⚡ Tech Tech & People — May 31, 2026
Anthropic's KAIROS always-on agent revealed in source leak, Meta's cloud signal reshapes AI infrastructure, Waymo's 6th-gen cuts sensor costs, and ADL study shows no model escapes bias.
Technology & People
AI Developments
Robots, chips, and the technology shaping our future.
⚡ Tech 
⚡ Tech MCP Is Dead? The AI Tool Standard That Ate Its Own Context Window
The Model Context Protocol was supposed to be the USB-C of AI. Then Perplexity's CTO abandoned it at his own conference, and Cloudflare showed it could be replaced with 99.9% fewer tokens.
⚡ Tech Free Home Cleaning — If You Let a Camera Film Everything to Train Robots
Shift will clean your home for free. The catch: a camera on the cleaner's head records everything you do, and that footage trains the robots coming for your job.
⚡ Tech Technology & People — May 30, 2026
Anthropic deploys to 276K KPMG staff, NZ's early adopters report real gains, Māori data sovereignty inspires new AI voice models.
⚡ Tech Daily Tech & People: May 29, 2026
Claude got smarter overnight. The AI bottleneck is now data centers, not GPUs. EAGLE 3.1 quietly makes every model faster. And new research shows AI agents can't actually do what we're paying them for.
⚡ Tech Daily Tech & People: May 28, 2026
Anthropic opens the containment kimono. DeepMind solves what humans couldn't for 56 years — for a few hundred dollars. Vibe coding gets a security reality check.
⚡ Tech AI Will Read Breast Cancer Scans in NZ From 2027 — Replacing One of Two Human Readers
AI is becoming a breast cancer screener in NZ. Starting 2027, it will replace one of the two human readers in mammogram assessment. The government says AI 'supports, not replaces' clinicians. Privacy advocates have questions.
⚡ Tech The VibeSec Reckoning: Martin Fowler's Team Found the Cracks in Vibe Coding
ThoughtWorks' own marketing team built a vibe-coded video tool. When they tried to scale it, AI recommended insecure configurations at every turn. The VibeSec reckoning is here.
⚡ Tech YouTube Will Now Automatically Label AI-Generated Videos
YouTube is done trusting creators to label their own AI content. Its systems will now automatically detect and label photorealistic AI video — and you can't remove labels from YouTube's own AI tools.
⚡ Tech AI Makes You Write Better Code — If You Slow Down
Nolan Lawson's viral essay turns the 'AI slop cannon' narrative on its head: use Claude, Codex, and Cursor Bugbot to find bugs, not ship faster. The result is better code, more side-quests, and zero velocity gains — and 695 Hacker News points suggest a lot of developers agree.
⚡ Tech Technology & People: May 27, 2026
The Vatican just became a major voice in AI governance. Anthropic left 3,000 internal documents open by default. Uber's AI ROI problem goes public.
⚡ Tech Uber Burned Through Its AI Budget in 4 Months — Now It Can't Justify the Spend
Uber rolled out Claude Code to 5,000 engineers in January. By April, the 2026 AI budget was gone. Now the company president says he can't draw a line between token consumption and useful features. Is the enterprise AI spending bubble about to pop?