Igor Bubelov About Blog Photos

Self-Hosting as a Hobby

October 5, 2020

Most hobbies have certain practical benefits, but our free time is limited, so we should pick our hobbies wisely. In this article, I share my thoughts and experiences on self-hosting my own digital infrastructure for fun and profit.

Illustration by Sandra Ahn Mode

Table of Contents

Why Self-Host?

What more sacred, what more strongly guarded by every holy feeling, than a man’s own home? ― Cicero

Most hobbies have certain practical benefits, but our free time is limited, so we should pick our hobbies wisely. In this article, I share my thoughts and experiences on self-hosting my own digital infrastructure for fun and profit.

Surveillance capitalism is a relatively new term, and it used to describe a system where companies can track our activity and steal our personal data with the intent of selling it to the highest bidder. Many of us were also rightfully outraged when we learned about illegal and immoral surveillance programs waged by our governments.

The importance of digital privacy and being in control of your own data are the long-standing priorities of the open source community. All of those ideas aren’t new, but most people didn’t take it seriously until 2013 when the world was shaken by Snowden’s revelations.

The information which Snowden shared with the world was a wake-up call for many, but we’re still struggling to figure out the best ways of limiting widespread digital surveillance. Self-hosting can be considered an extreme form of fighting digital surveillance, since it allows you to keep all of your data locally without exposing it to any third parties.

Setting up, securing and maintaining your own server is the best way to protect your data, but it takes time, and you have to learn a lot before you can even start. Also, we’re constantly bombarded by all sorts of a corporate propaganda which talks lengths about how easy it is to start using their products. Not surprisingly, they aren’t that eager to tell you about their shady but very profitable practices such as collecting as much of your data as possible and selling it to the highest bidder.

It’s a shame that tech companies spy on us, but the incentives are too tempting. Now, let’s try to compare self-hosted open source software and “managed” or purely commercial closed-source products in order to figure out the pros and cons of each option and whether self-hosting is a thing worth trying.

Security Without Transparency

Most of the apps and websites we use are “black boxes” in a sense that it’s extremely hard and, in many cases, impossible, to figure out what they’re doing with your phone, laptop or any other piece of hardware you happen to run them on. It’s fascinating how much we have to trust the software companies and how eager they are to betray this trust if it gives them an opportunity to make an extra buck.

Imagine a fast food chain which has an option of making a huge profit by adding a little dose of poison to each meal served. It will most likely pass on that opportunity because we can expose it by auditing the food, and we also have certain laws which are supposed to make sure that anyone involved in such a scheme will end up in jail. Food is transparent in a sense that it can be examined with ease, and it’s really hard to get away with harming your customers. So, what are the incentives here?

  • Law – Breaking the law is the fastest way to lose your business and end up in jail

  • Profit – Poisoning food isn’t profitable, therefore it’s a pretty dumb thing for a company to do

  • Ethics – Harming people is bad, right?

Well, this example is a bit silly, isn’t it? We need to focus on the real world instead of an imaginary fast food chain operated by psychopaths. Let’s look at Big Tobacco from the same angles:

  • Law – They sold and advertised tobacco products knowing that it’s a deadly poison until it became illegal not to tell about it

  • Profit – Why bother telling your customers about the risks associated with using your product, that would hurt the bottom line

  • Ethics – No such thing, it’s all about money

By using closed source software which is hosted somewhere “in the cloud” we invite companies to act on their darkest incentives because it’s very profitable, and it’s really easy to conceal such activities. Most apps and websites generate small amounts of “poison” every time we interact with them and this poison slowly accumulates in various databases where it can be enhanced and used to harm us later. You see, this poison is personalized. It can’t be used on a random person, it works only on a specific person and using the right poison on the right target is an extremely profitable activity.

By poison, I mean personally identifiable data, of course. Does it seem too far-fetched? Let’s say you’ve got something valuable, and you want to buy a safe. Do you really want this fact to be known to anyone? What if this information can be cross-referenced with your real address? What if we can add some other data points such as your usual time of absence from home and generate a list of targets which would be pretty valuable to various criminals? What about state actors? Even the NSA don’t mind breaking the law and invading privacy of their own citizens. Those guys are angels compared to what less developed nations tend to do with people they don’t like and wish to get rid of.

It doesn’t matter who collects your data. The more data your leak, the more vulnerable you are. Companies use it to influence your consumer choices, politicians use it to influence your votes, state actors use it to deal with the dissidents, criminals need it to steal from you or blackmail you and literally no one is harvesting your data in order to make your life better: there are no incentives to do that.

Self-hosting is the best way to protect your data and to have a transparent infrastructure built by honest people and for honest people.

Distance and Empathy

Let’s say you’re a local baker. You know most of your customers personally. Many of them are friends of yours. We have much more empathy for the real people who we know personally than for the abstract “users”, and it affects our decisions in a big way.

We can’t even keep more than a couple of hundred people in our heads. Human communities don’t scale that well and humans aren’t known for being friendly and empathetic to total strangers. I believe that, contrary to the popular opinion, there is nothing glorious in scale. I’ve met many talented technologists who’re excited about scaling systems and building tools that help them achieve that, and I get it: building scalable and reliable systems is hard and talented people tend to like challenge. Of course, the likes of Google, Amazon or Facebook are scalability marvels but the problem is: they’re also merciless monopolies that aren’t shy of exploiting their scale to crush their rivals. Would it be beneficial for a society to let those companies get even bigger?

The bigger the company, the less it cares about an individual customer, but at the same time, their customers become increasingly more dependent on their new digital overlords. It’s easy to ban a YouTube blogger for any bullshit reason or even for no reason at all. It means nothing for a big company which owns the platform, but it can be devastating for its unfortunate victims.

When I hear “scalable” I tend to add “and also prone to centralization” in my mind. As engineers, we often obsess about technical challenges and tend not to think about the social consequences of our actions. Monopolies are bad for consumers, they don’t care about consumers, and it’s also pretty bad to support them by using their products or by working for them. As for challenges, there is a lot of inherently complex problems which don’t involve scaling monsters to monopolistic sizes. It’s even possible to scale stuff without causing centralization, take a look at Bitcoin if you need an example.

Self-hosting can remove your dependency on big platforms, allowing you not to participate in making big tech even bigger and nastier.

Do Service Providers Deserve Your Trust?

We all have to deal with the fact that the future is uncertain. Even the most trusted companies may decide to prioritize their short-term profits over their long-term reputation. It’s called reputation mining and a low level of transparency allows software companies to slow down their reputation rot, adding more incentives to engage in the behavior which damages their clients but pushes up their profits.

Self-hosting helps you to avoid disappointment when a yet another well known and adored brand becomes evil profit-at-all-costs monstrosity. Don’t make a mistake of trusting commercial companies too much. Even the best ones tend to rot by themselves or with the help of activist investors that can take them over at any time.

Companies Come and Go

I worked for many big companies as well as for several small startups. Most of those big companies are still in business, but they’re filled with ads and deceptive dark patterns. They’ve become user-hostile and their users can’t do anything about it. Most of the startups I’ve worked for have failed, shutting down their servers and abandoning their users. That’s what happens when you rely on a service you don’t actually own.

I really liked Google Inbox and I used it as an email client and task manager. It was a great service and it was also pretty popular. Guess what? Google killed it and tried to force all of its users to switch to Gmail which is a piece of garbage, and it’s full of ads. Can you opt out? Of course not. I’m not ready to self-host my emails, but it taught me a lesson, and I switched to an email service which is transparent and upfront about its business model. It uses encryption to provably minimize the amount of data which is accessible from its servers, which is a nice feature to have. Anyway, this situation was a good reminder that the costs of NOT self-hosting your services can also be pretty high.

Self-hosting allows you to use the software you like for as long as you wish to.

Conclusion

I tend to be optimistic about most of the things. Well, optimisticly-nihilistic, which means I tend not to give a fuck about most of the online and offline dramas but, for some reason, resisting data harvesting and raising awareness of these shady practices seems like a right thing to do.

It’s hard to argue that automation and the division of labor are wrong, and it’s not what I wanted to imply here. There are reasons why many of the services we depend on are centralized and why self-hosting is not for everyone. It has many benefits, but you need to invest some time and money in order to enjoy those benefits.

I see self-hosting as a productive hobby, similar to upgrading and customizing your house. I would argue that our digital spaces are no less important than our physical spaces and taking care of your digital castle and taking time to understand how it works might be both productive and fun. Isn’t it cool to have an autonomous house which has its own water and energy supply? In my view, self-hosting enables us to build self-sustainable digital fortress, what’s not to like about it?