The GitHub Abandonware Era

Last weekend I built something that would have taken me two weeks in 2019.

It had auth. Tests. Decent structure. Typed responses. A tidy UI. Nothing revolutionary, but solid. The kind of side project I would have been proud of a few years ago.

It took a few hours.

That’s not a humblebrag. It’s just the reality of working with AI tooling now. Sit down with Claude, sketch the idea, scaffold the boring parts, refine. The friction to get to “something working” has almost disappeared.

And that’s exciting.

But it’s also made me slightly uncomfortable.

I used to feel like if I built something, it really had to earn its existence. Time was the filter. Energy was the filter. If I pushed something to GitHub, it meant I cared about it enough to wrestle it into shape.

Now the bar to start is tiny.

So I’m starting to question the instinct to publish everything.

GitHub increasingly feels like a stream of output. Another note app. Another Things3 clone. Another minimal component library. Another starter kit. Most of them are technically fine. Typed. Tested. Sensible structure.

But do they need to exist?

When I think about the open source projects I’ve relied on for years, they weren’t weekend experiments. They were tools someone stuck with. They solved a real problem. They were maintained. They matured.

AI removes the cost of building.

It doesn’t remove the cost of maintaining.

Maintenance is still slow. Stewardship is still long term. Backwards compatibility still hurts. Community still takes time. Longevity is still expensive.

I’m not against building quickly. I love it. I’m more productive than I’ve ever been.

I just don’t know if productivity alone is a good enough reason to publish.

More code is being written than ever. I’m not convinced more durable software is being created at the same rate.

Maybe the constraint now isn’t skill.

Maybe it’s deciding what’s worth sticking with.

Not everything I build needs to exist in public.