We’re in June 2026, and I’m (still) feeling like the citizen of a dystopian universe. Not that things were ever that great with this world of ours, but lately it’s looking more and more like the string theory explanation of the Mandela effect.
I don’t want to focus too much on the negatives - there are people in the world feeling everything that is happening more acutely than I ever will.
In this stream-of-consciousness post, I’m trying to:
- digest or document a few things;
- figure out some positives in situations that currently concern me or my projects and plans;
- just write something, damn it, without too much fuss and formality, because I miss writing.
EU getting out of US-based clouds
One of the positive trends I noticed is EU finally waking up about the state of its tech stacks. I used to say that the whole world were selling their souls to either Microsoft, Apple, Google, and Meta. A bit dramatic, but recent events showed that there is no such thing as too dramatic and there is no “global” citizenship as long as there are state actors.
I can’t imagine why anyone ever thought it was a good idea to keep their data in US-based clouds. I think many US citizens have had that concern for a while, and in such a context, EU decision-makers couldn’t pull their heads out of their asses to realize that data is valuable, and even though we are allies “today”, you don’t know how the winds of interests (geopolitical and economic) change. Why ever risk a nation’s information by making it freely served intelligence to a potential enemy? You silly non-technical people…
Its a good thing that decision-makers in the private and public sector in the EU are finally waking up and generating demand for EU-based services. Better late than never.
I love to see the growth of the EU tech space. There are new services and apps popping up like mushrooms after the rain, and it is a very good thing. We will have more choice, more competition, all good for a healthy market. Now if only we had the money to use them and support the developing individuals and companies through sponsorship, retainers or some such arrangements, so that they could lower prices a little and make the services affordable enough for individuals and small companies too, without ridiculous limitations.
Hopefully, with more advancement and diversification of the EU-based tech portfolio, we will be able to have a solid European-created open-source ecosystem with tooling for everything we need.
Being online with a business, or anything really, has gotten mighty expensive for solo entrepreneurs and small businesses that don’t have in-house development teams or own infrastructure, and with current tech price trends, it will not get any cheaper. We need true alternatives in the EU, not just for static-generated sites, but also Next.js, Astro server-side rendered applications, à la Vercel, Netlify, Cloudflare Workers.
We also need content managements systems that aren’t reserved for enterprise customers only and with idiotic limitations.
Owning a website can still be cheap…
…if you’re willing to learn new tools, do the work, can compromise on features, and - most importantly - don’t actually need a CMS as such.
Speaking to the idiotic limitations I mentioned earlier, many CMS limit how much content you can have (the fuck?!), or how many languages you can enable for a multi-language site.
Not everyone needs visual site builders (in fact most non-technical people I know get blocked eventually and require developer help even with those), or content personalization, or other such ridiculous things.
Unlimited content, good editorial experience, and some contact forms will do for most small sites. Wagtail handles this well, albeit with some limitations, but they’re working hard to offer excellent editorial experience. Now, if only it could be hosted cheaply in the EU, by non-technical people, and not require a developer for basic content types… But otherwise, it’s very nice, free, great editorial experience.
You’d think Jamstack, headless content management systems (CMS) and static site generators, came to rescue us from WordPress, but even those static sites require hosting somewhere, initial development for a site application with integration to the headless CMS (or some git-based solution), plus the maintenance of one or more applications in tempo with CMS updates AND package updates and security vulnerabilities.
Headless CMS are a nightmare in my book - they add complexity, cost, and you end up with an infra with too many moving pieces. You either need to host them yourself - adding the burden of maintenance - or you overpay for less powerful features and plenty of limitations in a one-size-fits-all managed service, that may or may not support the front-end application development (or give you an off-the-shelf “theme” that you must customize yourself). On top of that, you get content over some API endpoint, whether REST or GraphQL, and get the performance hit (and potential security issues) that communication comes with.
If I were paranoid, I’d say tech hubs around the world conspired to take away the ease of getting a business or blog online for non-technical people and raise the price of services in the industry. But that’s ridiculous. Or is it?
Needless to say, the Jamstack architecture is a good fit for developers and organizations with in-house dev teams, or large pockets to retain external devs and consultants. But it’s not such a breath of fresh air for my 60-year-old neighbor or relative trying to have a website with 5 pages and a contact form for their very small family business, especially since they can barely afford their domain name.
On top of all this, many CMS products are darn expensive these days… Whether in licensing fees, support, customization or plain operation.
Take for example Statamic. While you could run a free instance on your machine as a developer, a non-technical person will be challenged in operating the CMS locally. So you need a server, and a developer keeping things running.
Cloudcannon - basically a fancy UI for editing markdown files (with some very nice bells and whistles to be sure, but not sufficient to be an actual CMS in my book) - costs 55 (fifty-five!) dollars monthly.
PayloadCMS requires development of all collections, blocks, hosting on a server or compatible serverless platform, juggling which databases you use based on whatever quirks your content (or architecture) has and what you can afford.
Wagtail is cute but a pain to deploy and host, needs a developer for ANYTHING other than content production, and going through changes (future maintenance fees?!).
And everything ads one bill after the other. I used to be able to deliver a decent small site in 10-15 hours at 30 dollars an hour. Now it’s painful to even write to a client that license fees or subscriptions have gone up, cause I know they’re struggling, and I end up eating the loss with a cherry on top. It huuuuurts.
There are some on-your-machine solutions - Frontmatter CMS, Publii, and so forth, but either they’re too technical, too opinionated, incomplete (relative to the needs of the client), making it difficult for a site owner to customize the behavior (and branding) of their site. Nonetheless, mad respect for their makers - single devs or small teams that really carry the torch for us small, insignificant individuals and small companies that want to have an online presence beyond friggin’ Facebook and “Insta”.
AI is pushing server (and other tech) prices up, as is the increasing demand for compute and storage throughout the EU. For example, Hetzner, one of the top-notch server providers in the EU, just increased their prices for new contracts by more than 2.5 times, 3 times if you add the backup feature at 20% of the server cost. And you must have server backups… Now, what’s a small business to do?
I can’t thank the gods enough for discovering Obsidian. It’s not “made in Europe” - it’s Canadian - but it works on my machine AND my tablet. Owning a server at home, I can work within our home network and never worry too much about who might be snooping on my brilliant ideas.
I’m working on a collection of plugins (developed by the community, and some new ones by yours truly) to help me manage content for multiple sites in a sensible way. I’m close to figuring out the right “recipes” for content front-matter, such that it allows me a solid, tree-like organization of content (almost Wagtail-style), including translations for the one or two sites I manage that require multilingual content. It’s high-effort work, but it will be worth it in the end.
I was already able to reduce my costs considerably - I now pay for domain names, phone subscription and email service (and losses from clients, FML). I might consider paying for Cloudflare’s image optimization service in the future. It’s good enough to justify it. Maybe.
Everything else runs on either the home server or local machine. This is good. Very good.
Whenever my vault recipes are consolidated and good to go, I might share them with some of my budget-conscious clients, assuming I don’t close my company before that.
I feel increasingly powerless and exhausted having to explain to people I can’t give them what they want, while they refuse to accept what they need, and don’t have the budgets to scale according to their desires. And this is very bad.
Given that there’s a higher risk than ever to operate at a loss overall, not just for isolated clients, this possibility is growing on me. Not sure if this is a good or a bad thing, but I’ll say that I’m used to having money-related stress, but not used to feeling powerless and trapped WHILE ALSO having money-related stress. So far, the cards are stacked clearly in a specific direction.
Proof of content ownership in the age of AI
This has been bugging me quite a bit lately. I’ve been busy figuring out the tech stack behind herbology.eco for the past half year or so, and I’m finally getting there.
I’ll be honest - I would love to have that website fully public and free. But the way AI operates today has me feeling completely and utterly hopeless.
Imagine putting thousands of hours of work into complex and difficult content that requires expertise, only for someone to end up appropriating your content and ideas through the use of AI in their web searches (AI overviews) or directly through generative AI use to create their content.
The content ends up being ingested by various AI agents, and spit out with or without reference links. Even if the links are included, few people verify them. Which is death for a content-based business. No clicks, you can’t monetize anything through subscriptions or anything else.
So what’s the solution? Well, it needs two components:
- a closed community for accessing the content, coupled with marketing to spread the word to the target audience that your content-based product exists. This is pretty bad news for me - I hate doing any marketing except on page SEO, which becomes a bit “meh” since most content will be locked away from Google’s eyes. Perhaps semi-public index pages that show “member-only content” locks for the valuable sections will do.
- a way to PROVE that I, MYSELF, wrote the content, rather than AI-generated, the content, time-stamping a document’s history in an immutable way. I have a feeling that most people will appreciate quality, human-made content. And, ooooh boy, did I find the goodies for this! It’s called Semoi and it comes with a very handy Obsidian plugin for capturing proof of writing . After you install it, it keeps an eye on your typing for each note in sessions. When you’re done writing, you “mint” the content. It will show how long you actively type (approximation) and how many keystrokes you performed. An AI-generated bit will show many keystrokes (say 200) for under 10 seconds or some such numbers.
There are downsides to these approaches.
One is that now, what should have been a content site, becomes an online community, with all that entails. Fewer visitors, probably fewer clients, less income opportunities. But perhaps this is the only way to survive as a content-based business until politicians figure out how to keep AI in check.
How do you manage an online community, while not using subscription-based services or keeping a database? … Pfft. I still need to solve this. But I’m confident the solution is right in front of my nose. Perhaps I just need to choose to trust at least one other company beside the makers of Obsidian. Cloudflare? Not EU-based, but at least I can set the jurisdiction on buckets, D1 databases, and durable objects. And, other than not allowing me to buy or transfer .eco domains to the Cloudflare Registrar, this company has yet to disappoint me (throwing salt over my shoulder that I didn’t just jinx it).
As for Semoi, since most people can barely type 80 words a minute, having too few keystrokes for a long piece of content will rat out an AI-creator. Of course, that also means, that, ideally, I will use this solution only for completely new content. Sure, you could trick it by copying key by key AI output, but I mean, how messed up is that? Either way, in such a case, at least some amount of effort would be put into it, lolz.
For now, this is my last post without Semoi proof-of-writing certificates - hopefully it’s clear enough that I wrote it myself ;)
Rounding up with gratefulness
I’m grateful that I forced my left side of the brain to be good at more than language learning and processing, and that I’m able to more-or-less build my tech stack. A custom solution these days, with maintenance, can get over 4000 - 7000 Euros annually in infrastructure and maintenance costs, not to mention custom development.
While I know some will say you can develop much with AI, allow me to be skeptical - for complex application the investment in AI tokens will be too high, and the quality does not necessarily follow, absent human supervision.
AI is like Trump - completely removed from both humanity and reality. It just doesn’t get it, confabulates, and it leaves us all with debts of some kinds, while making us darn st00pid.
There is a time and place for it (translations are better with AI than without, grammar checks with a locally-running AI are better than without and cheaper than a subscription-based service too). But as a content-generating tool, I’d rather not. As an architect, I’d rather not. As a code reviewer - okay. Everything else needs just as much work to verify the output as it does to do the work myself and then ask it to review.
Nonetheless, in addition to being grateful that I learned how to build my own website applications and weave the necessary infrastructure, I am also very grateful for our second-hand, AI-capable, home server my better half purchased. We can host our very own private models, and our websites, no fuss, no hundreds of big money flying into corporate pockets. I married a visionary!
Last updated: June 22, 2026