Software Is Content Now

Software has gotten cheap enough to build that a lot of it now behaves like content. It can be made for a moment, earn what it can, and then get absorbed into a model or someone else's feature. That's breaking business models. It also changes what product-market fit means when growth comes this fast, and which moats still hold, at least for now. Those questions matter to anyone deciding where to put their money or their time.

This morning I almost downloaded a phone game. Duolingo's free tier keeps showing me a playable ad for a color-sorting puzzle I enjoy. (I'm around 1,200 days into a streak, and at this point Duolingo is a growth hack cosplaying as learning.) The real game is freemium, so it would likely push me to upgrade. It might carry tracking. Its business model might be to exploit my privacy and data. A lot of potential downsides for a pretty simple game.

So on the way to drop off my son, I asked Claude what kind of game it was and had it build me one as an artifact, playable on my phone. By the time I was on the bus home, I had Crumb Hunt, a puzzle where squads of ants eat a pixel-art picture block by block. It was easier to build it from scratch than to contend with those what-ifs.

All of those ad-driven games are going to take a hit. There will also be far more of them, because anyone can dream up a game and publish it. App Store submissions jumped 84% in a single quarter earlier this year, the biggest jump in a decade, according to The Information. Developers complain that reviews which used to clear in a day now take weeks. Apple says it still reviews 90% of submissions within 48 hours. Either way, that whole stream of income is in for some trouble.

Everything can be rebuilt

Games are one corner of it. An open-source tool called REA, short for Reverse Engineer Anything, lets a coding agent take apart an app it has no source code for. Point your agent at an app and REA will explain how a feature works and, if you'd like, build a version for your project.

Kotaku reported hundreds of browser ports of games like Halo and GTA: Vice City appearing within a couple of weeks, decompiled with Claude's help, alongside mashups like Super Mario 64 dropped into Elden Ring. The team behind ArtCraft has published open-source alternatives to Photoshop, Premiere, Lightroom, and most of the rest of Adobe's lineup, plus Word, Excel, PowerPoint, and Figma. ArtCraft says every app is written from scratch, and describes several as clean-room reimplementations, built from how the original behaves rather than from its code. The game ports show complete disregard for copyright or business model. Either way, people are tearing down the old walls of software to make whatever they want.

It's crazy to watch, and it's scary. We don't know which of these companies make sense or which business models will hold up. We don't know which jobs will still be relevant in nine months, or what education to invest in for our kids. There is so much change and uncertainty.

Even the analysis meant to answer these questions is getting cheap. Lenny's Newsletter recently ran a guest post by Dan Hockenmaier on how to tell whether a startup has a moat, written for people deciding whether to join one. He argues that distribution, speed, and brand don't count. The moats that matter are compounding advantages, and he names four. Switching costs, network effects, scale economies, and data effects. The post walks you through finding which one a company has, gauging how strong it could get, and measuring whether it's working. I don't think those are the moats forever, but they're the best map we have right now. In May I wrote that your app is not the moat, and that the durable part is what a system learns from its data. His data effects are the same idea, with a method for checking.

Then he shipped the framework as a skill. You download it, add it to Claude once, and start a chat with something like "Does [company] have a moat?" It asks one or two questions at a time, reads any deck or investor update you hand it, and gives back a verdict, a scorecard, and questions to ask the team. That analysis would have been the pinnacle assignment of business school, or a McKinsey project. Now it's a throwaway skill. The creative destruction is insane.

Software is content now. I can spin something up for just social media likes, or for pizza money for a month (maybe lots of pizza), until it gets baked into a model or another company makes it a feature.

A moment or a moat

The next day, Nihal Mehta, a friend in venture capital, posted a note from a founder. "Took 170 days to have our first $1M GMV day. It took 8.5 hours to have our second." His caption was "what PMF looks like."

I replied that the speed at which those results can be reached now is insane, and that week to week the whole world can change, for better and for worse. I asked how his best founders manage to stay in that pocket.

What I kept wondering was what product-market fit even means when you can grow this fast. You can recognize an opportunity, be making bags of money in three months, and three months later no one can remember your name. The speed that gets a company to a $1M day is available to its competitors too. Nobody should pass on a moment like that. What I don't know is how you'd find a sustainable advantage once you're in it. Maybe in a lot of places there isn't one. You run as fast and as far as you can and hope you keep pace or outrun everyone else. Or you take as much advantage of the moment as you can, and then the moment is gone.

Hockenmaier's post holds the sharpest objection to that. "A company that has to sprint forever is not a good business." The ability to sprint isn't a moat. But it could be a lucrative effort for a time. Most people don't do Porter's five forces before they launch something. They work from where they are and make the best of it. A durable moat with a reinforcing feedback loop is better, and if you're deciding whether to join a startup, the question his skill is built to answer, it's what you should look for. When you can't get one, you make the best of what you have.

A lot of what's getting built now is closer to content, made for a moment and worth something while the moment lasts. The mistake may be assuming a $1M day is proof of a moat rather than just a moment in time.

The same tools change what happens after the moment, too. We can now run zombie businesses. Automate the email, automate all the things, and scale the business down until one person can run it. If the economics work. There's also the risk that the automation lets you hold on to a thing for longer than you should.

Back to first principles

We have to reassess so many things. For a long time, we've taken shortcuts based on the status quo about what's valuable and how we work. That a polished app takes a team. That a strategy framework is worth paying for. That fast traction means you've found fit. Right now we need to go back to first principles and consider what we're trying to accomplish, and the best way to accomplish it.

That goes for the long-term bets we invest in and count on, and for the short-term opportunities we should move on now. Both mean experimenting more and re-interrogating our priors. I don't think there are answers we can carve in stone, because the world is still changing so fast. We're not done, and things are still very fluid.