Claude Gets Cheaper, Meta Gets Hacked, Amazon Gets Mad
Anthropic just dropped a model that performs like its flagship but costs forty percent less. Meta's new AI assistant has a zero-day so bad it contradicts every security claim they made. And Amazon just banned Meta's shopping agent entirely. It's Monday in AI: launches, vulnerabilities, and a grudge match over who gets to decide what you buy. 1. Anthropic launches cheaper flagship-level model (Anthropic News) 2. OpenAI adds two more GPT-6 models (GitHub Copilot RSS) 3. Xiaomi open-sources new omnimodal coding models (TLDR AI) 4. Meta's Muse AI has critical security flaw (Ars Technica AI) 5…
The skinny
The skinny isn't ready yet — notes appear once the transcript is processed.