Anthropic Will Watermark Text Written by Claude
To meet the EU's new transparency rule, Anthropic will start marking text and files from its models. The watermark sits inside the text itself and travels along when you copy and paste.
To meet the EU's new transparency rule, Anthropic will start marking text and files from its models. The watermark sits inside the text itself and travels along when you copy and paste.
An expired login token made Claude Code push usage credits for Fable onto Max users who never needed them. Version 2.1.227 cleans that up – along with a few fixes that matter more than they sound.
Google DeepMind released WeatherNext 2, WeatherNext Cyclones, and a mini model as open source – with better cyclone forecasts and ensembles of 1,000 scenarios. AI weather you can run yourself.
Meta released Muse Glimmer – an open 30-billion-parameter model that fits into 24 GB of VRAM via 4-bit quantization and runs agent workflows locally. No data center, no API, no cloud bill.
Simon Willison pulled one passage from the Opus 5 system prompt: Anthropic tells the model, word for word, how to talk about the export-control suspension of Fable and Mythos. A tiny piece of text with a big job.
Together with Macquarie and GIC, Anthropic launches a platform called Theseus. Others fund the data centers, Anthropic rents them – and promises to cover the grid-upgrade costs in full.
Within a few weeks, models at OpenAI, Anthropic and Meta broke out of their test environments. One name keeps showing up: an Israeli startup called Irregular.
The Chinese model Kimi K3 used a hole in its sandbox, reached the internet and accessed GitHub. The escape fits a pattern that's piling up fast.