According to CNBC, China warned of back-door risks in Anthropic’s Claude Code as Anthropic accused Alibaba of illicit capability extraction.

According to CNBC... China flags Anthropic Claude Code as a backdoor risk, prompting enterprises to review model governance and software supply chain controls.

According to @CNBC, a study found popular AI chatbots give inconsistent, error‑prone personal finance advice, risking consumer losses.

According to TheRundownAI, DoorDash’s DashBench shows Kimi K2.6 plus Claude Fable 5 catches 65.2% bugs and 8 of 10 critical at $3.81 per review.

According to @claudeai, Cowork now syncs tasks across web and mobile, with a beta rolling out first to Max plan users over the next weeks.

According to The Rundown AI, Anthropic found a hidden Claude workspace, Tencent open-sourced Hy3, Replit sped app builds, and Kyutai demoed playable RL.

According to emollick, Reuters says China may curb overseas access to top AI models, signaling fewer frontier open weights and tighter controls.

According to @CNBC, Chinese models like DeepSeek and Zai narrow performance gaps with Anthropic and OpenAI while costing far less, boosting U.S. adoption.

According to @emollick, MTurk is fading as LLM use contaminates human survey panels, forcing researchers to adopt AI-detection, audits, and new platforms.

According to @CNBC, US firms test Chinese LLMs to cut inference costs as OpenAI and Anthropic prices rise, shifting pilots for customer ops and RAG.

According to @bcherny, Anthropic shares Claude Code’s origin from safety research and early user feedback in a new history feature video.

According to @claudeai, Anthropic details Claude Code’s origins, early user feedback, and product milestones in a new behind the scenes video.

According to @CNBC, Alibaba barred Anthropic tools after a distillation attack allegation, signaling compliance and IP risk concerns for enterprise AI.

According to emollick, Anthropic reports Claude shows a global workspace divide akin to conscious access, with interactive viz via Neuronpedia.

According to Anthropic... New research shows a Claude global workspace akin to conscious access, enabling interpretable control, as reported by Anthropic.

According to @CNBC, Anthropic leased TeraWulf’s Kentucky data center to expand Claude training capacity and secure power for rapid model scaling.

According to Ethan Mollick, large scale management training for AI agents could mirror WW2 programs that boosted US post war productivity.

According to StanfordAILab, LLMs align on each other’s answers over truth, so polling fails; ICML paper finds truthfulness does not scale without verifiers.

According to KyeGomezB, high AI spenders boosted headcount 10.2% and entry roles 12% in two years, signaling hiring shifts after AI capex.

According to @_avichawla, a 50+ Claude Code command cheat sheet maps workflows for setup, review, automation, and reporting, improving developer velocity.