How You Do Anything Is How You Do Everything

It’s a beautiful principle—often undone by the reality of “everything.” There’s too much to do. Priorities shift. Corners get cut. Our pragmatism chips away at our perfectionism.

When we think of AI, the first benefit that comes to mind is usually productivity. Faster output. Cheaper labor. Fewer people (and ideally fewer meetings). It’s the usual pitch. We’ve seen this movie.

After the dotcom bust in the early 2000s, software development underwent a massive shift. Work migrated overseas, enabled by global fiber networks laid in the ‘90s and bolstered by rising technical education abroad. It was the birth of outsourcing at scale.

The outcome? Software became commoditized, and quality took a backseat. The firm I co-founded back then found its footing by helping companies clean up the mess. Our tagline was “Software as Craft”—basically a polite way of saying, “stop shipping garbage.”

That same pattern is repeating now with AI. If we’re not careful, we’ll focus on using AI to produce more of the wrong things—faster than ever. But there’s another way.

AI gives us back the space—mental and temporal—to do better work.

Here’s a quick example: I recently tripled the performance of a piece of software. It already worked, so under normal conditions, I wouldn’t have invested two days writing comparison scripts and tuning optimizations for a hypothetical gain.

But with AI, I validated the hypothesis and implemented the fix in 20 minutes. That’s not a humblebrag—I don’t have magical powers. Unless you count forgetting why I walked into a room.

We’re just getting started with AI as a thought partner. As we move forward, let’s shift the focus from doing more to doing better.

I love this quote:

If it falls to your lot to be a street sweeper, sweep the streets like Michelangelo painted pictures, like Shakespeare wrote poetry, like Beethoven composed music…
—Martin Luther King Jr.

Imagine the satisfaction of improving everything you touch—every day.
That’s what AI makes possible.
That’s what we should aim for.