The Leverage Myth

There’s this story going around that AI levels the playing field. Junior devs shipping senior-level code. Designers without training making professional work. The gap is closing, everyone’s getting democratized access, etc.

I don’t buy it.

What I actually see

AI multiplies whatever you already have. That’s it. A senior engineer with Claude can explore ten architectures before lunch because they know which outputs smell wrong. They catch the subtle bugs that won’t surface until production. They’re doing work that wasn’t even on the table six months ago.

A junior engineer gets code that compiles. Passes the obvious tests. Ships. And six months later someone’s untangling a maintenance nightmare because nobody caught the design flaw early. The junior couldn’t catch it because they’ve never maintained anything for six months. You can’t shortcut that.

More output? Sure. More value? I’m skeptical.

The math changed

I used to think a strong senior was maybe 3x a junior. Gross oversimplification, but useful enough. With AI tools, that ratio is closer to 10x. Could be higher, honestly. I keep revising it up.

I see the same thing with PMs. A senior PM with AI can run customer research and positioning work in parallel, stress-test multiple strategies before committing. A junior PM gets nicer meeting notes.

What this actually means

Here’s where I’ll probably annoy someone: if you’re early in your career, AI isn’t your shortcut. The judgment these tools require is exactly the judgment you haven’t built yet. That hasn’t changed and I don’t think it will.

If you’re experienced, you’re probably still using AI to do the same work faster. That’s fine. But the real unlock is doing things you couldn’t do before. Most people I talk to haven’t made that jump yet.

And if you’re hiring, recalibrate. AI won’t make inexperienced people productive. It’ll make your experienced people produce way more. Budget for fewer seniors doing more, not juniors doing senior work. That second thing isn’t happening.