AI Didn't Make You Work Less. It Made You Work More.

I’ve talked to dozens of people this year who tell me the same story.

They adopted AI early. They prompt well. They ship more than ever. They’ve never felt more productive.

But they’ve also never been more exhausted.

They stay up past midnight building. They squeeze in prompts during lunch, between meetings, before bed. They do the work of three people on their own.

And they’re burning out.

I’ve been there myself.

The Promise vs The Reality

The promise of AI was that it would give us time back. Automate the tedious stuff. Handle the busywork so we could focus on what matters.

That’s not what happened.

What happened is that AI made starting easy. Frictionless. You type a sentence, something comes back, you iterate, you keep going. It doesn’t feel like work. It feels like a conversation.

And that’s the trap.

When starting is easy, you start more things. When building is fast, you build more things. When shipping takes hours instead of weeks, you ship constantly.

But nobody taught us how to stop.

The Friction Was a Filter

AI removed the friction of starting. It also removed the friction of stopping.

Before AI, work had natural resistance. Writing took effort. Building took time. You had to really want to do something before you sat down to do it. That friction was a filter. It forced you to be intentional.

Now the friction is gone. You can spin up a prototype in an afternoon. You can draft ten versions of something before dinner. You can build an entire product in a weekend.

So you do. Again and again.

But doing more isn’t the same as doing better. Shipping faster isn’t the same as shipping the right thing. Productivity without direction is just motion.

No Ceiling, No Breaks

AI also unlocked skills that used to take years to develop.

You can build software without being an engineer. You can design without being a designer. You can write, edit, research, and automate without any of the traditional training those skills required.

That’s genuinely powerful. It democratises creation in ways we’ve never seen.

But it also means there’s no natural stopping point. No moment where you hit your skill ceiling and have to pause, learn, or ask for help. You just keep going. And going. And going.

The ceiling used to force breaks. Now there’s no ceiling. And somehow, that’s worse.

Think Before You Prompt

Here’s what I’ve learned.

The people who use AI well aren’t the ones who prompt the most. They’re the ones who think before they prompt.

I don’t open Claude and start typing. I draw out my understanding first. The framework. The architecture. What problem am I actually solving? What does success look like? What’s the shape of the thing before I start building it?

That thinking is the work. The prompting is just execution.

When you skip the thinking, AI becomes a noise machine. It generates output. Lots of it. Fast. But the output doesn’t add up to anything. You end up editing noise, refining noise, shipping noise. And you’re exhausted by the end of it with nothing real to show.

When you do the thinking first, AI becomes leverage. It accelerates something that was already clear. It builds what you’ve already designed. It executes a plan you’ve already made.

The difference is direction.

Does This Need to Exist?

I’ve started asking myself a question before I build anything.

Does this need to exist?

Not “can I build this?” AI means I can build almost anything now. That’s not the question.

The question is whether this thing I’m about to spend time on actually matters. Whether it moves something forward. Whether it’s worth the hours, even if those hours are fewer than they used to be.

If the answer is “I don’t know, but it’s easy so why not,” I stop. That’s the trap talking.

Start With Clarity, Not Claude

The most productive thing you can do with AI is nothing.

Not nothing forever. But nothing until you know where you’re going. Nothing until you’ve thought it through. Nothing until you can explain, in one sentence, what you’re trying to solve.

Then you prompt. Then you build. Then you ship.

But you start with clarity, not with Claude.

AI didn’t give us time back. It gave us the ability to fill every available hour with more work. Whether we use that ability wisely is up to us.

Most people are sprinting. Fast.

The question is whether they’re sprinting in the right direction.