A developer rebuilt Photoshop, Lightroom, Premiere and four other Adobe apps with AI and put them all on GitHub for free. By the project's own roadmap, the Photoshop clone has about 76% of the features, about 45% is "ready for real work", and Photoshop still doesn't reliably open the PSD files it saves.
You can't deny it's disruptive. If I were running one of those companies, I'd be looking at the stars piling up on those repos and wondering what else could be copied this way.
But is it that easy? Menus and tools are visible from the outside, so they get rebuilt fast. Most of the value sits under the shell of the product, in a huge amount of domain detail you can't see from there, like color science, file formats, camera profiles and years of edge cases someone found and fixed one by one.
I've seen a lot of posts saying software is dead and code costs nothing now. I partly agree, because writing code got much cheaper and I build some of my own tools with AI, but I think the conversation swung too far the other way.
The most sensible take I've read in the middle of all this hype came from Fabio Akita, who wrote a long piece on the clones. He put it like this:
"PhotoCraft today is that Ferrari replica built on top of a VW Beetle: the body has the right shape, you take a photo and most people say "wow, a Ferrari". Then you turn the key and out comes the Beetle engine."
I see a version of this in the industrial robotics software I work on, where the ground truth is the physical world. You can generate a lot of code very quickly, but without a faithful representation of that world to test against, it's hard to even tell whether you're looking at the 76% or the 45%.
https://akitaonrails.com/en/2026/10/13/ai-reverse-engineering-is-adobe-dead/