Back to Blog

Why Game Discovery Is the Real AI Bet

by MrPhil
games-industryaidiscoverygame-design

A few days ago I wrote about the last techno-scribes — about how AI is doing for programming what the printing press did for literacy. I called it "the flood." When anyone can make a game, everyone will.

I believe that. I also believe almost nobody is thinking about what comes after.

Here's the thing about floods: the water isn't the problem. It's the fact that you can't find anything once the water rises.

The Creation Bottleneck Is Dissolving

Right now, making a game requires a rare combination of skills — programming, art, design, audio, project management, and the stubborn refusal to quit when everything breaks at 2am. That's a high barrier. It's also a natural filter. The difficulty of making games has, for decades, served as a crude quality gate. Not a perfect one — plenty of great designers never shipped because they couldn't code, and plenty of mediocre games shipped because the team had technical chops. But the barrier kept the total supply of games within a range that storefronts could sort of manage.

AI is dissolving that barrier. Not tomorrow — now. People who couldn't make a game last year are making games this year. People who could make one game a year will make five. The tools are getting better every month, and the people using them are getting better at using them every week.

This is wonderful. I mean that sincerely. The games we're going to get from designers who were previously locked out by technical barriers will be extraordinary. Some of the best games of the next decade will come from people who have never written a line of code.

But the math is unforgiving. Steam already has nearly 20,000 games released per year — a new record. What happens when that number is 100,000? 200,000? A million?

Discovery Is Already Broken

If you're an indie developer, you already know this. Getting your game in front of the right audience is the hardest part of shipping. Harder than development. Harder than marketing. The Steam algorithm is a black box. The press covers a tiny fraction of releases. Social media is pay-to-play. Wishlists are the new currency, and earning them is a full-time job that has nothing to do with making a good game.

The tools we have for finding games are stuck in the past. Genre tags. User reviews. Curator lists. "More like this" algorithms that mostly recommend whatever's already popular. These systems were designed for a world with thousands of games, not hundreds of thousands. They're already groaning under the weight. When the AI-driven supply explosion hits, they'll collapse.

And here's the cruel irony: AI is going to make it trivially easy to create games, but it's not going to make it any easier to find the one game in ten thousand that's perfect for you.

What If Games Had DNA?

This is the problem I've been chewing on. Not how to make games — we've got that covered — but how to help players find the ones they'll actually love.

The core insight is simple: genre is a terrible way to describe a game. "Strategy" tells you almost nothing. Europa Universalis IV and Into the Breach are both strategy games. They have almost nothing in common. One is a 1,000-hour grand campaign across centuries of history. The other is a tight puzzle you can play in thirty minutes. Calling them both "strategy" is like calling The Shawshank Redemption and Airplane! both "movies." Technically true, completely useless for recommendations.

What if instead of genres, we described games by their essence — the combination of mechanics, pacing, complexity, themes, aesthetic, and feel that makes a game what it actually is? Not a single label, but a multidimensional fingerprint. A DNA.

That's what I'm building at GameLegend.com.

Gameplay DNA

GameLegend maps every game across nine dimensions:

  • Core Mechanics — What you actually do. Combat, building, exploration, puzzle-solving, automation design. Twenty-one categories, because games are complicated.
  • Feel/Pacing — The moment-to-moment experience. Is it meditative? Frantic? Methodical? Tense?
  • Progression — How you advance. Roguelike loops, open world exploration, linear story, skill mastery.
  • Social Mode — How you play with others. Solo contemplative, co-op chaos, competitive ranked, async community.
  • Aesthetic — Visual identity. Pixel art, photorealistic, hand-drawn, minimalist.
  • Themes — What the game is about. Cosmic horror, political intrigue, wholesome friendship, survival against odds.
  • Complexity — From pick-up-and-play to spreadsheet territory.
  • Session Length — Five-minute runs to multi-hour deep dives.
  • Strategic Scope — For strategy games specifically: map scale, economic depth, warfare emphasis, emergent narrative.

When you describe a game this way, you get something powerful: a real basis for comparison. Not "these are both strategy games" but "these both have methodical pacing, high complexity, deep economic systems, and multi-hour sessions — but one emphasizes political intrigue while the other focuses on warfare." That's a recommendation you can actually use.

It's live now with a handful of strategy games — Europa Universalis IV, Crusader Kings III, Hearts of Iron IV, Factorio, Total War: WARHAMMER III, Sekiro, Hollow Knight. Enough to see the concept work. Enough to see how games that share genre labels can have wildly different DNA, and games from different genres can share surprising similarities.

The Map Needs a Crowd

Here's where it gets interesting. Right now, the DNA profiles on GameLegend are my own assessments. That's fine for a proof of concept, but it's also the weakest part of the system. My perspective is one perspective. I might think Factorio is "meditative." You might think it's "methodical." Someone who speedruns it might call it "frantic." They're all right.

The real power of game DNA emerges when it's shaped by the crowd. When thousands of players weigh in on a game's pacing, or complexity, or themes — not with a single 1-to-10 score, but across every dimension — you get a map that reflects how games actually feel to the people who play them. Not a critic's assessment. Not an algorithm's guess. A collective understanding of what makes each game tick.

That's the vision. Let players contribute to a game's DNA. Weight their input. Build a map of the entire gaming landscape based on the wisdom of the people who actually play these games. A living, breathing taxonomy that evolves as games update and communities grow.

Why This Matters Now

The AI boom in games is going to be incredible for creators. The economic signals are already clear — smaller teams, bigger output. But it's going to be brutal for discoverability. When the flood hits, the games that survive won't necessarily be the best ones — they'll be the ones with the biggest marketing budgets, the most viral TikTok moments, the luckiest algorithmic breaks.

Unless we build better tools for finding them.

That's the bet I'm making. Not that AI will create amazing games — it will, and I'm excited about that. But that the real unsolved problem, the one that will determine whether this creative explosion actually reaches players, is discovery. And the solution isn't a better algorithm guessing what you want. It's a deeper understanding of what games are.

GameLegend is just the beginning. A handful of strategy games and a nine-dimensional model. But every map starts with the first few landmarks. If you want to see it — or if you want to tell me I've got a game's DNA wrong — come take a look.