mailwal

Should your game discovery plans change due to AI & decompilation?

Simon from GameDiscoverCo

Should your game discovery plans change due to AI & decompilation?

A serious question for devs today. Also: the latest news & the new GameDiscoverCo Show.

[The GameDiscoverCo game discovery newsletter is written by ‘how people find your game’ expert & company founder Simon Carless, and is a regular look at how people discover and buy video games in the 2020s.]

As we start the week, yep, Steam Next Fest has launched, and has 7,100 playable demos (!), significantly up on the 5,000-ish that the ‘trend line’ predicted, and with a new navigation UI that helps filter the ‘random’ part of the Next Fest - Mon to Wed - by categories. We’ll have a full report of results next Tuesday, btw…

Before we start, an ex-Nintendo employee who grabbed about-to-be discarded art from a dumpster is now selling her trove, including the U.S. arcade art for Donkey Kong, some Disney-themed Game & Watch art and even the 1985 Nintendo Xmas card. (And we’d look great in the satin Nintendo jacket from 1990’s Winter CES.)

[NEED THE BEST DATA? You can get a free demo of our GameDiscoverCo Pro company-wide ‘Steam deep dive’ & console data, inc. MCP server, by reaching out today ->110 orgs have it. Or, signing up to GDCo Basic gets extra newsletter and Discord access, plus more.]

Game discovery news: heya, Phantom Blade Zero…

Let’s start the week with a whole bunch of game platform and discovery news, starting with this:

OUR SPONSOR: PWN Games — PC Paid Media Done Right

Advertise your PC game with PWN Games. With our CPA model, you only pay when a player wishlists, installs, engages, or pays.

250+ PC games promoted. 42M+ players acquired. 10 years in the PC UA business.

Trusted by top PC game publishers worldwide. Explore PWN Games today!

Should your game discovery plans change due to AI & decompilation?

Listen, we hate doing non-data or ‘case study’-based newsletters, but we’ve got to break into the programming here. There’s been multiple higher-profile games canceling or delaying demos due to concerns over AI ‘cloning’ of their concepts:

  • ‘Roulette drinking roguelike’ Drunken Goddess Reflux (est. 570k wishlists) canceled its public demo, saying: “We are concerned that, based on the content of the demo, works imitating or replicating this game's presentation and gameplay could be quickly produced using AI or other tools and released ahead of the game's launch.”

  • ‘Tower defensevania’ ReVamp (est. 250k wishlists) canned its Steam playtest & switched to a closed-ish test, noting: “We've been concerned by recent news of games being reverse-engineered, cloned and published before the original is even out. With launch still months away, sharing a build with tens of thousands of players felt like too big a risk for ReVamp.”

All this is going down alongside devs actually decompiling other people’s ‘clone’ games using AI to find stolen code in them, presumably grabbed via AI. (This example, which is Needle In A Haystack accusing A Game About Needle In A Haystack while filing a copyright claim with Steam, is believable but isolated.)

However, everyone - take a breath. We’re basically at the Peter Venkman explaining scenarios to the Mayor of New York in Ghostbusters stage of ‘AI x ripping off your game’ memetics right now. I’m going to suggest that’s a little bit overwrought:

So, some opinions. It’s fine to be worried about people ‘stealing your game idea’, but a) if you have momentum already, or b) if your idea is non-trivially complex and requires balancing and artistry (i.e. if it’s not super-low barrier to entry!), then the chances of someone having a smash hit and locking you out of the picture are low.

The worst-case scenario here is if you have a meme-worthy, but very low-lift idea, such as the mass proliferation of ‘needle in a haystack’ job/incremental games. Just to give you an idea - it appears* (*we think!) that Nas Nakarus from Studio BitDot was the originator of this concept, and released videos around it in late August.

There was no public playable (AI-decompilable) version at the time. It’s the idea, and the fact that people can copy the ‘descriptive’ (searchable!) name of it easily that made clones appear, quick. (We’re pretty darn sure most of these are non-decompiled idea ‘borrows’.)

Just for the record, here are the released ‘needle/haystack’ games - the gameplay is generally first-person, and about gradually automating the ‘needle search’ process:

What’s a little ironic is that a player review of A Game About Needle In A Haystack indignantly says, after the code plagiarism allegation: “i literally did not realize this wasnt the original game DO NOT BUY.” The… original?

Actually, ‘the game made by the guy who did the OGshort-form videos’ is #2 of unreleased titles in this new (?) micro-micro-genre (??), with ~200k wishlists. Everything else is pretty much nowhere, wishlist-wise:

So, just repeating: if you have a short, meme-worthy idea that’s a) great b) trivial to reproduce & has c) a searchable keyword, you will get clones, and within weeks. Most of these clones will go nowhere. The vast majority will not AI decompile you, at least not currently. (And the ‘original’ Needle x Haystack should still have a good launch, when it comes out…)

But when your game is complex, already semi-popular, and is called something else - like ReVamp, or Drunken Goddess Reflux - we don’t think you should delay demos. Why?

  • There’s nothing in either of these games that indicates people could search and find similar games to them & be interested. (They’re not ‘one single idea’-y - and they have a lot of personality, design decisions, and players already bought in. And trailers have already been out there, if people want to hand-copy ‘em.)

  • If a dev ‘fast follows’ your game on Steam, they’re likely not doing it via AI, they’re doing it majority by hand. The resulting game is also unlikely to have scaled wishlists or sales, since you’re the OG. And if it’s not popular… just ignore it. (Hey, it shows you are on the right track, people might find you via them, etc…)

  • If somebody clones you blatantly by decompiling your actual code, AND it’s getting very popular (v. v. unlikely?), go de-compile theirs, document, and report it to Steam as a copyright infringement with proof. This shouldn’t stop you from putting out a demo that could be v. helpful to your discovery scaling.

It’s tricky - let’s not wave away legit concerns. But we need an action plan, rather than abstract dog/cat co-habitation fear. (For example: two potato-themed games fighting about who originated the potato concept.) The above three bullet points are what I think is a sane response to an otherwise all-consuming problem. YMMV.

Of course, if you do have a simple-to-make, casual-facing game with a dynamite idea that you think is going to go viral and cloned the second you reveal it ( ‘punch walls to escape’): you might want to decrease the time from ‘reveal’ until paid launch. (And the ‘moat’ is getting smaller and smaller for these type of games...)

However, that’s a specific - and smaller - subset of the entire Steam game landscape. For most games with significant levels of complexity, we wouldn’t recommend you change your route to market significantly. (Demos & streamer outreach earlier: still good.) Nobody is $ Steam-cloning e.g. Dead As Disco via decompilation soon…

Finally, a couple of recent links worth reading about AI & games: Sulka Haro on where the moats are for all of us in the age of AI: “Brand gets more valuable as supply grows… trust and accountability stay scarce.” And Dylan Collins on why he finds some of this new AI/games energy exciting - tho he does conclude, due to this avalanche of content: “existing games aggregator platforms are almost all undervalued.”

GameDiscoverCo Show: is AI coming for game demos?

Timed neatly to the newsletter, here’s the latest The GameDiscover Show, ft. myself (Simon) andPost Games’ Chris Plante talking trends - sub to the podcast via Apple Podcasts, Spotify & elsewhere (audio), witha video version on YouTube (above).

The latest GDCo Show talks about - yes - the surge in AI decompilation of games and whether releasing demos early is now bad, some learnings from Forbidden Solitaire from a long-term indie game dev, the state of the Meta Quest game scene, and if Epic's vibecoded Epic Games Launcher is going to change the velocity of the PC store.

[We’re GameDiscoverCo, an analysis firm based around one simple issue: how do players find, buy and enjoy your PC or console game? We run the newsletter you’re reading, and provide real-time data services for publishers, devs, and other smart game industry folks.]

© 2026 Game Discovery Now LLC
Bay Area, California, USA