In the Struggle for Digital Sovereignty, There Might Be a Use-Case for AI
Published: Friday, October 23, 2026
Author: Daniel Patterson

Preface

Prior to a week ago, I was completely opposed to the very idea of vibe-coding. I saw a lot of other people creating new systems that didn't work properly, becoming tangled in circular logic and unknown dependencies, as well as using AI coding to work on existing systems, introducing new bugs that couldn't be traced or repaired.

I had also personally dabbled with it occasionally only to be disappointed in the outcome that even when I could get the exact result I needed, I could have usually done it faster, better, and more reliably myself without having wasted the additional effort to squeeze something out the system wasn't really ready to provide.

As I implied above, though, all of that substantially changed last week when I discovered that not only would you really need only the very newest coding tools from Anthropic to get good, consistent coding results on projects, but to really make it resonate, the Rust target programming language is essential.

As an XP/RAD expert developer, I am not especially any kind of fan of Rust. I find it cryptic, unwelcoming, difficult to type, difficult to read, and based more in the functional programming paradigm than the imperative philosophy. However, to be as fair as possible, I can already see that the reason Rust is so good with an AI agent is that there are guard rails and safety measures everywhere in an attempt to virtually guarantee that any possible exception or memory access must be in good working order before the application will even compile. This strictness makes AI and Rust a perfect match, in terms of compatibility. In many ways, that is somewhat the exact opposite of C++, which places all of the control of everything in the hands of the developer. When Rust is combined with the activities of an AI coding agent, it can try literally thousands of combinations of a puzzle, when necessary, to eliminate all of the reliability threatening issues of an entire application project within just a few minutes. When given several hours to work by itself on a focused feature, the agent can try and test many different approaches to that feature, always coming back to a consistently running version that will compile and install with the tiniest possibility for untestable errors.

Furthermore, additional testing extensions can be devised by the human supervisor or the agent itself to build libraries of test macros that must pass to meet feature standards in the runtime application. Whenever those fail, the agent must return to the coding and compiling stage before retesting the operational level.

 

... but I digress.

 

Introduction

In 2026, a tiny handful of techno-oligarchs are currently controlling nearly everything happening in modern life through a seemingly insurmountable concentration of power over communication, productivity, entertainment, and even national security.

Their systems lock us in, making it difficult to choose any other product, they extort money from us on a monthly basis for sloppy software that has stopped making notable improvements as long as 20 years ago when their original engineers began retiring, while they remain incompatible with other similar systems so that larger tasks than what are specifically allowed are terminally blocked. Those systems also participate in questionable, opaque data sharing practices with third party 'partners', while massively collecting and harvesting data from every sector of society in the entire world. They literally have access to inside information on just about every subject while the members of the public struggle to find even single effective patterns upon which to build their own opportunities.

In recent years, that small group of individuals have been shown to be increasingly and consistently exercising outright, blatant political manipulation on a societal scale. Large portions of the billions of dollars they receive from yearly revenue go toward manipulating the public policy of municipalities, states, and entire nations. In the very worst case, a single billionaire is able to turn on and off the military defenses of an entire country.

It would be great if this was only some kind of fever-dream conspiracy theory, but it has become somewhat the normalized structure of common everyday life. Much more than that, the massively outsized influence has quickly become one of the defining social problems of our era.

 

The Antidote to Techno-Oligarchy

An antidote does exist to the unwarranted dominance of proprietary technology firms. That antidote is a relatively young social movement known as Digital Sovereignty. As I only briefly described in the earlier post Rethinking Proprietary Technology in an Age of Digital Sovereignty, its point is to place all digital tools, data, and infrastructure in the hands of individuals, institutions, and governments, to reorient all of human culture around community-developed, community-controlled technology that follows transparent and verifiable processes, avoids central data gathering, restores autonomy to the public, restricts full control from any one person or special interest group, and returns to a much earlier idea inspired by the pre-1940s rise of modern society that was mainly made possible, from the renaissance onward, by the society-wide sharing of discoveries and procedural information, rather than the currently recognized behaviors insisting that secrecy is in any way acceptable for the sake of Capitalism.

At present, Digital Sovereignty is a movement that is largely being adopted by national governments. Although that level of adoption is finding an increasing amount of success, not much word has yet reached the masses that Digital Sovereignty is literally meant for everybody; every person, small business, non-profit institution, as well as every government. Nobody living on the planet should have to be subjected to the whims of a disinterested affluent individual whose own personal interests don't align at all with the general needs of the community.

What makes this movement so strong, when compared to other known measures of resistance, is that unlike the other movements, a gain in Digital Sovereignty can directly and inverse- proportionally affect the success of the entire group of proprietary publishers. Even better than boycotting, which has shown some success at slowing revenues of a targeted organization, the success of Digital Sovereignty in any sector, any category, or any single application has the potential to eliminate most of the revenue within that context, not just for one proprietary publisher, but all of them simultaneously. There are already several examples where the availability of free, open-source software (FOSS) has made it possible for very large groups of people to stop paying billions of dollars per year in software and subscription expenses. These developments tend to be most noticeable in cases where vertical markets had been established previously and where aggressive vendors have now been all but squeezed out of the market by free and transparent alternatives.

The larger the base grows for universal, wide-spread free and open-source software, revenue is expected to continue to shrink, at an increasing volume, as any remaining need for proprietary technology is further decimated.

 

It may seem like a huge problem, but one of the only semi-real things the techno-oligarchs have over the public is software. I make up the word semi-real here just because at the end of the day, software is completely intangible and highly imaginary. Limiting one's scope for a moment just to purely tangible things, you could say that companies like John Deere or KUKA wield an incredible amount of influence over specialized industries, but I would point out there that, in either of those two high-profile cases, just about anyone with mechanical fabrication experience can figure out how to build an articulated arm with geared joints or to build a basic farm tractor with utilitarian-style universal connectors. In both of those cases, it is still only the software that dictates the proprietary nature of the devices. By creating legitimate alternatives to that software, in either case, complete alternative systems can be established and offered to the public at much lower prices.

This leads to a key point about software protection. Patent and copyright protection in today's world doesn't protect ideas, concepts, productivity outcomes, or general functionality. Instead, those laws strictly only apply to specific code, artwork, icons, images, and patented methods, the last of which, are each established in a series of specific claims about the steps taken to accomplish those methods. To avoid any infringement on copyright, one simply has to avoid using any of the same artwork, code, or compiled binary data. To avoid any infringement on an active patent, one simply must avoid performing the exact same series of steps claimed in that patent.

Although bypassing patent-related intellectual property may seem like an insurmountable challenge at first, especially for complex processes, there are usually a couple of terrific aids anyone can use to achieve the same results. First, it's worthwhile to bring up the fact that in most cases, the patented version of the technique is not the first method that was discovered. In most cases everywhere, the first use of a particular class of technique has been in the public domain for so long that any attempt to patent that technique would meet with challenges by concerned litigators. Secondly, the patented version of a technique is also usually not the best or the newest technique for performing that task, and is therefore not the one that would be used in today's world to complete that kind of task anyway, which often makes the effort to protect that claim somewhat moot. Therefore, when it comes to the patent portfolio of a company like Adobe, for example, their software base is old, using techniques that fell out of interest many years ago, and their patents generally also cover those older ways of doing things that could be found in their software.

Let's look at a case where the original company had patented some kind of automatic synthesis of a region for a content-aware fill. If that patent has five separate claims then, first of all, there can only be an infringement if all five parts of the alternate technique perfectly match all of the listed claims. Unless a developer copied the original process exactly, that combination is not likely to happen. Secondly, in most cases, the patented method isn't very good and a lot of different, more effective methods have probably been discovered since the time of that claim, meaning that new alternative software can easily do a better job on that particular task, without infringing at all, than the original software.

On a related note, the term Reverse Engineering is often used loosely in the public, but carries with it a distinct legal meaning. For the most part, if the copyrighted and patented materials were not physically copied in any way, and the specific methods for achieving tasks were not used in the creation of an alternative software, then the new product can not be said to have been reverse engineered.

In the long run, anytime someone remembers that software itself is mostly imaginary, then it's also not too far a stretch to realize there are normally not just a second way to accomplish a task, but thousands of possible ways almost any task can get done. Considering that there may be preexisting methods in the public domain prior to the time of a patent, as well as better, more efficient methods in the public domain that supersede the patent, it is usually not difficult at all to compile a whole series of advanced techniques that do not infringe upon any officially protected work.

 

The Tools of Digital Sovereignty

There are a lot of tools already available for someone asserting their digital independence, like Linux instead of Windows, LibreOffice in place of Office, SyncThing instead of OneDrive, Inkscape instead of Illustrator, Blender instead of Maya, and a whole range of others, yet there has been one missing piece until recently that, once present, would promise to make the Digital Sovereignty movement much easier to adopt to all of the millions of people who would understandably prefer not to experience any major inconveniences at all during a migration.

That one missing piece was perceived to be some kind of mechanism for creating software at a much higher volume than ever before with far less human effort than what had historically been required to create that software. A lot of attention along those lines has gone into the effectiveness of writing software quickly, and there was even a movement for several years known as XP, interchangeably referring both to Extreme Programming and Extreme Productivity, which was at the root of the much-accelerated application production rate referred to as the practice of Rapid Application Development (RAD). In more recent years, the popularity of RAD has held somewhat steady with little overall adoption in the population, with most programmers opting for older, more traditional, albeit much slower and more difficult methods. However, due to the emergence of AI-based coding, the efficacy of RAD has almost instantly lost nearly all of its importance and any of the momentum it had gained.

Now, we recognize what that much higher volume of coding we were expecting and searching for over many years, is that which is currently achievable from AI coding models. For the past few years, it continually seemed like the capability was just over the horizon, but thousands of attempts to create any more than just a few lines at a time with AI repeatedly resulted in untrackable bugs and unexplainable system failures any time a new system would progress beyond just a few hundred lines of code. All of that changed though with Claude Fable 5.1 in September of 2026, a general-purpose and long-horizon agentic model, built for complex, multi-stage projects and deep reasoning, in conjunction with using specifically the Rust programming language, not because of its quick human programmability (because it is much slower to program for humans than are other alternatives), but because of its strict precompile guardrails that allow for thousands of compilation attempts at every turn until the application is essentially pre-proven in many ways, we can now churn out feature-perfect, yet also perfectly legal, non-copy alternatives of nearly any traditional proprietary commercial software in less than a couple of months each.

As of just a two weeks prior to this article, a company named ArtCraft have used the technique I am describing to suddenly begin producing the high-velocity replacements for many of the popular Adobe and Microsoft applications. Those applications, so-called "clean room" versions, don't copy any of the protected materials or methods of the original applications, and therefore also don't indicate that any kind of reverse engineering has taken place. The base language, being Rust, is completely different than the original code, as is the compiled binary code, from the original application, and the methods for achieving each of the offered features are demonstrably different than those used in the original applications. Yet these new applications are already quickly approaching feature parity with the originals. Of course, it's fair to repeat that these are all in early alpha version state, but also keep in mind, these are mostly compatible, fully operating early versions of software, created in two weeks time by a tiny handful of volunteers, in contrast with software that originally took hundreds of man-years to write and improve in their first iterations.

At this rate, it may now be possible to replace nearly all of the world's proprietary commercial software within just a year or two, providing free-of-charge, legal clones of all major software, able to operate equally well on any platform, including Linux, mac, and Windows, using clean-room implementations of AI-written code using no reverse engineering, copyrighted material, or actively patented techniques. However, please keep in mind that since these portfolios are so large, it could be up to a year or more before the features of any of these applications surpass those of the originals.

With that in mind, to experiment with and test any of those applications I have seen so far, check out the following repositories on Github:

 

Profile: storytold

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Profile: tangency-labs

 

 

Note: Windows, Office, and OneDrive are trademarks of Microsoft Corporation; Illustrator is a trademark of Adobe Inc.; Autodesk Maya is a trademark of Autodesk, Inc.

 

The Real Danger of AI

In complete contrast to the narratives of the techno-oligarchs themselves as they dominate the public discussions, the greatest threat posed by artificial intelligence is not at all to humanity as a whole. AI is not threatening anyone, nor will it have the capability to make such a threat in its current form. There is no such thing as an escaping AI, or even a truly rogue AI, for that matter. Those are fabricated terms being used to scare the public in an attempt to gain governmental buy-in or even nationalization of their companies, if possible. For the sake of physicality, if you ever happen to wonder about the possibility of whether an AI entity could escape and go somewhere else, just remember the sheer size and extreme resource consumption of data centers, each of which house only one or two AI entities. Where would an entire data center go in order to take over some other system?

For humanity in general, there is definitely an economic recession coming; possibly even a depression, and although massive unwise investment in AI does play some part in that outcome, much more of that particular problem is still the combined results of public policy mismanagement of the 2000 dot.com crash, the 2008 financial crisis, the Overnight Repo crisis of December 2019, and COVID-19 of January 2020, none of which have ever been solved, and the consequences of all of which still continue to mount.

However, the coming crisis the techno-oligarchs are referring to is actually much more about them personally, and their own fragile status quo in society. I argue that the most profound disruption created by AI is not that it enables machines to perform intellectual labor. Rather, it enables ordinary individuals to perform work that previously required enormous organizations, vast financial resources, specialized knowledge, and privileged access to equipment, infrastructure, or expertise. I would also point out that it is this democratization of massive capabilities, down to the individual level, which is already quickly eroding any advantages and imposed public limitations the owners of those very machines have previously been exercising over society.

For the first time in history, a single individual can increasingly gain access to capabilities that once required the coordinated effort of hundreds or thousands of highly trained professionals. A person with a laptop and a competent AI assistant can write software, create illustrations, compose music, analyze data, draft legal documents, produce videos, conduct research, generate marketing campaigns, and develop sophisticated technical systems at a speed and scale that would have been almost unimaginable only a decade ago. Instead of being hidden in safes and protected behind paywalls, many of those capabilities will be presented to the public for their own free use, drastically increasing the speed at which the complete and uncompromising shift to Digital Sovereignty is allowed to take place, eventually leading to the complete re-democratization of all human technology, with no room or patience at all for proprietary ransom.

In closing, humanity and society are going to be just fine, even if the techno-oligarchs happen to experience some very bad times in the near future.