Post by @glyph@mastodon.social

One of the recurring tropes that bums me out *so* much about the process of AI sludge devouring software _specifically_ — with practitioners of the craft enthusiastically participating more than the practitioners of any other — is that the *self*-developmental potential of writing software is completely ignored.

I couldn't really fit this in to my blog post "Programming Isn’t Special" https://blog.glyph.im/2026/10/programming-isnt-special.html although it touches on many of the same themes, so that's good background for this.

View

0 Boosts

0 Likes

This federated comment is not yet available here.

Glyph

@glyph@mastodon.social 16 min. ago

At one point I inadvertently saved a business a million dollars in completely pointless costs by noticing something like this, and deleting a feature from their software.

Glyph

@glyph@mastodon.social 13 min. ago

And I'm leaving the "money" story in there because that's how our society values stuff but I have also had deep revelations about managing my own energy and focus by understanding how an asynchronous multiplexer works. I've developed better appreciation for the beauty of morning dew on a flower by understanding the way you have to do lighting in OpenGL.

Glyph

@glyph@mastodon.social 13 min. ago

Understanding something in your brain well enough to encode it as software means understanding it *better*, in the same way that writing isn't just putting words on a page but understanding what you are trying to say, programming can be about learning what you are trying to *do*.

Glyph

@glyph@mastodon.social 26 min. ago

For an example: Story time.

Once upon a time I worked at, let's say, a vertically-integrated business process automation vendor. Our customers were running hundred-year-old businesses, and we had to codify their business logic into software.

I, as a scrappy youth with _zero_ experience in the relevant problem domain, had to learn a bunch of stuff about how their businesses were supposed to work in order to encode those rules.

Glyph

@glyph@mastodon.social 22 min. ago

So, I learned. I learned from interviewing the customers, I learned from the senior members of the development team, I learned from books. I wrote a bunch of code to automate certain processes.

And it turned out… some of those processes didn't make sense. They weren't repeatable; they didn't actually make sense. At one point, I had an argument which was almost *exactly* the famous bodybuilding forum "how many days are in a week" thread. With a customer who owned a $30MM business.

Glyph

@glyph@mastodon.social 20 min. ago

I learned enough that I knew how to fix serious problems with their businesses, as a very young, inexperienced programmer who knew only the very basics of double-entry accounting and just about nothing about running a business. I was not even *good* at this job; I was fired from it.

Because thinking systematically enough to encode things in software is a valuable intellectual practice, and it's one our society does not do enough of because we delegate it to specialists.

Glyph

@glyph@mastodon.social 18 min. ago

I was very, very interested in end-user programming. The thing that drew me to Python was the CP4E ("computer programming for everybody") initiative associated with the language. Python seemed so high-level and so simple that if we could just build a tower of abstractions to separate users from archaic technical concerns like "sockets" and "files" and "IOCTLs", and allow them to write code in terms of the nouns *they* cared about, we would unlock that fabled bicycle for the mind.