Vibe-coding is not a strategy. It is a speed tool.
Yes, you can get a working prototype into users' hands in hours. But someone still has to decide what to build and why. Someone still has to talk to users, find the real friction, and determine what problem is actually worth solving. AI does not do that. It executes.
UX research is not a luxury you cut when you go AI-forward. It is the thing that tells you whether what you are about to build this weekend is worth building at all. Also, will there be a return!
Speed without direction isn't working fast it is working in a hurry.
That is what I am helping fix this week with a client.