The tools are
ready. Are we?
Why I think the hard part of AI adoption is now us.
A year ago, building the Whist Scorer would have been a weekend project: schema, state management, deployment, the lot. Last month it took twenty minutes. The tools got there while I wasn’t watching.
Yet most people I talk to use the same tools to write slightly faster emails. The gap between what’s possible and what people actually do with it is the widest I’ve seen in my career, and what’s missing is a skill most of us haven’t learned yet. Now the hard part is describing what you actually want.
The people I see getting the most out of these tools are good delegators. They know what they want, hand it off, check what comes back and correct it without micromanaging. That’s a management skill.
What changed
The models got good enough to hold a whole small app in their head. A wrong turn now costs seconds rather than hours. And you can say what you want in plain English instead of writing a spec.
So trying an idea is cheap now, and I try a lot more of them. That’s most of what’s on this site.
The tools are ready. Are we? Going by the people I talk to, not yet.