Quiet morning, loud evening. The compute wars escalated, Anthropic published research that turns Claude’s “thoughts” back into English, and the […]

GPT-5.5 Instant goes live, Anthropic strikes a SpaceX compute deal and ships Wall Street agents, CopilotKit lands $27M, and Nvidia B300 prices double in China.

Big money, new defaults, and a quiet flex from Anthropic — here’s what hit the AI world while you were

Anthropic locks B in Google compute, Microsoft/Google/xAI hand models to US security testers, the Trump admin sharpens AI regs, and Reddit roasts Brockman in the OpenAI trial.

OpenAI lands on AWS, Anthropic teams with Wall Street, Claude stays ad-free, AI beats ER doctors in Harvard study, and Nvidia B300s hit $1M in China.

Markets are closed, models are still shipping. Here’s what hit the wires while you were eating dinner. 🏦 Anthropic links

Your fast-read briefing on what shipped in AI in the last 12 hours. Models, agents, and a courtroom curveball —

OpenAI x OpenClaw vs Anthropic, GPT-5.3/5.5 defaults, DeepSeek-V4, NVIDIA NeMo RL speedups, plus the hot take from r/ChatGPT.

GPT-5.5 hits AWS Bedrock and Databricks, the Pentagon signs classified AI deals with seven companies, The Atlantic argues the AI bubble might not be a bubble, and Moonshot ships FlashKDA. Plus the community verdict on GPT-5.5.