People Are Vibe-Coding Clones of Every Paid App. What Actually Survives?

Earlier this month, a developer in Atlanta got fed up with Adobe’s cancellation fees. So he used Claude to rebuild Adobe’s Creative Suite from scratch in Rust: seven apps, from Photoshop to After Effects. He put them on GitHub for free and declared that “software is over.”

He is not alone. Clones of paid apps are appearing everywhere. There are creative tools, CRMs that claim to “drop Salesforce”, and all sorts of commercial software being copied by individuals who point AI agents at problems that once needed a whole company.

Let’s be honest: most of these clones are rough. They are early, buggy and missing the hard last 10% of polish that makes software usable. The Adobe-clone creator says so himself: “not anywhere close to ready.” If you are waiting for a free Photoshop that fully replaces Photoshop, you will be waiting a while.

But judging each clone on whether it is good yet misses the point. Something bigger has happened. The wall that protected a handful of giants for twenty years has just become much easier to climb. So what actually survives?

The moat was never the code

The giants were never protected by genius. They were protected by cost.

Building something like Photoshop or Salesforce took armies of engineers, years of work and tens of millions of dollars. Plenty of people knew what a photo editor should do. Almost nobody could afford to build one.

That is what let the giants behave as they did. They could charge subscription rent for ever, raise prices at will, lock you into their file formats and make cancelling painful. Where else were you going to go? GIMP existed, but the fact that “the free alternative” was a joke for twenty years shows how safe the moat was.

AI did not make the clones good. It made them cheap and fast, and that is the variable that mattered. One person with an AI agent can now attempt in weeks what once took a funded team years. A clone does not need to be finished to matter. The wall only needs to be climbable, and now it is.

What survives (and it is not the clones)

Most of the clones will die. They will stall in alpha, or the maintainer will lose interest, or they will never close the gap between an impressive demo and something you trust with your work. AI makes it easier to start a project, not to finish and maintain one.

What survives is this: the giants can no longer assume they are the only option. For twenty years, “build your own” was not a sentence anyone finished. Now it is, credibly and cheaply.

A monopoly’s power comes from having no alternatives. Once a credible alternative is cheap to build, the monopoly has to behave as if it faces competition, whether or not any one rival wins. The possibility is what counts, not the polish. The cost wall has fallen, and it is not going back up.

Why this is genuinely good

This is not just about enjoying Big Tech’s discomfort. There are real reasons it is healthy.

Monopoly pricing weakens. Even if few people actually build an alternative, the giants know they could. That caps what they can charge and how badly they can treat customers. The escape hatch is real now.

Leverage returns to users. For two decades the choice was: accept the subscription, accept the price rise, or leave for nothing. Now there is a credible “or else”.

Innovation spreads out. When serious software needed armies, it was built to serve shareholders and quarterly targets. When it needs one motivated person and an AI, thousands can try. Some will build things the giants never would, because they are solving their own problems rather than tuning a sales funnel.

The most exposed companies earned it. The developer built his clones because of cancellation fees and resentment. The firms most at risk are those that leaned hardest on their moat, with aggressive pricing and hostile lock-in. Companies that treated people well have the least to fear. That is not revenge. It is the market doing its job.

The irony

The giants pushed AI harder than anyone. They built it, put it in everything and sold it as the future. Many also cited AI efficiency while laying off thousands of their own engineers. The message to workers was clear: AI makes your skills cheap, so you can be replaced.

Now the same AI is making the giants’ moats cheap. The tool used to cut labour costs is wearing away the wall that protected their pricing. You cannot spend years telling the world that AI makes expertise cheap and replaceable, then act shocked when it does the same to your product. The symmetry is simply fair.

AI as the great equaliser

AI collapsed the one thing holding the monopoly structure up: the cost of building. For the first time, building ambitious software does not depend on a company’s balance sheet.

There is a double edge. The same AI helps Adobe defend itself, and the giants have more computing power, data, reach and lawyers. But the benefit is lopsided. Giving a new weapon to someone who already has an army changes little. Giving it to someone with a stick changes everything.

The honest counterweight

If I only gave you the optimistic half, you would be right not to trust me.

The legal fights are coming, and they will be ugly. “Clean-room” reimplementation, which means copying what software does without copying its code, is a real and defensible idea. But trade dress, trademarks and the line between “inspired by” and “copy of” are not settled, and the giants can afford to make that line expensive to approach. Some clones will be sued out of existence whatever their technical merit.

A free clone is not automatically a win either. Software nobody maintains is a security risk. Without support, updates and a community, a clone becomes a trap the day its only maintainer gets bored. The open-source graveyard is full of “free alternatives” that were better for about eight months.

None of this undoes the core shift. But anyone telling you “software is free now and the giants are finished” is just being overconfident from the other side. The wall fell. The giants will adapt, fight in court and survive. What is gone is their uncontested power.

So what survives?

Not most of the clones. They will do what early, unmaintained, one-person software usually does.

What survives is the end of an era: one in which a few companies could build an unbreakable wall out of engineering cost, charge rent behind it for ever, and treat users and workers as disposable because there was nowhere else to go. AI has knocked a hole in that wall. The giants are not finished, but they will never again enjoy the same uncontested dominance. A threat that is cheap to mount never fully goes away.

This is not the end of software. It is a change in who gets to make it, away from the few with war chests and towards the many with an idea and an AI agent. Most will fail, and that is fine. The point is that, for the first time in a long while, the giants can lose. A world where they can lose is healthier than the one we had.

Two honest questions, and I want real answers, not comfortable ones. Is AI collapsing the cost of cloning software actually good, because it ends monopoly rent and lock-in? Or does it just leave a graveyard of abandoned clones and a decade of lawsuits? And the harder one: after years of “AI will replace you” aimed at engineers, is there real justice in AI coming for the giants’ moats, or does it just feel good to say? I go back and forth.

Context: this piece draws on the early-October 2026 release of ArtCraft, a suite of seven open-source, AI-assisted “clean-room” reimplementations of Adobe’s creative apps by Atlanta developer Brandon Thomas (Photocraft, Vectorcraft, Filmcraft, Lightcraft, Printcraft, Effectcraft and Designcraft). It was built with Claude Opus 5.5 in Rust and covered by Gizmodo, PetaPixel, KDnuggets, Phoronix and others. He describes the apps as early alpha and “not anywhere close to ready.” The wider “SaaSpocalypse” idea and related clone projects come from the same October 2026 reporting. Details are changing fast, so treat the trend as the story, not any single clone’s readiness.