Software never had a soul(jmduke.com) |
Software never had a soul(jmduke.com) |
I never give a name to a car until I’ve done something substantial to it and it rewarded me with a decent trip in return. My wife’s Subaru will likely never have a name because I haven’t cut myself fixing it or replaced anything major.
The old SRT8 300C? It has a name.
Even the original author Ryo's rambling post on X [1] conflates the personal web (personal blogs), window manager UX, AI-generated code, commercial / enterprise software, testing, and boutique or hobby apps, concluding with a plea for a "path forward" that is already available today, while painting everything with the same brush.
Even back in the 1990's real software (e.g., I used VxWorks every day) was built for consistency, functionality, and repeatability, with strong QA, and the thing was expected to work.
Yes, this is the same that happened to other forms of expression decades or even centuries ago and is extremely well understood. Why blame the canvas?
I think what people actually want is attention and praise for their individual efforts, and that's a very different problem altogether.
At the root of all this is ego and mortality. People much more clearly see their own insecurities now. The inner voice to confront them only gets louder the longer this cultural constipation drags on. The anxiety to feel validated holds it in.
So many great potential artists are afraid to bare their souls and be vulnerable. They're afraid their song will just sound the same as everyone else's and nobody will care. Worse still if everyone hates it or doesn't age well.
Yes, that's what makes art so hard. Of course it is! You do it anyway and stop taking yourself so seriously. Not everyone has to be or should even want to be a rock star. You join the choir at least. We all lose when nobody sings at all anymore.
This software has a soul - at least by my definition.
It's not software. It's the fact that distribution is owned and taxed by outsized players that live as gods and control the experience for the rest of us.
You might not care about Google and Meta, but your customers and parents will be bound up by them. You'll have to pay a tax to reach them. You'll have to jump through their arbitrary rules and give up more than you wanted.
They're the ones deciding to let privacy encroaching governments continue to erode our rights. It better facilitates their profit making opportunities and helps maintain their high walled moats.
Your little blog might have meant something in 2004, but today it's nothing against the titans.
The internet of 1990-2008 was not "indie". It was "free".
The internet of 2000 was the undiscovered country. The internet of 2026 is 1985 surveillance coupled with Brave New World meets Thunderdome algorithms.
The other reason the author misses - the internet was a much smaller place. A personal website or forum would be seen by a large percentage of the internet. The "indie" web was the web. (Drop the "indie".) Now you have to go live on a platform and be an ephemeral engagement sink.
If you don't obey all the arbitrary rules (no external hyperlinks, no videos under thirty seconds, no website references in your images, no green texts, no edits, no posts after 11AM), you won't be selected by the algo lottery for content farming to the horde. Nevermind that they'll ban you if you're a problem to any important powers.
You're just a consumer now. And if you're a creative person, your wares are content to be algorithmically milked.
Isn't that just a little bit dystopian? Doesn't it fall just a tad shy of the dreams we had twenty years ago?
What does that even mean? Seems like empty words to me from people too accustomed to tv commercials.
There is the trivial meaning, where the subject of the sentence is apparently whiling away time, achieving nothing of note except pretending perhaps to be in an imaginary land.
Then there is another sense, one that includes the thrill of experimentation, the disappointment of failure, the doggedness of persistence, and the satisfaction of victory and success when the puzzle is complete, understood, and the whole thing is working as desired or expected. This is why we call programming "fun" and if you are having fun doing it for yourself, you should perhaps be very careful where you end up doing it for work, if you do.
You could do that on computers of the 1990s, and still have the feeling of a broad system, but one which was not unfathomably deep. That's because those systems could be completely understood by one human brain, and being able, striving to be able to do that, was indeed enormously engaging, but people who waxed lyrical about such things were often seen as weirdos, and humans don't like that, generally, so instead they reach for a word that has universal meaning: "fun". Of course, words that have universal meaning, and for which everyone has their own interpretation (though they may not be aware of it), in this manner ironically tend to lose all shared meaning in the strictest sense.
What's sometimes overlooked in the Smalltalk story is that Alan Kay was leading the "Learning Research Group", which is why he refers to educational theorists like Jean Piaget. In some of Alan's talks he goes into some detail showing how children can learn about calculus by watching and visualizing the acceleration of a ball as it falls and bounces. This sort of thing is a serious kind of fun because it actually has a positive benefit, much like sport does for many people.
On the other hand, the use of the word in "making breakfast fun for children" in the advertising sense is a disgusting perversion, and is no way reasonable comparable to the idea of "computers being fun in the past".
Now, if you'll excuse me, I'm going to have my breakfast consisting of dippy eggs and soldiers, and marvel at the viscosity.
Some of this might just be demographic shifts, i.e. "normal" people using the internet more. The people who are on the internet now would likely never have been interested in reading some indie blog, they just weren't online in 2000.
I could be wrong, but I suspect the absolute number of people who read this blog today is larger than it would have been in 2000, just in a smaller corner of the internet.
Sure we used AIM and MSN Messenger but we also used IRC, visited forums and looked at newsgroups.
So many people these days don't even own laptops. Their entire digital footprint comes through apps on restrictive mobile platforms.
Also, you dont need to consume the algo content. You can choose to disengage.
That's a choice.
You don't actually have to be an innocent passive bystander; that's a choice. A lot of these nostalgic threads read like old men shouting at clouds at this point "Get off my lawn!". I'm part of the generation that built the internet (though I didn't build any of it). It wasn't taken from us; we just got old. That's actually on us. Nobody ever stopped anyone from blogging, publishing RSS feeds, or whatever it is people did when they and the internet were young. I was there as well. The only thing that changed is that we sunk in a collective mid life crisis when Google Reader pulled the plug. Alternatives never filled the vacuum. Social networks took over and then imploded. And boohoo the evil corporations took our toys. Give me a break!
If you choose to get your news from algorithmic ad factories, that's a choice. Sure, it's convenient. I do this as well. But actually we're empowered to do whatever we want. More so than ever. There are plenty of tools and content creators out there. And with AI it's trivial to build your own tools. Try it, it's fun. The only limitation is your imagination and apathy.
I've been working on a Google reader style reading experience but running locally with codex. It uses locally running LLMs, embeddings, and entity extraction to stay on top of the fire hose of news. Embeddings enable semantic search, clustering, related articles. Entities identify common topics. LLMs allow me to turn this into a personal news agent and editor. Grand vision and I'm a bit geeking out. But completely doable with all the modern toys we have available. Use openclaw/codex/claude cowork/whatever agentic thing you want, make it parse opml and sync and store feeds. Add some skills and you have yourself an AI news agent. Or skip all of that AI nonsense and do it by hand. RSS readers never were rocket science. This stuff no longer is hard.