• [1]Home
  • [2]About
  • [3]Archives
  • [4]Bookshelf
  • [5]Projects
  • [6]Stories
  • [7]

The time to unmaintainable is very low

January 23, 2024

It’s so easy nowadays to get up and going on a project. I can burp some npm
commands into my terminal, burp some more to setup a deployment pipeline and
blam! Website. The time to product demo is so low. You can get far on your own…
very quickly… but then… you’re on your own. And it’s possible you’ve built
something way past your ability to maintain.

To fix that problem or to allow you to do more business’y work, you add people
to the process. But people cost lots of money. Unless what you’re building
makes lots of money already, you probably have to raise VC. With a squad of
developers burping npm commands, now you’re moving super fast. Hopefully money
starts coming in, but the whole point of investment was so you could price the
application artificially cheap to grow a user base that you upsell later. As
the investment coffers tick down, the pressure goes up to land a new feature or
gimmick so you can get another round of investment.

You can, of course, ride the hype cycles. This year that’s AI. Investors might
throw money at that. But now you’ve bolted on a feature from a highly volatile,
emergent, non-deterministic space into your overgrown application. You’ll
either need to hire more expensive people who know the space or divert existing
resources who were performing maintenance over to the new feature.

Oops, AI costs more than you can charge for it. You burned through the
investment even faster and must do a round of layoffs. Now your app maintained
by 10 people has 8 people… then 5… then 2…

Paying people to burp npm commands is expensive, could AI do that? Vercel’s [8]
v0, for example, farts out entire UIs. Great. In a day I have twelve thousand
screens built. I have more UI than some developers will code in a lifetime. But
I’m getting the itch to update it and the machine isn’t doing what I want it
to… perhaps I’ll take some VC to hire some people to clean up these robo-farts.

And by the time you finish all that work, it’ll be right in time for a major
version update on a core dependency. Good luck out there.

I realize I’m complaining about moving too fast but that’s not my intent.
Although, I could argue that while driving 200mph is fun and exciting, you’re
one small fuckup away from a major fuckup. My point is that a key factor of
sustainability is making sure maintainability stays on par with growth. At the
risk of sounding like a Luddite – [9]which I am – the ability to fancy
copy-paste your way into an unmaintainable situation is higher than ever and
that’s a trade-off we should think about.

© 2024 Dave Rupert • [10]Mastodon • [11]Twitter • [12]Github


References:

[1] https://daverupert.com/
[2] https://daverupert.com/about/
[3] https://daverupert.com/archive/
[4] https://daverupert.com/bookshelf/
[5] https://daverupert.com/projects/
[6] https://daverupert.com/stories/
[7] https://daverupert.com/atom.xml
[8] https://v0.dev/
[9] https://amzn.to/3NA62WE
[10] https://mastodon.social/@davatron5000
[11] http://twitter.com/davatron5000
[12] http://github.com/davatron5000/