Steam and Surplus

The creators of Myst – the revolutionary 1993 puzzle adventure game that made the CD-ROM a must-have in the home computer – have been on a bit of a PR campaign about their latest, purportedly abandoned effort (1). They also released a teaser video.

I’m a cynical sort, but the whole thing feels a bit like a planned PR campaign rather than a concession speech, but whatever. I mean, just launch your crowdfunding effort or whatever it is you seem to be pivoting to.

Regardless, I use it as a kicking off point only because the Wired article includes the following non-sequitur that somehow pins the blame on AI-

At the same time, gaming platforms like Steam—which indie and double-A games once depended on for discovery—have been flooded with low-quality dreck. According to the AI Transparency Index, more than 18,000 games on Steam now feature AI usage disclosures. “User trust in platform stores as a curatorial layer has been all but erased,” says North Cook. “When you feed people enough ads and slop, they just check out.” The loss of discoverability has made it even less likely for publishers to take a risk on a mid-budget game because of the additional marketing costs required to stand out from the crowd.

This lazy blame-AI slop1 can be found everywhere now, but this one amazed me in how spurious and laughable it really is.

Let’s revisit the history of Steam2.

Steam – the electronic software distribution platform – was an exclusive platform in its early days. It saw only a few hundred new releases per year, with those titles coming from either Valve itself or a few major publishers that had established a close relationship with the company.

In 2011, for instance, some 281 new games hit the platform.

In August 2012, the platform introduced a new mechanism called Steam Greenlight as a means of broadening its offerings, opening up the service to the many talented indie gamedev creators that didn’t have the resources or connections to go through a traditional publisher. This was an initiative that allowed games to be curated by the community, and those that made it through the gauntlet were added to the platform. Stellar offerings like Stardew Valley and The Stanley Parable got on Steam via Greenlight.

By 2014 the number of new games on the platform had risen by multiples – 1,713 new games that year – largely due to Greenlight. It wasn’t perfect. Not only was there heavy manipulation of the process, many end-users of Steam still had the perception that the platform was responsible for the quality of offerings, as if every selection on the site was hand approved and certified fresh by Valve.

As greenlight let lots of dubious content through, this led to a lot of friction and mixed expectations3.

To solve this Valve decided to fully be hands-off – to leave no doubt that they were merely a market, and not a curator – via Steam Direct. This mechanism allows anyone to publish on Steam if they pass some identity checks and pay the $100 USD fee per listing. The fee is refundable if the game grosses more than $1000 USD (recall that Steam gets their 30% cut), which the vast majority of these submissions never will. Most will wallow in obscurity, while Valve flaunts their fat stack of benjamins 💵4.

Soon many thousands of games were published on the platform each year. Every kid with a copy of Unity could start with an example project, customize it a bit, and publish it on Steam5. Loads of XXX and bizarre niche content started rolling through the new page daily as well.

Advances in game development tools – most now fully free for small devs, where once there was a significant up-front cost – and community forums like Gamedev Stack Exchange, made game development accessible to legions.

And this is entirely a good thing. While there’s plenty of slop – as I’ve covered many times, humans have generated lazy slop long before AI ↗, and we can see intellectual slop in this lazy Wired article citing AI disclosures6 as somehow the cause of discoverability issues – there is also a lot of good content. There are games pandering to tiny niches, which is exactly how a robust market should work.

Someone wanted a Korean War Memorial on Steam, so they added it. Loads of similar small-audience projects can be found on Steam.

All in all, a great improvement. It is whiny, entitled nonsense when people hold this as a negative. People who want to serve up low-value slop get annoyed when users have options and vote with their dollars. Not unlike network television that for years generated endless volumes of slop and still got millions of viewers that were just passing time.

If the marketing plan for a game is “hope people notice us on the new page on Steam” – which seems to be what the source of that quote is implying – it was doomed long before ChatGPT hit the scenes. Utterly doomed. AI is a negligible factor in this story.

But in the end it’s clear that the complaint in the Wired article was never really about “dreck”. Their complaint is, at its core, that with more talented devs and designers and artists in the space, you have to do better to get the word of mouth and hype going to succeed. When you’re asking for millions in funding for a game and your real competition are solo developers creating excellent games with $0 funding, there is a disconnect and a change in reality.

Claude Code, Forking Agent Supreme

Another AI-related story over the past few days includes a very weird tale where someone seemingly cloned an existing project and claimed it as their own. And to be fair, the original project is an MIT project that you can clone, presuming you give the required attribution.

The original project leaned on AI coding assistants. I mean, basically every project on the planet uses coding assistants in their workflow now, so this isn’t a particularly interesting element to the story, though many seem to really hang on it.

It’s just a norm of tooling. There are very few exceptions.

But the weird thing is that in creating an almost identical solution, the author claims it wasn’t him, but Claude did it. This is a growing trend that when things go sideways, AI is to blame7.

If you explicitly tell Claude Code “I want a replica of this MIT-licensed project on Github”, Claude Code may very well say “would you like to clone it and use it as the foundation?” This would be extremely obvious, and there would be zero confusion on what is happening. Loads of projects are forked like this all the time and that is literally how things are supposed to work in open source world. It’s an ideal situation where projects diverge off of each other.

Simply unveiling your own project that is a clone, however, with no attribution, and even causing confusion with the product name, is super scummy behaviour though.

And this incredibly weird situation gets odder still.

Apparently the person submitted an Astrology / Tarot app to the App Store, it was denied as an oversaturated field that is no longer taking submissions, so he then cloned this project and went crying to some influential people like John Gruber whining about being persecuted, pointing to their astrology astronomy related application as proof that they are the victim. So the cloning of the project seemed to just be cover for some other bit of nonsense, but it blew up a bit more than they expected.

All in all, what an incredibly weird story that yielded a lot of burned bridges and professional self-sabotage.

But again, it’s amazing how “AI did it” got the blame, and how many people simply accepted this, including loads of commentators on HN who praised how classy it was to “own up” to the mistakes.

AI has become a go to for all sort of malfeasance now, because it works.

And none of this is to say that AI isn’t sometimes, if not often, a problem. The Epstein administration of pedos and grifters post laughably stupid AI slop on the daily. Loads of menus, banners, posters, and so on is AI slop. But slop and malfeasance didn’t originate with AI, and people probably need to be a lot less accepting of dubious “AI did it” short-circuit slop so many resort to, however much anger they feel towards AI.

AI is the new Ambien lazy out for bad people doing bad things.

Footnotes

  1. My wording is intentional and my punctuation correct. I am not talking about blaming AI slop, I’m saying that a lot of worthless, lazy writers are authoring organic slop where they find a way to blame AI. Their position relies upon being incredibly ignorant and ahistorical. ↩

  2. I first installed Steam over the Christmas holidays in 2004. I wanted something new to play and the stores were closed, but this relatively new Steam thing offered up the ability to purchase and install online. So I bought and installed Half-Life 2 and was hooked. Since I’ve amassed a library of hundreds of Steam games, many of which I’ve never played. ↩

  3. Not unlike the Apple AppStore. Apple likes to talk up their rigorous demands and careful processes, but in reality the store has always been full of enormous volumes of garbage. ↩

  4. 0% of the text content of this blog has been authored by AI – it’s 100% hand authored, letter by painstaking letter – yet I insert Unicode for colour and use and misuse em-dashes. How can this be? ↩

  5. I was going to link some samples here, but that felt mean-spirited. But rest-assured, there are many tens of thousands of examples, and almost none of them relied upon AI at all. ↩

  6. And honestly, the whole AI disclosure thing is such performative nonsense. Projects released in 2026 that don’t disclose AI tooling simply haven’t admitted to it. ↩

  7. Always weird that when people are caught doing something, their lesson is to tell everyone else what rules they should be following. ↩

See Also