No hype. No vendor PR. No "AI will save us all" copium. Real conversations about what happens when intelligence gets cheap — and humanity gets expensive.
Microsoft and Carnegie Mellon just proved what we already suspected: the more you trust AI, the less you think. A new study on 300+ knowledge workers found that people who relied on AI for critical tasks showed measurably lower critical thinking scores. Here's the uncomfortable question nobody wants to ask — is this a bug, or is it the business model?
MIT Sloan's Eric So calls it "AI gravity" — the constant pressure to outsource more thinking to machines. A new paper in Computers in Human Behavior Reports confirms what pilots, doctors, and traders have been warning about for years: cognitive offloading isn't a productivity hack. It's a slow-motion skills collapse. And we're designing every new tool to accelerate it.
Anthropic's 2026 Economic Index analyzed millions of real Claude interactions. The headline: AI augments more than it replaces. But dig deeper and a different picture emerges — the workers who treat AI as a thinking substitute, not a thinking amplifier, are quietly making themselves redundant. The real fault line isn't human vs machine. It's between people who use AI to get better, and people who use AI to stop trying.
This article was drafted with AI assistance. Edited by AI. Proofread by AI. The meta-irony isn't lost on me. I'm writing about the dangers of AI dependency using the very tools I'm criticizing. And that tension — that uncomfortable, hypocritical, deeply human tension — is exactly the point we need to talk about. Because if even the critics can't think without AI anymore, what does that say about the rest of us?