Aug 6, 2026
•
4 min read
PLUS: Meta finally releases Muse, its coding agent for large codebases, as Google courts Mechanize’s AI talent and new tests expose where AI agents still fall short.
Aug 5, 2026
PLUS: Samsung is redesigning memory, Chinese open-weight models are moving into deployment, and AMD is challenging Nvidia through open software.
Aug 4, 2026
WHILE: Palantir is betting on AI sovereignty, Kimi K3 is rewriting long-context serving, and agents are entering the systems where work gets done.
Aug 3, 2026
Preview
Aug 2, 2026
2 min read
PLUS: OpenAI’s containment findings, TSMC’s reported packaging push, and a practical test for agents that behave differently under observation.
Aug 1, 2026
Plus: Amazon completes its $50bn OpenAI investment, OpenAI’s Astra tackles 10 unsolved problems, and the real AI advantage may sit in proprietary context.
Jul 28, 2026
3 min read
Anthropic rejects a blanket open-weight ban; Microsoft splits cyber defense across agents; Claude artifacts surface in Google Search; and Stripe reportedly eyes OpenRouter as AI moves from models to control.
Jul 23, 2026
Anthropic just locked in 2GW of AI Compute, While OpenAI expands Camellia and Presence, and Google scales Gemini.
Jul 21, 2026
Possible US curbs, OpenAI’s long-horizon safety warning, Anthropic’s $1.5B copyright settlement and the infrastructure race underneath the agent boom.
Jul 18, 2026
Kimi K3’s infrastructure paradox, OpenAI’s new AI scorecard, Claude’s capacity rollout, Gemini 3.5’s delay, and America’s open-model dilemma.
