I've got a filter I run every AI headline through: who does this actually help, and who pays for it? It's simple, and it cuts through most of the noise.
Three questions
When a new tool or model drops, I ask:
- Does it solve a problem a real person has? Not a demo problem — a Tuesday-afternoon problem.
- Who's on the hook when it's wrong? If nobody can answer that, be careful.
- What does it cost the people around it? Jobs, energy, attention, trust.
Most of the hype cycle can't survive those three questions. The stuff that does survive is usually worth paying attention to.
The block test
The other thing I keep coming back to is what I call the block test: would this hold up if you explained it to the people on my block? Not dumbed down — just honest. If the value only makes sense in a pitch deck, that tells you something.
None of this means being anti-AI. I use these tools every day. It means refusing to be sold fear or fantasy. The technology is real. So are the tradeoffs. We're going to keep talking about both.