Two themes: (1) commercialization and scaling strains—AI products face quotas, ad-driven monetization, massive compute costs, shutdowns, and huge data‑center builds; (2) capability versus safety—agents enable powerful automation (useful and risky), but fragile harnesses demand stronger engineering, governance, and public scrutiny.
-
The Register: Anthropic admits Claude Code quotas running out too fast (Mar. 31, 2026)
Claude Code users report rapid quota exhaustion, disrupting work, and Anthropic is investigating. -
Neowin: Microsoft Copilot is now injecting ads into pull requests on GitHub (Mar. 31, 2026)
Microsoft’s AI tools are injecting ads into pull requests via hidden Copilot comments, promoting services like Raycast, Slack, Teams, and IDEs. -
Claude: Computer use is now in Claude Code. (Mar. 30, 2026)
Claude Code can open apps, click through user interfaces, and test what it builds, all from the command line. -
Axios: Everyone's worried that AI's newest models are a hacker's dream weapon (Mar. 29, 2026)
Anthropic’s unreleased model, Mythos, reportedly lets agentic AI autonomously plan and execute large-scale cyberattacks, making such strikes far more likely. -
WSJ: Sayonara, Sora: OpenAI Says Fun Time Is Over (Mar. 29, 2026)
OpenAI shut down Sora after massive compute costs and chip shortages forced a move from free spectacles to revenue-generating uses. -
WSJ: The Sudden Fall of OpenAI’s Most Hyped Product Since ChatGPT (Mar. 29, 2026)
OpenAI shut Sora, its consumer video app backed by Disney, after it proved unprofitable and consumed too many AI chips. -
WSJ: Ex-BP CEO Bernard Looney to Lead Data-Center Developer (Mar. 31, 2026)
Bernard Looney, former BP chief, is taking the CEO role at Prometheus Hyperscale to lead massive, AI-focused data-center projects in Wyoming and Texas, aiming sustainable, liquid-cooled facilities. -
NY Times: When Will Japan’s Cherry Blossoms Bloom? A.I. Can Help Answer That (Mar. 30, 2026)
AI helps Japanese forecasters analyze decades of weather data and thousands of crowdsourced photos to predict cherry blossom timing more precisely. -
Simon Willison: A quote from Georgi Gerganov (Mar. 30, 2026)
Local models often fail not because of the model itself, but due to fragile harnesses, chat templates, prompt construction, and inference bugs. -
Center for Humane Technology: “The AI Doc” Opens Friday. It's The Film We Need In This Moment. (Mar. 26, 2026)
The AI Doc opens in U.S. theaters, showing AI’s promise, peril, and power in a shared, urgent experience. It calls for joining the Human Movement, using the discussion guide, and acting together for a humane AI future. -
Microsoft Copilot: Copilot – Terms of Use (Mar. 31, 2026)
Microsoft’s legal team is clearly at work: “Copilot is for entertainment purposes only. It can make mistakes, and it may not work as intended. Don’t rely on Copilot for important advice. Use Copilot at your own risk.” -
WSJ: Buying the Dip? This AI Agent Will Do It for You (Mar. 31, 2026)
Public is rolling out AI agents to automate investing, execute trades, and apply orders like stop-losses, with users setting, reviewing, and pausing workflows. The tool passed Series 7 exams. -
The Fine Print: "CEO Said A Thing!" Journalism (Mar. 30, 2026)
US media often parrots CEOs without context or challenge, amplifying false, self-serving claims. This “CEO said a thing!” coverage props up corporate power, normalizes myths about tech leaders, and blurs news with marketing.