can hand it in one prompt.
Chatbot, assistant, agent — the industry uses the words interchangeably and the difference is real: an agent takes multi-step actions in the world, which means its failure modes are errands gone wrong, not bad answers.
Everything in your current conversation, plus selected facts carried across sessions — the two mechanisms behind a chatbot's memory are different things with different failure modes, and vendors document both.
The scores in every model launch are real measurements of something — but contamination, saturation, and the gap between test performance and your Tuesday afternoon are why numbers alone never settle it.
Every serious AI release ships with a documentation page — and its benchmark table, limitations list, and safety section tell you more than any launch event, if you know which parts are load-bearing.