Last month, I joined a panel hosted by Liquid & Grit. The topic: What happened in 2025 and what’s coming next?
One critical takeaway we all agreed on: change isn’t just coming, it’s here right now.
Here are the things I wanted to highlight — and the one nobody wanted to say out loud:
The Attention Economy: The noise problem is about to get much worse
Going Small: Small teams aren’t a trend. They’re the new default.
Progression Compression: Players are more impatient than ever — and that breaks your progression.
AI or Bust: If you’re not using AI, you’re ngmi.
👉 Watch the full panel on YouTube below!
🎧 Listen on Spotify, Apple Podcasts, or Anchor
Speakers:
Oscar Clark, CEO at Fundamentally Games
Cody Hard, Design Director at EA
Joseph Kim, CEO at Lila Games
Brett Nowak, CEO at Liquid & Grit
Ariel Navon, Head of Competitive Intelligence at Playtika
Below, I’ll break down what stuck with me — and what I was thinking but didn’t say on the panel.
The Noise Problem Is About to Go Parabolic
Bruh, check out this chart of App Store submissions from a16z below.
Post-AI, the curve is bending massively upward! After relatively flat growth over three years, new iOS app releases surged 60% year-over-year in December. They attribute it to “agentic coding” — what people are calling “vibe coding.”
By the end of this year, how much higher do you think this goes?
Here in Silicon Valley, nobody codes anymore. That’s not hyperbole — it’s just how people build software now. Web apps, SaaS products, internal tools — all AI-assisted, built in days instead of months. Games are harder, and game dev will likely be the last software category to go fully AI-native. But we’re already seeing it accelerate.
If you’re interested in why, I can post on this in a future episode.
Think I’m being hyperbolic? Here are just 2 projects ( I actually have others I am not publicly sharing), I personally coded from zero as an old guy who hasn’t coded in over 20 years, and over just a few days:
ClawTalk: A remote TUI into OpenClaw that enables context-specific conversations, model usage data, multiple agents in context-specific discussions, live voice chat, and more. It’s open source; check it out, especially if you use OpenClaw.
Synthetical Research: A service we use at Lila to synthetically test art concepts. I created an easy way for anyone to run this research paper: https://arxiv.org/abs/2510.08338
The cost of software and content is trending to zero. Not reaching zero — there are still operating costs, infrastructure, human judgment — but the marginal cost of producing *something* is collapsing by orders of magnitude.
And it’s not just a supply-side problem. On the demand side, players are more impatient than ever. We saw this clearly with ARC Raiders — the progression loop has been compressed dramatically. You don’t grind 20 games to unlock the next map anymore. You play 5, and you’re in. You craft something after your first battle. The “do something, get something” loop has been squeezed to near-zero delay. That’s the attention economy at work — if your game doesn’t deliver value immediately, players bounce to one of the billion other options that will.
So studios are caught in a double squeeze: more competition flooding in from the supply side, and pickier, less patient players on the demand side.
What happens when anyone can build an app, and nobody has patience? You get a billion apps and seconds to prove yours matters.
Oscar Clark put it well on the panel. He compared the moment to the rise of folk bands — people seeking authenticity and trust in response to the synthetic flood. When everything looks the same, the thing that’s clearly made by humans who give a damn stands out.
But I’d push that further. Authenticity alone isn’t enough. You need to signal three things almost immediately: 1. quality, 2. a strange attractor (that’s a Hollywood concept — the thing that makes your project undeniably unique), or 3. mystery (think Simon Stalenhag). If you don’t hit at least one of those in the first 15 seconds (for mobile), you’ve already lost the player’s attention. And that was true before the AI flood. Now it’s existential.
—————————————————————————
The New Winners: Small Teams That Build Communities
I published a post last September (see Alpha Insight #2: The Great Western Game Development Divergence) that looked at what the most successful Western games had in common. Not the $200 million AAA titles — the breakouts. And a clear pattern emerged, especially on Steam and PC.
The winners were small teams. Teams of 1 to 20 people. Lethal Company — one developer, $70M+ net revenue, zero marketing spend. Schedule 1 — one developer, 8M+ units sold. Project Zomboid — about 10 people, hundreds of millions in lifetime revenue over 13 years of early access.
Two things they all had in common:
They engaged their community obsessively. Not as a marketing tactic — as a development methodology. Building in public, iterating based on player feedback, making players feel like co-creators. The community wasn’t a distribution channel. It was the product.
They moved fast. No 18-month development cycles. No waiting for approval from seven layers of management. Ship, learn, iterate, ship again.
And here’s the thing — it’s not just small teams winning. Delta Force generated 60-70% more revenue than PUBG with a fraction of the DAU (over 22X less) — through layered LiveOps that stack events, progression, and monetization in ways PUBG never attempted. The winners aren’t small or big. They’re fast and smart.
Here’s what this means for 2026: if you’re a new studio, your path to breaking through the noise isn’t a bigger marketing budget. It’s a deeper community and faster iteration cycles. Build something people care about, with people who care about it, and let the community do what no amount of paid UA can.
—————————————————————————
The Thing Nobody Wanted to Hear
I said on the panel that I’m an AI maximalist. I compared the current stigma around AI in games to the early days of online dating — everyone thought it was gross, nobody wanted to admit they did it, and now it’s how most couples meet.
AI in gaming is on the same trajectory. Players care about it today. They care less about this issue for mobile. And in one to five years, it won’t matter at all. You can’t fight the wave. It’s like trying to fight the Internet.
There’s going to be a split: studios that lean into AI to create compelling experiences — faster, cheaper, better, and some things that aren’t even possible without AI — and studios that don’t.
I believe the second group simply doesn’t make it.
—————————————————————————
Next week, I’m going deep on the GDC State of the Industry report — because the data in it confirms a thesis I’ve been developing about what the gaming industry looks like when its moat disappears. The industry isn’t in a downturn. It’s losing its moat. Stay tuned.
