Does that software really need to be built?

February 21, 2026 at 3:51 PM (UTC)

Of course I'm following the discourse about agents generating software and how it's suddenly really good and programmers are doomed and what not now. Software is my literal job. But also, the chatter keeps butting up against my craft, filling the spaces I used to use to practice it.

Maybe it's a programming nerd form of doomscrolling.

Anyway, one thing I've read has been sticking with me recently. I'm not even gonna quote it in any particular post because frankly, it's showing up a lot.

It's a narrative about software developers who have a lot of abandoned personal projects. Story goes, they are now suddenly able to tell Claude Code or somesuch, "make that project happen, I'm going to bed"—and they wake up and lo, their once-grand vision has been made real. Ish.

I am not doing this. In fact, I am finding my joy practicing my craft. I have created several things, all by hand, expressing myself in every line of code. It gives me life to be able to build these things I definitely want to see in the world, with my own hands and mind.

But sure, I have lots of abandoned code as well.

Last year, I zipped up the majority of it and stuffed it into my archive folder, so I could move it away from active stuff. Every last thing I filed away, I once had—at the very least—marked interest in. Some, I was really excited about!

Here's the thing. Not all of it needed to be built.

That's very different from saying that I didn't need to make a fresh directory, init a project, read through a bunch of websites, and write some lines of code. Everything I did was good for me. I learned, I thought, I imagined.

And usually, I moved on.

That's a good thing!

It shaped me. It trained me. It influences my craft, even today.

Some of it's really embarrassing to look at. I hoarded the zips anyway because sometimes, I actually do remember a thing and want to look at it, whether it's because I think it can help what I'm doing now, or maybe just to remember an idea.

Legendary awful person Marc Andreessen once said "software is eating the world". In the Wall Street Journal, natch. He was right, but it's also a fucking invasive species.

There's too much software. It's everywhere and it doesn't need to be. We're living in a world where it's (sometimes literally) killing us, and driving our humanity out. Some software gives us new and wonderful capabilities. Most of it flattens us.

There's too much pressure to justify our craft by the requirement to ship. It is not enough to try and learn and practice, gaining wisdom. We must be delivering, delivering, delivering, all the time. Deciding to end an effort, right thing or no, is failure.

I really like that I have evidence that I have worked through many ideas over my (too) many years in this craft. It lets me trace how I am learning and growing, revisit old thoughts from time to time. They did not need to become a Product™ to be worthwhile.

I needed to give them space in my mind for that time. That's all.

So no, I am not particularly eager to drop a couple hundred bucks a month to activate part of a datacenter nobody wants in their backyard for a bit, just to inflate my ego by adding some more software to the pile that's already too big. Not everything needs to take form in the universe.

I am quite happy having had those experiences—and moved on.