What if building an alternative to expensive software no longer required a team of hundreds and years of development?
That question is at the heart of ArtCraft, an experiment in using AI to recreate familiar creative and productivity tools, challenging the dominance of Adobe and Microsoft.
For years, the software industry has increasingly pushed users toward subscriptions, recurring payments, and ecosystems that make leaving difficult. Adobe’s subscription cancellation practices have even attracted a lawsuit from the US government, while Microsoft’s growing integration of Copilot into its products raises broader questions about the future of software ownership, choice, and pricing. When they started stealing private data, with no opt-out option, that was a step too far.
ArtCraft is a protest against this direction. It explores what becomes possible when AI puts software development within reach of more people, allowing one developer to challenge companies that once seemed untouchable.
The idea is simple: powerful software should be accessible, alternatives should exist, and users should have the freedom to choose the tools they use without feeling locked into a corporate ecosystem.
AI makes this challenge more realistic than ever. It can accelerate development, reproduce familiar workflows, and help turn ambitious ideas into working applications at a fraction of the traditional effort. The results still need refinement, but the experiment itself sends a powerful message.
ArtCraft is proof of a possibility: the giants of software no longer have to be the only ones capable of building the tools we depend on.
But after all the fanfare around the media, regular Adobe users trying ArtCraft products were immediately disappointed.
I’m not shocked. Because every time I used Claude to code – it was extremely frustrating once you scratch below the surface. I did try: the products are usable but they are not on par with the big software yet. These ArtCraft programs are using Rust – which is another hype almost as big as AI that is suspiciously slow every time it counts. This is no exception. These apps are crushed whenever they encounter a moderate task.
Yet another case of AI doing a fine job building something that looks very close to perfection – and no matter how many iterations, reworks, new models – never reaching the desired outcome.
