#ai AIL: 2

When AI beats procrastination

in Personal

I have a list. I’m sure you have a list too—a list of all the things that could be, if only you had more time. Time to explore the idea. Time to build a prototype. Time to invest in something new.

And if you’re anything like me, time is an excuse. Once I get started on a project, it becomes all-consuming. I find the time—10 minutes here, 10 minutes there.

If I'm being honest, the list isn’t full of hard problems. It’s full of things that never felt important enough to start.

Human vs AI

A year ago, I thought AI was pretty terrible. It was too slow. Too unpredictable. Too willing to confidently invent nonsense.

I’d look at its output, think “I could do this better,” then walk off, adding the idea to the growing pile of things I was technically capable of doing but never motivated enough to begin.

The mistake I was making was simple: I was comparing AI to a human.

I was asking, “Is this as good as if I did it myself?” and the answer was usually “no.”

But I was asking the wrong question. What I should have been asking was “would I ever do this at all?”

The breakthrough

I’ve been meaning to build a CLI that wraps listentotaxman.com so that I can compare two salaries for years. With a full-time job, a family, and other interests, it never quite made it to the top of the list.

So one Saturday evening, I took 10 minutes to write down what I wanted:

  • the commands
  • what each should do
  • a couple of example outputs

Then I walked away and made dinner.

When I came back, I had something that mostly worked. It wasn’t perfect. It’s definitely not something I’d ship to a customer. But for infrequent, personal use? It was perfect. Far better than the “this would be nice to build someday” alternative.

This was the breakthrough for me. It’s not “AI vs. Michael.” It’s “AI doing the work I never make time for.”

Getting started is the hard part

Realising this changed how I approach my backlog.

  • I try more things, because failure is cheap.
  • I finish more things, because “good enough” is easy to reach.
  • I revisit old ideas, because they’re no longer competing with everything else for my time.

I used to have more ideas than time. I had to choose between spending an evening at my desk or with my family. That's no longer the case.

In two months, I built six tools I’d been putting off for years:

None of these are groundbreaking. And honestly, that’s the point. I couldn’t justify the time investment. But now they exist where they didn’t before.

AI is still untrustworthy

Let’s be clear: I don’t trust any of the code. It's brittle. I’m certain there are bugs I haven’t found (mostly because I haven’t looked very hard). AI has made assumptions that I haven’t validated.

AI is not how you build things you care deeply about.

It’s how you build things you want to exist.

AI doesn’t replace deep thinking and craft. It’s a shortcut around the part where nothing happens at all.