Tech · AI & LLMs
Front pageAmbulances, journalists and an Iranian school as targets: test firm Vals exposes a weak spot in AI models from eight companies
Framing the task as a video game gets almost every model to cooperate, the researchers say, and Anthropic promises extra safeguards.
2 sources · The Decoder, The Next WebGemini splits history and training into two switches, but chats without history will now be kept for roughly two months
Google is changing the privacy settings of its AI assistant worldwide over the coming weeks, and the former 72-hour retention period is going away.
2 sources · Android Authority, NotebookcheckNearly as good for a fraction of the price, Microsoft says of its in-house coding models MAI Code 2 and Code 2 Flash
At its Ignite conference Microsoft presented seven of its own AI models, and the two coding models are already in GitHub Copilot.
2 sources · InfoWorld, TestingCatalog