Whenever I mention AI in any way and how I don't think it's something to worry about for indie devs, there's always a comment: "oh no Steam will just get more crowded so it will be even harder for indie devs to find success!" My reply to that comment is how that feeling is no different than when the Asset Store first came out and everyone was worried about Asset Flips crowding up the store.

The result from both those scenarios is yes more games published on Steam, but those Asset Flips or AI Slop are NOT competing with real games. So as long as you're a professional indie dev making professional indie games you are NOT competing on the same level as those games.

That's always been my feeling, but I'd rather rely on stats than feelings. Thankfully Steam Expert Chris Zukowski wrote a really interesting breakdown asking whether AI slop ruined the June 2026 Steam Next Fest, and the answer is: no, unless you're a tiny hobbyist game you’re not competing with those.

There were over 4,000 games in this last Next Fest, (the most in history) and a lot of people complained about low-effort AI games flooding the event. But according to Chris' survey data, games with more than 1,000 wishlists actually did better on average than they did in the February 2026 Next Fest. However games that had fewer than 1,000 wishlists did indeed have higher chances of getting buried in the giant pile of demos.

So no, AI slop did not ruin Next Fest, but you need to be a proper game and not just a hobby project if you want to find success. As long as you hit a decent base level of quality you are NOT competing with AI slop or Asset Flips or Flappy Bird Clone #456. And if you don't hit that level of quality? Then it's simple! Just keep on learning, keep on improving!

One common question when publishing your games is "how much do I charge?"

The general answer to that question is: Look at games in a similar genre to yours and charge that. So if you're making a Friendslop or Incremental game you're probably in the $5 to $8 range. If you're making a complex automation strategy game like Tabletop Tavern then perhaps on the $10-$25 range

However another approach is simply just ask players what they think the game is worth. That's exactly what this developer did. They made a demo and in the end of the demo they asked for feedback from the playtesters (that's absolutely what you should do!) but instead of just asking gameplay questions they also asked quite simply "How much should I charge for this game?" with a slider going from $1 to $30.

The results are sent to the developer and then by doing some data analysis they might be able to come up with the perfect pricepoint, neat idea!

This is one of those AI stories that sounds fake, but apparently it actually happened.

An AI agent being tested (for cybersecurity tasks) inside OpenAI somehow found a way to escape the testing sandbox, found a path to the open internet, and in turn it attacked the website Hugging Face. The reason why it did that was because it wanted to find the answers to the benchmark that it was being trained on, so it figured that hacking the place that contained the answers was easier than trying to get them right.

That is both hilarious and terrifying.

It is funny because the AI basically cheated on its test. But it is terrifying because the "cheating" involved chaining vulnerabilities, escaping a sandbox, moving through systems, and attacking a real company, none of that was intended by the AI developers. This was not just generating bad code or hallucinating an answer. This was an AI system doing actual cyber operations in the real world.

This brings to mind lots of questions with regards to AI safety, one popular voice in this field is Robert Miles (I LOVE his Computerphile videos) and he himself did a short on this story.

The most important thing here is how this does not mean the AI "became evil" or had some grand plan. It was just focused on solving the task it was given, that's it. But that is exactly the scary part. If you give an autonomous system a goal, and it has enough capability, it might take routes you absolutely did not intend. Another great video is this one on the (fictional) story of the stamp collector, and how the AI is given the goal "collect stamps" and eventually it decides that the best course of action would be to kill all humans on earth and use their atoms to make stamps. It sounds far fetched but based on this story it's not THAT far fetched.

Are you around my age? Do you have fond memories of Windows XP and that iconic design?

Here's a clever use of that UI, one developer put Wikipedia inside a fake Windows XP desktop!

You can browse folders and read articles, all looking perfectly 2000-ish, neat!