In the space of a week, two stories about AI and reverse engineering took over developer feeds. One is a set of free, open-source apps built to replace Adobe's Creative Cloud. The other is a GitHub project called REA that promises to reverse engineer almost anything. Both are real and both are interesting. Neither is as finished as the headlines suggest.
The Adobe clones
According to PetaPixel, developer Brandon Thomas released seven free apps under the ArtCraft name, covering stand-ins for Photoshop, Illustrator, Premiere, Lightroom, After Effects, InDesign and Acrobat. They run on macOS, Windows and Linux. TechSpot reports the motivation: Thomas said he was "bit by the dark pattern 'cancellation fee' one too many times."
Coverage describes the apps as written in Rust, with WebAssembly builds for browsers, using an AI model. Ars Technica's report, as summarised by other outlets, names Claude Opus 5.5.
"Reverse engineered" or "rebuilt"?
Several headlines, including Futurism's, say Adobe's suite was reverse engineered. The reporting we've read describes building clones of what the apps do, which is a different thing from taking apart Adobe's own code. Which one it is matters a great deal, legally and technically, and we haven't seen it established. Treat the phrase as shorthand until someone documents the method.
What isn't proven
- Parity. Thomas has said the apps will reach "100% feature parity within a month." That is a stated goal, not a result.
- Maturity. One analysis calls the Photoshop stand-in an alpha with around 800 open issues. Commenters have listed many gaps.
- Professional use. Nothing we've read shows these tools opening real Adobe project files faithfully, which is the thing working designers would test first.
REA: the toolkit
REA, short for Reverse Engineer Anything, climbed GitHub's trending page the same week. KDnuggets groups it with the clone wave. A more sceptical write-up describes it as a toolbox that still needs you to bring the binary and the Ghidra disassembler, and argues it can't do the dramatic things people imagine, such as breaking into a bank. In other words, it speeds up a skilled person's workflow rather than replacing the skill.
Is Adobe in trouble?
One essay asks exactly that. Our reading: not this month. Adobe's moat is file compatibility, plug-in ecosystems, colour management, enterprise contracts and decades of edge-case fixes. A weekend clone can't copy those. What the story does show is that the cost of building a credible first version of familiar software has dropped sharply, and that subscription friction such as cancellation fees is now a motive people will act on.
What to watch
- Whether the clones reach real file compatibility, and how many of those 800 issues close.
- Whether Adobe or other vendors respond, in code, in pricing or in court.
- Whether REA-style tooling produces published, reproducible results rather than demos.
For anyone building in this space, the lesson is practical. AI agents make it cheap to attempt a rebuild, so the defensible part of a product is shifting from the features to the data, the trust and the details that take years to get right.