You might have missed this one, as it is very tech related, but really important nonetheless, so let me get you up to date.
Perhaps you already know that Artificial Intelligence agents are now powerful enough to build entire software applications that would have taken thousands or even tens of thousands of expert human programmers in hours or days at most.
Of course those agents need to be carefully managed and their output reviewed (sometimes even by other AI agents) to make sure the final product is correct, but so far the pros heavily outweigh the cons.
An unintended secondary effect from the ability of AI agents to create code and entire applications is their ability to take existing applications and reverse engineer them into source code that can be understood and analyzed.
And I want to stop there for a minute, so I can explain you what all this about source code and reverse engineering means in software.
Imagine you go to a restaurant, where they prepare a special dish you really like.
You visit them very often, because it’s really delicious and you don’t have the ability to prepare it yourself.
But even if you had the ability to prepare it yourself, you don’t know the recipe, and the restaurant keeps the recipe as a well guarded secret.
So you cannot hope to ever prepare that delicious dish yourself.
You talk about this with a chef, friend of yours, and he promises to accompany you to that restaurant, taste the dish, and using his knowledge about how food is prepared, find out which ingredients (even the secret ones) are part of the recipe, and also find out the recipe itself.
So, your friend tastes the delicious dish, and is able to find out some of the ingredients, and a possible recipe, but when you both try to replicate it, you realize it is not perfect, and it doesn’t taste the same.
Bummer.
Then you talk with a chemist, friend of yours, and he proposes buying the dish, and analyzing it at the chemical level, to try to find the secret ingredients.
And, of course, you do that.
Then your friend comes back telling you it is made of carbon, oxygen, nitrogen, etc. and its specific proportions.
But you immediately realize it is completely useless, because even if it is the precise composition at the molecular level, you don’t really know which ingredients it has, and more importantly, the recipe.
You had lost all hope of being able to prepare that this by yourself, when a friend of yours that works in a frontier laboratory tells you they have developed a food disassembling machine capable of taking in an already prepared dish, and turn it back into its ingredients!
But not only that: the machine had a huge amount of data on how every dish in the planet is prepared, so once it knew exactly which ingredients went into the dish, it could perform billions of simulations to find out which exact recipe, with those exact ingredients, would produce the dish you crave so much.
So you buy the dish, feed it into the machine, and the machine outputs: the recipe and the actual ingredients what were part of the dish!
So now you call your chef friend, and following the detailed recipe and using the ingredients you got back from the food disassembling machine, you prepare the dish, and it is perfect!
Well, in the software engineering world, the already prepared dish (what the computer “eats”) is called object code.
It is your Adobe Photoshop, that you double click and use.
Your Microsoft Windows that boots up your computer.
Your computer only understands object code.
But if you want to be able to create another Adobe Photoshop or Microsoft Windows, you need the recipe and the ingredients.
The recipe and the ingredients are called source code.
And the process of finding out the source code from the object code is called reverse engineering.
As you can easily see from our food example, reverse engineering can be pretty hard.
When your chemist friend came back with the molecular composition of your food, you couldn’t actually use it to create the dish on your own.
Even if it was very precise.
The same happens with software.
Some reverse engineering techniques (called decompiling) can bring you back the detailed composition of the original source code, but in a way that is not understandable for a human programmer, so no very useful to replicate the dish.
It works, it is precise, but it isn’t useful for what you want.
But our current AI has SO MUCH data on existing programs, on every domain in existence, that it can actually take object code (an already cooked dish) and find out what you had to do to a specific list of ingredients to make them end like that.
The source code.
So now you know what is reverse engineering, why it was a hard problem to tackle, and why our current Artificial Intelligence has been able to solve it.
Let’s go back to the news.
Someone took a swarm of AI agents (think of your small software development shop, but comprised only by AI agents) and tasked it to reverse engineer Adobe Photoshop.
And, of course, the AI swarm came back with the source code (the recipe + ingredients) that would produce an Adobe Photoshop if compiled (cooked) back.
But not happy enough with this result, they tasked another group of AI agents to read the reverse engineered Adobe Photoshop source code, and describe it in an abstract way, like the plans or architectural design of a building.
The kind of work actual software engineers do when you ask them to build an application for you.
And finally, they tasked ANOTHER swarm of AI agents to create an application from scratch, using the specifications and plans produced by analyzing the Adobe Photoshop’s source code in another programming language that wasn’t the original one (that programming language is called Rust).
So now they had a completely new photo editor software, that does almost everything Adobe Photoshop does, built in a different programming language.
Happy with their results, they did the same with the rest of the Adobe suite, reverse engineering them and then rebuilding them in Rust.
And, of course, making them available for free, with their source code included, for everyone to tinker with it.
So, in less than a week, and for a few hundred or perhaps thousands of dollars cost in Artificial Intelligence tokens, he brought the 60+ billion USD Adobe suite to the ground.
(Adobe as a company is valued at around 99-100 billion USD, and their Creative Cloud Suite represents something around 73% of their total revenue, so being conservative, let’s put the total value of the suite around 60 billion USD).
I know my Intellectual Property lawyer friends are probably already squirming in their chairs, thinking about all the copyright infringement stuff being performed during this reverse engineering process.
But this guy was no fool at all.
And that’s the reason he first tasked one agent swarm to decompile the initial Adobe software, then a different one to read the source code and create an abstract specification from it, and finally another different AI swarm to create a completely new product from scratch using only the abstract specification.
In the software engineering world, this is known as Clean-Room Reverse Engineering (also The Chinese Wall Technique).
If one day you decide you want to become a competitor against Adobe Photoshop, and you sit in front of your computer, and start taking notes of EVERYTHING Adobe Photoshop does, creating a design document, then you start programming from scratch an Adobe Photoshop clone without ever having looked at either the object code or the source code, you are not committing IP infringement.
And decompiling the software into source code is usually prohibited in most License Agreements (those enormous legalese documents nobody reads and just clicks OK when they “purchase” new software).
But, as with everything, depending on each jurisdiction, there are exceptions that allow decompiling the software if you legally acquired it, and it is for learning purposes, etc.
So as long as you don’t take actual pieces of the original code, and put them into your new product, you are not stealing anything.
And also it is very important that the “team” that actually saw the original source code, and wrote the abstract specification, doesn’t participate in the development of the new application, because it is assumed it brings the knowledge with them, and inadvertently will program similar stuff.
So you bring a whole new “team” that only reads the specification, and codes a whole new application from it.
Hence, the Chinese Wall.
Again, my lawyer friends will be thinking: “but these laws were made thinking in human teams, not in AI agents that can be spawned in swarms for a few bucks…”.
And you would be absolutely right.
This is an actual hole in the existing regulations.
So I guess lawsuits will come, and at some point this thing will be settled and integrated into the existing legal systems.
We can only sit and watch how this plays out.
But it is a very interesting development, because so far Adobe is not the only company affected, but also entire games are being reverse engineered in the same way, and much more.
Even some notorious Open Source Software advocates, like Eric S. Raymond, are already calling the demise of big Closed Source Software businesses, like Microsoft.
But I think we are far from this.
The most probable outcome is that big software corporations will lobby their way out of this mess, forcing the legal system to patch the hole and prohibit the process that allows to reverse engineer their products.
But even if it gets prohibited, we cannot deny the enormous impact this new capability shown by our current AI could have not only in the software industry, but everywhere.
What do you think about this?
Let me know in the comments!
Conclusion
Thank you for reading my publication, and if you consider it helpful or inspiring, please share it with your friends and coworkers. I write weekly about Technology, Business and Customer Experience, which brings me lately to write a lot also about Artificial Intelligence, because it is permeating everything. Don’t hesitate in subscribing for free to this publication, so you can keep informed on this topic and all the related things I publish here.
As usual, any comments and suggestions you may have, please leave them in the comments area. Let’s start a nice discussion!
When you are ready, here is how I can help you:
By subscribing to my publication and reading this article you have already taken the first steps into redesigning your business to succeed during the Age of AI, .
But if you want to go further, here are several options:
If you want to start right away, you can join me in my 3-day Future-Proof Entrepreneur Challenge, where you will learn how to analyze your Profit Model (how your business makes money today), look around for the Exponential Technologies you will need to make it work in the Age of AI, and finally create a plan to make it happen.
You can check my FREE guide “7 Profit Models to Succeed in the Age of AI”.
And also my FREE guide “7 New Technologies that can wreck your Business like Netflix wrecked Blockbuster”.
The Entrepreneuring AI is free today. But if you enjoyed this post, you can let me know that my content is valuable by fueling my coffee addiction, so I can keep producing it for free 😸☕️