The Vintage Software Engineer

I tend to downplay new things in software which people are generally excited about. Perhaps unfairly.

An old Apple II computer sits on a table at a computer museum.
Vintage. Like a good wine. An Apple II computer on display at the Computer History Museum in Mountain View, CA

I’m an old curmudgeon. I like my old stuff just the way it is. VS Code? Ha, not for me. I use Sublime with the VIM keyboard shortcuts turned on. It’s called “vintage mode”, like a fine wine. I’ll take a virtual machine over a Docker container any day. Better yet, stack that hardware right on the rack plz.

I tend to scoff at new things now. Perhaps unfairly, but it seems like new things were a bigger deal when I was learning the ropes back in 1999–2010 ish. Maybe that's an artifact of getting older. I think it probably is. But, still, these things nag at me.

Amazon Web Services; that was a game changer in the early 2000s. Mind blowing for those of us to tried it out back then. I was at a local Linux users group when someone slid over a laptop, we all gathered around, and grown adults screeched with glee as an EC instance spun up with SSH access. Twitter was blowing up. Online forums and IRC were the watering holes. I got to work on Mozilla / Firefox source code, just because someone let me. Node.js and the CommonJS module standard changed the way we do front end web development forever. You could stop into a local users group and get high as a kite on it all. I couldn’t sleep when I got home.

I’m jaded on all the new stuff today. Either I’m getting old and grumpy, or the industry is maturing. Maybe some of both?

We can decentralize the web without blockchain. Hell, the World Wide Web has decentralization built in. And member owned and operated organizations have been around forever. They just weren’t called DAOs. Blockchain has not been able to add anything of value to it.

And crypto is like any other financial instrument that bubbled and burst. Options contracts, ETFs, credit default swaps, railroad bonds.

I have to admit that the NFT craze generated some pretty sic artwork. It just happened to be a modern tulip frenzy. (I would really encourage you to read that Wikipedia article on Tulip Mania — fascinating).

AI is not going to replace programmers, at least not the ones who build valuable stuff. How can a machine create innovative software when it is bound to learning from what was done in the past? It logically cannot. It can create software, just not innovative software. AI is a tool. Nothing more. The autocomplete in my Sublime editor is really dang good at figuring out what module I’m trying to import. And “AI”, if we want to call it that, is going to make those tools even better. I recently heard of an artist who won an art show with a painting generated by AI. It caused quite a scuffle. But, here’s the thing: He had to iterate over and over to finally generate anything that was any good — the idea he had in his head.

Garry Kasparov has made peace with being beaten at chess by AI on the world stage in 1997. Asked if computers will achieve general intelligence:

What’s the difference? We have always invented machines that help us to augment different qualities. And I think AI is just a great tool to achieve something that was impossible 10, 20 years ago.

I don’t believe in artificial general intelligence. Machines will be dominant in the closed systems, whether it’s games, or any other world designed by humans.

That’s it, really. AI is just a tool. It’s one we will get better at using, and perhaps design systems within which it can thrive; kinda like Google on the World Wide Web. AI will continue making it’s way into our lives; just little bits here and there until suddenly it is used all the time in an unremarkable way. And, I’m not sure it will even be AI as we’ve been imagining it. Just regular computer software written by regular programmers with good autocomplete and a set of well designed vintage VIM key bindings.

Everything old is new again.

What I’m Excited About

I’ll need to save the details for another post, but I’m very excited about the potential what is being called the “Fediverse”. It’s mostly just World Wide Web technologies being used in a decentralized way — as designed. That’s an oversimplification, of course, but it’s an example of how the WWW is probably the best software platform we ever came up with.

I’m really excited to explore and write about this more in the near future.


Related Reads

AI Looks like a Bubble by Evan Armstrong

AI and Search Will Make Each Other Better by Shubham Agarwal

Natural Language is a Lazy User Interface by Austin Henley

The AI Crowd is Mad: The misinterpretation of AI by Tim Daubenschütz