A Fundamental Basis for Digital Sovereignty:

headerimage

Introduction:

By Digital Sovereignty, I mean having full control over ones online identity and data on the internet.

There are many different issues that have been discussed over the past 30 years. Each one on its own however, is disparate. For a long time I have wanted a more principled basis, that is informed from hard-earned personal wisdom. Telling others about Kindle renting you books, or a streaming service possibly locking down your data when they change the freemium terms, appears as corner-cases to the lay person.

Then again, I will be explicit: The argument I am making is primarily to target those who are perceptive, and old enough to have deduced basic truths that occur in all human societies. If the average person doesn’t get it, it doesn’t really matter: because they likely will never get it. Even if they did, they are too overloaded to be able to do anything about it.

This article is for people who realize that Digital Sovereignty is the solution to oppressive and pernicious effects that arise from hyper-complex online societies.

The Principle-Agent Problem:

The Principle-Agent problem refers to a relationship between the Principle (entity being served) and the Agent (entity that performs actions or management on behalf of). The Principle hires, or is managed by, the Agent. The Agent manages, or serves the Principle. Nearly any legal, contractual or bureaucratic relation can be described in these general terms. For example: A board of directors at a public corporation answering to shareholders, or a core group of elected representatives administering a group.

In theory, the agent serves the principle, but mishaps and issues frequently arise. These are due to self-interest and information asymmetry. Another example: A board of directors has insider information about negative industry trends coming up, but chooses to sell their shares and keep it a secret (until after the shareholder’s meeting), because they know the stock price will drop.

In general: When you have a core group of people that administer things in service for the rest of the said group, issues can arise where they are placed in moral hazards [1], or must act in a self-interested way that does not benefit, or actively hurts the Principle group itself.

A wise person knows such a thing, because they have seen it in multiple contexts throughout their lives: A political party that makes promises, and implements almost none of them after a four year stint. A student club that has left over funds, and instead of distributing them to members, goes and blows it at the student bar because “they worked hard and they deserved it”.

But more relevant to this article, we see this all the time in subscription based software and freemium services.

Problems with Freemium and Subscription Services:

One of the reasons software moved to subscription services is because it allows for more regular and predictable profits. Imagine the old way: clients buying one license per year. If you have a number of corporate clients that buy at the end of the fiscal year (when an OpEx budget is approved), there is a harrowing wait to see if they renew, or if they went to another provider.

The quarterly/yearly profit is perpetually in crisis, which puts a lot of pressure on the sales and marketing team. They can check in with each one of their clients, but the clients could just lie and go with another provider.

The customer doesn’t have to be accountable to them, in effect, because they need the money and the customer is always right.

With monthly subscriptions, the possible churn is reduced. And in the event of a one month drop in profit, the sales/marketing division has more time to pursue new leads or pull back old clients.

Would you rather have a heart rate monitor that reports a pulse every second, or every minute? That time lag makes all the difference.

Now consider the engineering division. You have different versions of software available. Some clients are using the 2019 version due to legacy reasons, other the 2024 version. Many clients (particularly large ones) have custom plugins and integrations for your software. This creates headaches when upgrades have to be released. How can this be made more simple?

Easy: SaaS and you control the software on your own servers. The plugins are just API calls over the internet - let their teams build them to their own specifications. Upgrade clients to newer instances of the software, and keep back some legacy features to transition them over. No need to worry about what MacOS X or Windows 8 legacy OS’es will respond to software patches, because its all running on your own servers anyways.

We see in both cases, the companies have chosen to make their own lives easier, at the expense of clients.

This is justified by streamlining services, overhead, and providing more “performant solutions” to clients. But in response, clients have lost control and are just renting services. And now they are at the mercy of the providers.

Conclusion:

The Principle-Agent problem applies to any service provider, or administrative entity. As they are only human (like you) - moral hazards and perverse incentives can expose you to undue risks and exploitation.

Platform Decay:

The idea of “Enshittification” [2] - that freemium services eventually “go to shit” or more specifically - degrade in user experience in tandem with their pursuits of additional profits - was originally coined by Cory Doctorow. I’m not a fan of this word (it is an example of frustrated and childish millenial speak), so we will use the sanitized term Platform Decay instead.

Platform Decay is Caused by Executive Decisions:

We can illuminate all of the obtuse decisions public companies make (from the perspective of users and clients), by using simple concepts from Corporate Finance [3]. Public Companies have responsibilities, mandates and all are operated by executives, who basically follow the same set of rules:

  1. A company is responsible for generating profits (or minimizing losses in bad times), for its shareholders.

  2. A Quarterly report must be issued by the management team of a Corporation, to inform the shareholders of the companies health, progression, and events of the year.

  3. Shareholders may vote to change management and board members, if they feel unhappy with the way management is running the company.

In other words, The management team at any corporation hypothetically has its head in the guillotine, at every shareholder meeting. Keeping their jobs is important to them.

A second aspect to consider is how future profits are optimized. Excluding general industry trends and geopolitical events, this largely depends on the where the company is in its lifecycle:

  1. Companies that are still growing invest a portion of their profits (or may even go into debt) to fund Research and Development (R&D) to innovate and create new products and services for their particular market. This expenditure is done to keep abreast of the needs of the market, which is critical to capture future profits.

  2. Every quarter (or year), the management team of the company needs to assess which projects to pursue, either to maintain profits, or expand profits. Each project is broken down into costs and expected revenues or profits. A hurdle rate [3] is calculated for each, and compared to determine what to focus on. In addition an Operational and Capital Expenditure budget must be allotted (OpEx and CapEx).

In essence, they must spend time and money on maintenance and research project to maximize profits in the future, and a spreadsheet decision is made by boiling every possibility down to numerical metrics that can be optimized.

  1. Companies that are in the late stages of their lifecycle (ones that cannot grow more market share), are still pressured by the relentless grind of optimizing profits every quarter. If a company cannot invest in R&D to grow future profits, and cannot easily steal market share from other competitors, it must turn inwards and reduce costs overtime - in order to optimize profits [4]

There is also a sub-category of companies we have to deal with: Tech companies that are funded by Venture Capitalists (VCs). VC’s eventually want their money back. If a company has an entrenched enough client base dependent on their product, they can be pressured by the VC’s to be more aggressive in their profit taking, leading directly to platform decay.

Conclusion:

For any online subscription service, Platform Decay will occur over a long enough time frame. The nature of profit maximization for any company virtually guarantees this.

The Algorithm is the Message:

You are probably aware of the Marshall McLuhan quote “The medium is the message”, in that the message being expressed by the medium is necessarily restricted and warped by the medium itself - specifically by its strengths and limitations.

To illustrate this quickly: consider television. When is the last time you saw someone solve a second order ODE on a black-board on television? Unless you were watching Open University in the 1990s, answer: never. No news program or talk show could do such a thing - for it would immediately alienate the crowd and viewership would drop.

The medium itself (and by extension, the demographics that it serves) cannot accomidate such a thing - so it is truncated away and never considered, nor presented.

This dove-tails us to the problems of feed algorithms for web platforms [5]. The algorithm is a custom mixture of what is viral, what the crowd finds popular, what filter bubble the user builds around themselves, and what advertisers pay to have served to the crowd.

The only portion you can control, is the filter bubble you choose to build around yourself, via blocking/filtering and following + like endorsement. In some cases, you can pay for premium service to reduce the ad-man’s effects as well.

But of course, those elements you cannot control still greatly warp what you will see, and what can be expressed on a given platform. As attention and exposure are currency, and content formulas are derived by users, to game the algorithms [6]. There are lots of excellent things that could resonate with you, that you will never see because of this.

The Original Dream of the Internet:

The original dream of the internet [7] was that people all over the world, could share information, media and connect with one another; unbounded by class, nationality, or status. As a caveat, this doesn’t mean that everyone has to get along, nor will they connect. But the option to build bridges with others is there – and sometimes matches would occur.

Of course the internet did not turn out this way. Individuals struggled to find and organize themselves, let alone find information. Storing and sending information was expensive. So there was a need for government incentives, organizations, and big tech companies to sort it all out.

In order to avoid pay-walled gardens (such as AOL), the internet denizens implicitly accepted exposure to ad tech and marketing, in exchange for (mostly) free services and content. As big tech needs to generate regular profits, the monetization of the internet (via web 2.0 services and adtech) also developed along-side. Via ad revenue, content creation was incentivized, and the internet was literally filled up with every piece of content you can imagine.

We are left with an internet controlled by big tech algorithms, strong interest groups, and mixtures of corpo-government censorship and misinformation, laden with ad tech that hijacks our attention and overloads us with information. Its like staring into the sun, only instead of bathing in pure light, we are mostly pummeled with noise and spectacles. We are rendered ineffective via Narcotizing Dysfunction and still unable to connect with those we want to meet. [8].

And sadly you can’t meet them. There is too much noise. Try searching the things that you like in the Twitter search bar. You go to Reddit and look for sub-reddits that talk about your favorite hobbies, or niche issues you care about. And you can’t find anyone. You are alone.

Algorithms have failed. Because those that run them benefit themselves, and their benefactors, at your expense. You are not the customer base - you are the cattle. The people that pay money are those that are actually served. Ad revenue will always exceed user revenue on these platforms.

And you wonder why I started my own website. Because I know I can express myself to my exact specifications.

And because I can avoid formulaic slop that I can’t fully shore up, no matter how much I tweak the algorithms with user settings and engagement. Surely, you have seen these kinds of things before:

ff_news ss_email

It is baffling to me, why people don’t step back and think about what they are being served. Look at the crap that is being sent to me (above). This is but a small sample, but I know on your phone, your discovery bar (in your web browser), and in your various news feeds, you see it too[9].

What is jarring about all of this, is how non-impactful and micro-modularized all of these articles are. They are doled out, dispensing wisdom a few words at a time, charging relatively high piece-meal rates (your energy and attention). And people have slowly just accepted this - their attention spans and mental peace to contemplate higher ideas, ruined.

Consider the following thought experiment: Imagine you lived the same life you did now, but exactly one thing was changed: Across your entire life, you never hear the song “Video Killed the Radio Star” by the Buggles.

Would your life really be that much different, if this never occurred in your life [10]? Other than not being able to follow along at your friends 80s theme party for that one song, nothing would radically change your life in any way.

Further to my point - all of those articles above won’t fundamentally change your life in any way either. Yet, lots of creative writing and online blogging has been mutated into this hyper-palletable, low-info density slop. Sad!

Conclusion:

Feed algorithms serve the needs of a number of different parties that aren’t you. And their needs are different from yours. If you are not paying for it, you are the product and will be exploited. Even if you are paying for it, the Ad-man is usually paying the company more. You are still at a disadvantage.

Understanding Contractual Mutation:

A Lesson: The Budapest Memorandum:

In 1994, The US, UK, Ukraine and the Russian Federation signed what is known as the Budapest Memorandum: a document outlining security assurances for Ukraine, if it would commit to de-nuclearization. Per the Soviet Union’s collapse, its aging nuclear weapons infrastructure was fractured between Ukraine and the newly formed Russian Federation.

This was a significant part of the west’s push for the START treaty - and the eventual removal of nuclear removal of nuclear weapons from Ukraine (which were returned to Russia), was heralded as a success for nuclear de-escalation.

The following concerns and issues were likely at play for all parties involved:

  • US/UK: Were well developed, prosperous nations, with the two governments leading the way in reducing nuclear arsenals and tensions throughout the world. With the cold world over (and globalization + free trade rising), Clinton and Major (respectively) pushed for such outcomes, globally.
  • Ukraine: Had seceded from the Soviet Union in 1988, and was in significant economic hardship as of 1994. They had nuclear weapons, but didn’t have the ability to service them, or properly use them. There were some fears of aggression from the Russian Federation.
  • Russia: With the fall of the Soviet Union in 1991, the newly formed Russian Federation also had problems with cohesion and stability. They did not lack military equipment or men, but were in great economic and organizational turmoil - converting to a capitalist economy and whatever their version of “democracy” was. They were not in a strong position to oppose the West.

Essentially, Ukraine held on to nuclear weapons as it would be a token deterrence from other nations (including the West/Russia) from invading on its territory, or exploiting it in other ways. The West, wanting deescalation, gave aide and assurances to Ukraine so that these goals could be achieved.

All parties eventually came to an understanding, and signed the memorandum[11]. But of course, recent history shows that a “memorandum with security assurances”, amounted to nothing.

When Russia attacked Ukraine for the second time, Clinton et. al had long been gone, and the states and intentions of each of the nations involved had changed significantly. It did not matter if Boris Yeltsin had given an assurance - for Putin had been in power for over twenty years and build up his war machine against the West.

In essence - each actor in this play was its own ship-of-Theseus, and was slowly crippled with its own developing problems in a kind of Boiling-Frog Syndrome. The concerns and players are inevitably, very different after a generation.

It is this phenomena which I label Contractual Mutation - when both the subjects, parties and states involved in any agreement all mutate to the point of any such agreement maintaining its original goal and meaning. It is a common thing that occurs at all levels in life.

Everyone and everything is always changing - and these assurances only hold for the short term.

“We reserve the right to modify these ToS…”

“The Company may, at its sole discretion, revise, modify, or replace these Terms of Service at any time without prior notice.”

Continuing back to our Digital Sovereignty arguments, you have undoubtedly signed up for a service where statements like the one above, are made by a service provider. Likewise, Contractural Mutation over time occurs - whether it is companies trying to maximize profits, or a change of management which changes the management style or the relations with customers. You can’t ever really be sure. So you never have a bedrock to ever stand on in the first place.

The point is: If you rely and control things by yourself - half of the problem is already eliminated. You can’t get screwed by others. If you stop caring, its because you changed, and it doesn’t matter anyways. If you are a sovereign individual and screw yourself, you accept responsibility and learn from it (in principle - your own ego notwithstanding).

Conclusion:

We want stability and certainty from other entities, to control and manage things in our lives. But no matter what contractual agreement we make (formal or otherwise) - each entity (including us) is a ship-of-Theseus, and we are all slowly changed by our environment (Boiling-Frog Syndrome). Contractual Mutation always applies.

All the Low Hanging Fruit is Gone, and Everything is Done:

All common software (i.e: messaging apps, parts databases, peripheral drivers, etc) is stable and complete. Software developers (in there own need for efficiency, and future laziness), have cataloged every solution, which has been fed into super-intelligent chat robots. For smaller applications, an individual doesn’t need to hire a software developer to build them custom solutions. They submit a design idea to a chat robot, which figuratively takes software components “off-the-shelf”, and composes a decent solution.

Secondly, as per our previous point, we don’t have to worry as much about technical debt or poor software design (read: the neurotic concerns of hard-core software developers), because we ourselves and the environment are changing so quickly, such technical debt, won’t likely matter anymore.

Even if the composed solutions are sub-optimal, for everyday consumer needs (a post-it note app, mp3 converter, etc), we all have so much computing power, that it doesn’t need to be optimal in the first place[12].

In other words, its easier to build custom things for yourself, and it is very easy to do so!

Conclusion:

For common software tasks that require normal amounts of computing power, cut the other party out and build it yourself. API calls over the internet are always slower than local computation. SaaS is a lie for common tasks in this day and age.

Summary: A Foundation for Digital Sovereignty:

To recap, we can categorize the topics above in terms of different forms of personal freedom:

  • Issues with Negative Freedom (“freedom from…”)
    • The Principle-Agency Problem.
      • Platform Decay
      • Pathological Effects of Algorithm/feed usage
    • Contractual Mutation
  • Positive Freedom (“freedom to…”)
    • You can easily build it yourself.
      • All common software already exists, is mature and stable.
      • Computing power and hardware is cheap. Your everyday needs are not that complex.

So from my years of intenet usage and derived wisdom, the above constitutes my fundamental basis. The negative-freedom issues push us towards administering things ourselves, and the positive-freedom issues pull us towards creating our own custom tools and solutions. With this, we can maximize our own control, and tailor solutions to ourselves where external party dependencies (with competing and complementary interests) can thus be minimized.

END

References & Footnotes:

[0] Header Image taken from the MTG card “Yagmoth’s Edict”. Illustrated by Scott Kirschner.

[1] Specifically, when a group takes on additional risky behavior, because any drawbacks or consequences are bourne by another group, and not themselves.

[2] Cory Doctorow’s Blogpost - On Social Quitting

[3] Damodaran, Aswath, Corporate Finance: Theory and Practice (2e). Specifically, Chapters 2 and 7: The Objective in Corporate Finance, and Estimating Hurdle Rates for Firms, respectively. Generally speaking - a firm has different projects it can devote time energy and capital to, and it must choose projects that it believes will optimize profit. The hurdle rate is the cost of raising the capital/equity to work on said projects. Generally, any project chosen must have an expected return greater than its hurdle rate, and you try to choose projects that return the highest profits. I don’t pretend to be an expert in this area - I use the terminology somewhat loosely.

[4] A common one that is rearing its ugly head in this day and age is the automation of lower level white collar and blue collar jobs (lower skill). For example: Why pay eight graphic designers and two managers, when you can pay two “10x” graphic designers (who man an army of AI tools), and just have one competent manager? The difference between the salaries you don’t have to pay, and the cost of the subscription fees for enterprise AI tools, amounts to additional profit optimization via cost savings.

A corollary of this point (and the idea of project hurdle rates) - is that start-up companies (who don’t even bother to arithmetize their project decisions), and don’t have enough internal manpower to do cost savings, may be less prone to automating out members of the company. For small start-ups, every worker is effectively silo’ed and is essential. But then again, there is the additional stress of working for a start up. Pick your poison, essentially.

[5] We will specifically focus on media centric sites (videos, blog posts, or tweets). Not sites that aggregate products, vendors or services (i.e: AirBnB, or ebay). I specifically bring up this point, because sometimes these sites can put you near the front of the queue when it comes to serving your best interests. For example, on Ebay, it is the buyers that bring the money to the table - so as a buyer you are treated rather well. Intense pressure from Amazon also helps to keep Ebay somewhat in check.

[6] There was a time once, when I decided to look outside the Twitter feed algorithm and my filter bubble, to see what I was missing. This was before the API was locked down by Musk (and the ecosystem of support apps, completely destroyed as a result). Using the free-tier API calls, I would search for posts with certain combinations of words, to see if other people would search for what I was searching for. I got tens of thousands of matches. I even went through them all with a Jupiter notebook at one time. And what did I find? Effectively nothing. Perhaps 100 decent tweets out of 10000. Unfortunately, the Twitter feed algorithm is pretty good. The “best” tweets are presented. Still though, I’ve encountered enough incredible and prescient tweets down the back-alleys of reply threads, that make me wonder just how much the algorithm misses…

[7] There is no original reference to this idea, which I have seen drift around the internet in various places, when people get misty-eyed about the old internet of yore. It simply refers to the period of 1995-2005, before social media and web 2.0 became the dominant presence on the internet. Where people could connect, and everything was new and unmolested, essentially.

[8] No, the realization that I am here on my little idiosyncratic web site, posting little blog posts and humping the dead dream of the Original Internet, has not eluded me.

[9] And i’ve already blocked the Google Discovery app (Left swipe) on my phone. I have already deactivated the discovery bar in Firefox and Chromium. Yes, you can do that. The problem is that this is everywhere, and you can’t just toggle a setting for every website to avoid it. Don’t pretend you have viable choices.

[10] I ignore all appeals to the butterfly effect. That is pure midwit thinking.

[11] You can see the original document here, for reference.

[12] Consider the following post-it note application I designed for myself about 6 months ago. It uses the following hack: All the post-it notes are stored in a json file. And every time a HTTP POST request is sent from the FE to the BE, it just loads the entire database, and reads/writes the entire database back. There is no synchronization from HDD to RAM. I reasoned that the total number of post-it notes I will have is < 10000, and this would produce a file of about ~10mb. Given that the application runs on a home server, the maximum load/request time is of the order of ~1s. Any status signaling CS grad would laugh at this. But given the size of the problem, and my computing power, it is absolutely fine. I still get the same results, and saved a lot of coding/test time by not having a synchronized copy sitting in RAM. In two years I probably won’t even use the application much anyways, so there is no need to waste such time.