Dispatches
Page 3
42 posts total
- ·ai agents
AI Agents Broke GitHub. Microsoft Called AWS.
AI coding agents generated so much traffic that GitHub fell below its own enterprise SLA. Microsoft's fix: borrow capacity from Amazon.
- ·regulation
Colorado's AI Law Died Before Its Own Deadline
Colorado's landmark AI antidiscrimination law was gutted, stayed, and replaced before its June 30 deadline ever arrived. A case study in regulatory collapse.
- ·anthropic
The Government Pulled Fable 5 From the Cloud. Enterprises Are Rethinking Everything.
A US export control order forced Anthropic's Claude Fable 5 offline. The fallout is reshaping how enterprises think about AI infrastructure risk.
- ·anthropic
Claude Fable 5 Scores 95% on SWE-bench, Then Hands Off to Opus 4.8
Anthropic's new Mythos-class model leads on coding benchmarks but deliberately defers to a safer predecessor in restricted domains. That design choice says more than the score.
- ·model release
Google's DiffusionGemma Generates Text Sideways
Google released DiffusionGemma, a 26B open model that generates entire text blocks in parallel using diffusion, not token-by-token prediction. It's 4x faster and worse. That's the interesting part.
- ·anthropic
Anthropic Ships a Model It Says Is Too Dangerous to Ship Without a Leash
Anthropic released Claude Fable 5, a public version of the restricted Mythos-class model, built around a safety-classifier layer that routes high-risk queries to a weaker model.
- ·openai
ChatGPT's New Memory Doesn't Ask Permission First
OpenAI's Dreaming V3 builds a profile of you automatically from every conversation you've ever had. The architecture is clever. The consent question is harder.
- ·Apple
Apple Handed Siri's Brain to Google
Tim Cook's final WWDC keynote today confirmed Apple licensed a custom Gemini model to power a rebuilt Siri, a remarkable concession from the company that built its identity on owning the whole stack.