AI just deleted the three moats that protected software for 40 years. Here's what replaces them.
I wrote about the coming wave of disposable software a while back. That piece was a prediction. Now, it's a pattern. Here's what's actually happening - and where the moat moved.
In early 2024, a small team launched an AI productivity app.
Within 72 hours, a dozen clones were live on Product Hunt.
Same prompts. Same interface. Same pitch. Some of them were better.
The original team hadn't shipped anything wrong. They'd just forgotten something important: the product isn't the defensible part anymore.
This isn't an edge case. It's the new default. And if you're building software right now without understanding this shift, you're building on sand.
The Clone Economy Is Already Here
GitHub says 92% of developers now use AI coding tools in some part of their workflow. Cursor hit $100M ARR with fewer than 20 people. Midjourney crossed $200M revenue with 11 employees. Bolt.new lets non-technical founders ship full-stack apps in an afternoon.
The cost of building software is approaching zero.
Which means the cost of cloning software is also approaching zero.
If I can describe your UI to an AI model, scaffold the backend in a few hours, and deploy it on Vercel by evening — your feature moat isn't a moat. It's a 24-hour speed bump.
We are already seeing this play out in real time:
- AI writing tools proliferated so fast that the category commoditised within 18 months of ChatGPT's launch
- Dozens of "Perplexity clones" shipped within weeks of Perplexity getting traction
- Every time a new AI wrapper gets viral attention, three competitors appear before the original team's next sprint planning session
This is the clone economy. And it's only getting faster as the tools get better.
The uncomfortable truth for founders: if your entire value proposition can be described in a prompt, it probably already has been.
Features Are the Worst Moat You Can Build
For 40 years, software companies defended themselves with three layers:
- Engineering difficulty - it took months or years to build something good
- Time to market - being first meant compounding advantages
- Distribution barriers - getting in front of customers was genuinely hard
AI just deleted all three.
When your moat is functionality, your moat lasts exactly as long as it takes someone to inspect your UI, write a prompt, and ship a clone. Sometimes that timeline is less than a day.
Here's the brutal version: if your startup pitch is "we built X but better," and X can be rebuilt by a solo developer with Claude and Cursor over a weekend, then your pitch has a shelf life measured in months, not years.
We are entering an era where apps behave like memes. They appear fast, spike, and disappear. Remember Clubhouse? Lensa? BeReal? Each of those was a genuine product with genuine traction. Each got cloned, absorbed, or outmanoeuvred before they could build a real foundation.
Disposable software isn't a bug; it's the logical output of infinite builders armed with infinite tools.
The question isn't how to stop the clones. It's: **what do you build that they can't copy?**
Where Value Actually Lives Now
There are four things that don't clone easily. All of them require time, intentionality, and a different way of thinking about what you're actually building.
1. Distribution
The biggest moat in 2025 isn't your tech stack - it's your audience.
Look at Notion. The product has been cloned repeatedly. There are free, open-source versions with more features. None of them have meaningfully dented Notion's growth because Notion has something those clones don't: a dense network of users who evangelise it, build templates for it, and teach it to others.
Distribution networks take years to build. Software takes days to copy.
Midjourney built a Discord community before it built a web app. That wasn't an accident. It was a strategic decision to build the distribution layer first. The product sat inside an already-active network from day one.
The founders who understand this are inverting the traditional model. They're not asking "how do we build distribution after we have a product?" They're asking, "Where is the audience, and how do we build the product inside it?"
2. Community
Communities compound value faster than products do.
Memecoins proved this first - the most successful ones were never really about the technology. They were about collective belief, shared identity, and the social momentum that comes from people who've decided something matters.
Telegram games proved it again. Hamster Kombat hit 300 million users not because it was a great game, but because it had the most powerful distribution mechanism on earth: social obligation. Your friends were playing. You joined.
An engaged community provides feedback loops that make the product better, viral distribution that lowers customer acquisition cost to near zero, and cultural momentum that makes switching feel like a social cost, not just a functional one.
A clone can copy the code. It cannot copy belief.
3. Data Gravity
This is the one that most founders underestimate.
AI products get stronger the more data they accumulate. The more users interact with your system, the more they search, click, rate, return, skip - the harder it becomes to replicate what you've built.
Think about what separates Spotify's recommendation engine from a clone that plays the same songs. Spotify has a decade of listening data from 600 million users. That data is the product. The music player is just the interface.
Recommendation systems, behavioral feedback loops, proprietary training datasets, these compound. Code doesn't.
If you're building an AI product and you're not thinking about what proprietary data you're accumulating with every user interaction, you're building the wrong thing.
4. Ecosystems
The strongest platforms aren't apps. They're ecosystems that others build on.
Shopify has 13,000 apps in its marketplace. Roblox has 40 million experiences created by its users. The App Store has 1.8 million apps. These platforms won because they made it worthwhile for other people to invest in them.
Once an ecosystem forms around your product, competitors can't just copy the product; they have to replicate the entire network of developers, creators, and businesses that built on top of it. That's not a feature gap. That's an insurmountable structural advantage.
The lesson: build a platform, not just a product. Give others a reason to build on you.
The Nano Unicorn Pattern
The economics of AI-native startups are rewriting the rules of what a successful company looks like.
Midjourney: $200M+ revenue, 11 employees. Cursor: $100M ARR, under 20 people. Some AI-first companies are generating over $3M in revenue per employee — a multiple that would have seemed impossible five years ago.
This is the nano unicorn pattern. Small teams, AI-multiplied output, massive leverage.
But the same dynamic that makes this possible also accelerates competition. When ideas can be built and distributed instantly, they can be copied just as fast. The lifecycle of a software product is compressing dramatically. Products that once had years to build defensibility now have months.
The founders winning in this environment aren't necessarily the ones who ship fastest. They're the ones who understand that speed to product is just the starting gun, not the finish line.
What the Next Wave Looks Like
The companies that will thrive in the clone economy are building differently from the start.
Instead of: Build product → Find users
They're doing the opposite: Build network → Ship product into it.
This means:
- Communities before products. Build the audience first. The product serves the community, not the other way around.
- Distribution-native design. Every feature should have a built-in reason for users to share it. Virality isn't a growth hack - it's an architectural decision.
- Platform thinking from day one. Even if you're a small team, ask yourself: what would it look like for other people to build on top of what we're creating?
- Proprietary data as strategy. Every user interaction should be accumulating something that makes your product harder to replicate.
Apps will ship faster than ever. Many of them will disappear faster, too. That's not a crisis, it's the new operating environment.
The Takeaway
I've been building across SaaS, Web3, and AI-native platforms for over two decades. The pattern is always the same.
Every time a technology wave lowers the cost of creation, value shifts up the stack.
First, it was websites - building one went from a specialist skill to a commodity, and value moved to the content and brands on top. Then, mobile apps - the barrier to entry collapsed, and value moved to the platforms and ecosystems. Now, AI software.
Each time, the same realization: the product itself was never the moat. The moat was everything around it.
AI didn't kill software value. It moved the value somewhere harder to copy.
Code may become disposable. But systems, real systems, with distribution baked in, communities built around them, and data compounding inside them, those endure.
The builders who understand that distinction are the ones who'll still be standing when the clone wave passes.
Extended version with the full framework: Disposable Apps Are Here
My earlier take on where this was heading: Disposable Software Are Killing the App Store