One of the earliest uses of generative artificial intelligence (AI) tools like ChatGPT and Claude was producing source code — computer language instructions that make up software such as word processors, your Internet browser, and everything else.
Code is structured, which makes it easier for AI tools to generate.
So could they help to make fancy software we usually have to pay for? Enter reverse engineering: using AI tools to break a piece of software down to component parts, and then build code to replicate those parts.
Unlike vibe coding, where you tell AI what you want a piece of software to do and it tries to create it, reverse engineering works more effectively for large, complex programs.
In recent weeks, one software engineer used AI tools to produce what are essentially free versions of Microsoft Office and Adobe programs, the maker of Photoshop. To make things more interesting, a freely available AI tool called Reverse Engineer Anything has gone viral in coding circles.
Why should you care? Some people are hailing this as a return to the early days of the Internet when software was free — both to use and to adapt.
Even if we don’t go that far, this new development certainly poses questions about how for-profit software will work in the future.
The cool kind of weird
In the early days of the Internet, there was a cool kind of weird. In the 90s, the Internet seemed to offer endless possibilities, and no one had figured out how to monetize it yet. People spent a lot of time talking to strangers about their niche interests on chat forums, and creating badly designed websites to share information with others.
Lots of software was free or open source. That meant if a piece of software didn’t function quite the way you wanted, you could download the source code and tinker with it until it did. Often, you’d do this with help from the Internet, where technology searches were the second most common search topic.
It seems almost impossible to imagine now, but Google didn’t exist yet. Microsoft did, and quickly became the best-known proponent of for-profit software: programs you have to purchase and can’t tinker with yourself.
The nineties were therefore a battleground between free and open-source software, readily available to tinkerers, and for-profit software, of which Microsoft was the most vociferous producer.
The birth of ‘big tech’
In late 1998, for-profit software got considerable protection in the form of the United States Digital Millennium Copyright Act (DMCA). It made it illegal to attempt to break the locks that prevent tinkering. The future of for-profit software was secured.
And the Internet went mainstream, no longer a clubhouse populated by early adopters. You didn’t have to be as “tech savvy” to buy and use computer programs.
It was also the beginning of the companies now known as “big tech”: Microsoft already existed, Amazon (having started in 1995) became profitable in 2001, Google launched in 1998, and Apple eventually went from a niche, struggling computer company to a tech juggernaut after the launch of the iPod and Apple Music store in 2001. Facebook, the last of the original tech giants, launched in 2004.
All of a sudden, free and open-source software was something used only by the technically adept. Developing it was a special interest. It was no longer possible to remix, improve, or personalize most software, because accessing the underlying source code was technically difficult and illegal.
A new reverse-engineering utopia?
Arguably, because reverse engineering doesn’t break any locks, it may not infringe the DMCA, though this may face legal challenges.
Combined with vibe coding, this new era of AI-powered for-profit software clones could be the beginning of a time where everyone can have their most longed-for function in Excel.
All is not as it seems in this new utopia, though. Some of the software being created is already reported as having weird bugs. Those bugs are likely to be hard to track down when so much of the coding was done with an AI tool as opposed to human work.
File reusability could be a problem, too. If everyone is tinkering with their own copy of Word, there’s no guarantee files will remain compatible. You might be able to open a file in the “free” version of the software, but someone else using an official Microsoft product may not.
The other big change since the early days of the Internet is the importance of data. Even if you reverse-engineer Photoshop, it won’t contain the underlying dataset that allows you to use Adobe’s proprietary AI tools. If you successfully reverse-engineered Facebook, you wouldn’t have access to the underlying behavioral data that makes its advertising so successful.
So, will reverse engineering take us back to the early days of the Internet? Probably not. But it will likely usher in a new era where traditional ways of monetizing software don’t work anymore.
What will happen next is unclear. We can only hope that some of the balance of power between consumers and creators of software is restored, and the future is at least a little rosier.
Dana McKay, Associate Dean, Interaction, Technology and Information, RMIT University
This article is republished from The Conversation under a Creative Commons license. Read the original article.
Follow us on X, Facebook, or Pinterest
