31 October 2025 – Baldur Bjarnason

My tech work has proven highly volatile this year. I had decent business in the first half because of the then incoming accessibility regulations in the EU, but those gigs have dropped off in the latter half.

Meanwhile web developer tools based on Large Language Models (LLMs) have, seemingly, stalled the shift away from React that had been gathering steam. LLMs can’t help you with the genuinely new and innovative: frameworks, libraries, and standards that have only recently been introduced don’t have enough critical mass in the training data, which leads to a shift back to React, which in turn prevents the innovations from ever attaining that critical mass.

(If you don’t know React, it’s the reason why whatever web app you use sucks on mobile and has odd navigation bugs on desktop.)

The reliance of “AI” tools has turned an industry that used to be obsessed with novelty to one that’s obsessed with finding novel ways to shit out outdated legacy tech.

This means I’ve been having to think about how to make up for that lost business elsewhere. Less work means more free time, which in theory means I can expand what I do as a writing or media business beyond what I’m already doing with this newsletter.

Thinking about my existing audience #

This blog (or newsletter in the modern vernacular) has been around since 2008 and even then it was a direct continuation of a sequence of earlier blogs I’d been writing since around 2000 or so, back when these things were just regularly-updated personal websites. Those blogs were in turn continuations of the online writing I had been doing regularly in the late nineties.

Most of my existing audience therefore has arrived here by way of reading commentary that’s written from a personal perspective, usually from the perspective of an expert, but with that bit of hoity-toity, overwrought, over-researched style that’s symptomatic of the extensive damage caused by years in publishing and academia. Combine that with a fondness for reading over-complicated and self-important texts and you have a recipe for, well, this.

This has meant that this audience – you lot – are almost certainly a fairly diverse lot.

At my best guess, this audience is a mix of people who have to work with tech and aren’t entirely happy about it and people who work in tech and aren’t entirely happy about it. The Mastodon crowd that follows me leans more free/open-source and activist, with a decent representation of zealotry. The Bluesky crowd leans more towards media and publishing types, with a decent representation of tribalism. The makeup of the RSS crowd, which is at least as as big an audience as the direct newsletter subscribers, judging from the analytics and responses, is anybody’s guess, but by virtue of it being the RSS crowd, they’re likely to lean towards being in tech.

All of these audiences share, I think, the common problem of suffering from the tech industry’s obsession du jour.

(I almost used the term idée fixe instead of obsession, because I’m definitely that kind of person and matching up two stolen French idioms in an English-language text is the sort of thing that makes me laugh, before realising that is also exactly the kind of bullshit “intelligentsia” affectation that’s fundamentally why my audience is, at a rough estimate, ten times smaller than that of most of the others who are writing for a similar target audience. I thought of using idéfix instead to turn it into a cheeky Astérix reference, which also makes me laugh, but that kind of mixing of snob and anti-snob just leads to explosions and loud noises, and most of you honestly wouldn’t have gotten the joke, oblique as it is.)

Some of the service this newsletter provides lies in helping you think about the problems you’re facing. These tech impositions are disrupting our lives and work and we all need to process them. Reading about other people’s informed thoughts and experiences with this processing helps ours, and it can provide catharsis.

Helping you think about a problem is meaningfully different from helping you solve it, which is a fundamental flaw for any media activity that needs to pay for itself.

Having a newsletter that focuses on expert commentary has its advantages. The audience seems to be generally tolerant to variations in emphasis and focus.

But it’s not a great business. Commentary – helping people think about problems – does not have the same kind of distinct and easily measurable value in and of itself that is provided by concrete problem-solving. The value of commentary lies in changes in, or reinforcement of, the perspective of the reader.

This has value, but isn’t worth a lot. For it to be a business, you generally need scale and scale requires a broader appeal than writing, in my style, with my approaches, is likely to ever have.

Commentary also runs the constant risk of pissing people off. Spouting opinions that aren’t immediately grounded in practical and concrete problem-solving will generally always annoy somebody and those somebodies will send emails and mention you on social media. Fortunately, most of the time that somebody isn’t one of your actual subscribers, but the possibility that it might be and that they might be in numbers is a thought that constantly sits at the back of your mind as you write, which itself isn’t conducive to interesting writing.

Beyond the commentary, in the interest of developing a sustainable business, I’ve occasionally committed bouts of writing that’s more explicitly geared towards solving problems – solving people’s problems is generally a better business for both the creator and the audience than commentary – but the long history of this newsletter and the ongoing volatility of the web dev industry has made that tough for me to maintain.

Every time I’ve considered shifting my focus to practical web dev writing – the kind of training and education work that used to be a part of my freelance portfolio before the rise of “AI” wiped it out – something has happened to make me doubt the future of web dev, specifically, as a sustainable industry.

I don’t think it looks healthy.

For example, the entire npm ecosystem looks shaky. The platform has demonstrated clear vulnerabilities and I think there’s reason to worry about the state of the open source packages.

This concern isn’t limited to npm. According to the Linux Foundation’s Census III from 2024 (obviously not the final answer on open source, but it’s a summary of a few data points), many of the popular open source packages aren’t just managed by individual developers, but a small number of developers is responsible for large portions of all popular open source code.

Given that the long tail of open source JavaScript packages, the stuff we build the web on, consists mostly of solo developers solving problems at work, the ongoing layoffs in the tech industry directly threaten the sustainability of web development. The complex web apps and web sites that are the norm in the industry are almost entirely built on open source frameworks. Most of our work is a thin layer of incompetence smeared on top of a thick layer of monopoly- and oligopoly-subsidised open source frameworks.

It’s rare for a web-based project at a company to be written directly against the platform provided by the browsers themselves, even though in most cases that would – in my opinion – result in a project that’s substantially more likely to work for the end user.

Development at all major browser vendors – Mozilla’s Firefox, Microsoft’s Edge, Google’s Chrome, Apple’s Safari – are concerning as well. They’re all pushing “AI” features heavily and, from the perspective of an end user, the stability and performance of the browsers I use has either been declining or continues to suffer from long-term problems (Safari).

The future web developer looks to be an overworked coder, working in an understaffed team, using inconsistent bug-producing LLM tools to barely keep up, building on broken and decaying open source software, on a platform that’s being left to rot as their owners turn it into an “AI” slot machine.

The safe thing to do, as a web developer is to lean towards the more robust parts of the web platform (HTML and CSS) and downplay the fragile and complex part (JavaScript), as that’s more likely to result in projects that tolerate these shifts – could even thrive – but I’m not optimistic that this idea is getting enough traction across the software industry.

If I’m going to be staying on a sinking ship, rebuild a business focused on web developer problem solving, it had better be one that’s proportionally rewarding enough for the time it takes for it to sink, and based on how many of my peers have dropped off from the web dev education and training treadmill, my current guess is that it’s unlikely to be rewarding for me financially or emotionally. (Might be great for you, if you’re so inclined. Have at it!)

Either way, none of that affects this newsletter, the audience of which consists only partially of web developers. While web dev commentary might interest – many of you are developers or in developer-adjacent roles – the more practical web developer problem solving, when I’ve tried that in the past, hasn’t really gelled.

So, for this newsletter at least, commentary it is.

The upside is that writing commentary is often a lot of fun, but the downside is that my writing practice itself needs work.

Y’see, I’m fundamentally a lazy person.

Writing when lazy #

I have the unfortunate tendency to publish the first draft of an article to my website and then to the newsletter. I sit down, write, edit a bit as I go, and when I type the period at the end of the last sentence, I scroll through to spot typos and spelling errors before publishing and sending it off.

This is obviously not a great practice.

This means there is a bit of a divergence in the quality and structure of my writing between the books and the newsletter.

Shitty blog writing isn’t that unusual. It’s arguably the norm for writing in the newsletter and blog genre to be rougher and less polished than other genres of writing, but it’s a problem for the kind of writing I lean towards.

My bit of self-mockery above about my writing style is based in fact: I default to opaque references, obscure humour, subtle (read: muddy) arguments, non-sequitur section-to-section transitions, and over-elaborate language. This results in disjointed texts that lack purposeful and informative repetition, callbacks, and emphases and the message I’m trying to get across is lost on many readers.

All too often I’ve had people respond to something I’ve written, clearly interpreting it to mean the opposite of what I intended, or missing a major point I genuinely thought I had been hammering on ad nauseam.

I can also lean quite academic in my thinking.

A silly Nietzsche mention should make it really obvious that in this context there is no such thing as something being inherently virtuous and that, at least for the purposes of this argument, all ethics, virtues, even truths are pragmatic, if fossilised, constructs of society, right? Explaining that would be too much, I think? Everybody’s familiar with how changes in the wider perception of the author construct changes the meaning of the writing associated with that writer through re-framing. I don’t need to mention the points Derrida and Foucault made, this is already too long and everybody’s familiar with Derrida’s take on parergon…

These are all issues that get solved in rewriting. Clarity, disambiguation, simplicity, and emphasis aren’t textual qualities that leak right out of the brain like the amorphous body of a deboned Athena seeping out of Zeus’s ear canal: they are formed and sculpted into the text through editing after the first draft.

The process of writing a book lends itself to this kind of iterative polishing that steadily shifts the writing away from the author’s foibles towards the intended reader – while at the same time maintaining the writer’s voice – but I have a harder time of putting that into practice in the newsletter.

So, I’ve been working on a couple of experiments.

The first one is I’ve been working on a compilation of the “AI”-themed essays I’ve published on the newsletter in the past that treats those essays as first drafts. It’s less of a collection of previously-published essays and more a collection of polished versions of those first drafts.

Turn your weaknesses into strengths, as they say.

Depending on how that experiment goes, I have others in mind I could try, all geared towards improving the overall quality of this newsletter in a sustainable way.

I’m also working on my basic impulse control. You know… Trying not to publish as soon as I’ve finished typing up the first draft. Bit of a novel idea, I know.

But it can’t end there. I have extra time at the moment, so I’m planning on trying a few more things in addition to the main newsletter.

Adventures in publishing #

I’ll spare you the maps, sketches, notes, tables, outlines, and ideas I used to try and figure things out, but the short version is that I’ve shied away in the past from building on many of my advantages and the resources I have access to.

Mostly for what looked like good reasons at the time.

I’m Icelandic, but I’ve spent most of my life living abroad. I’m a couple of years shy of fifty and for about half of my years on this planet I lived in the UK and later Canada. It never made sense for me to build on the contacts and knowledge I have in Iceland because I wasn’t living in Iceland. It’s obviously not productive to build on international interest in Iceland, for example, when you’re working from a pub in Bristol.

Post-COVID, I’ve also downplayed my experience and expertise in publishing because web dev looked like a stronger foundation for a career than, y’know, books. But my web and software development experience has always gone hand-in-hand with publishing.

• My first degree was in comparative literature.
• My PhD was on ebook interactivity.
• Some of my earliest web design gigs were for small publishing companies.
• My specialisation as a web and software developer has centred on publishing tech: ebooks, PDF rendering, annotation, CMSes, typesetting with web tech, building web-based reading apps, and the like.

With Out of the Software Crisis and The Intelligence Illusion I built on what I’d learned while solving complex problems for publishers and publishing-adjacent organisations, but without mentioning the fact that this is where I developed my understanding of Systems-Thinking, problem-solving, and management practices.

I’ve also been involved with publishing directly. I’ve been self-publishing since 2010 and helped others get their books out on occasion over the years and this year I’m helping my sister, Brynhildur Jenný Bjarnadóttir, get her graphic novel out in the Icelandic market..

This is a continuation of my earlier efforts where I’ve worked as a publishing collaboration, helping people get their books out to the market.

I think we can at least break even on this small project, possibly even generate a small return, but the important part is that it’s the first brick in a publishing business and I, and my sister, can build on.

My first instinct has always been to focus on web and software developers as a target market for the simple reason that I know the problems they face and I also know they spend money to solve their problems.

Given that I’m both concerned about the state of software development and, more importantly, the state of the expense budget developers have access to, it makes sense that I take a look at covering publishing and position myself as a sort of publishing collaborator – sort of a publisher, sort of a consultant – that can offer insights from one of the more unusual book markets in the world.

That is, do publishing work and advice but through an Icelandic lens.

Iceland is, proportionally, one of the more active publishing industries in the world that has fostered considerable talent. We have writers, illustrators, editors, and designers that compete on the world stage despite coming from a country with a population smaller than that of Croyden (Staten Island if you’re American). Information about this market is also quite inaccessible to those who don’t know Icelandic. I think attempting to bridge the gap between the Icelandic and English-language publishing industries might be an interesting project to pursue.

I’m not sure what shape it’ll take in the long run, but it’s starting with me publishing a graphic novel, first in Icelandic and, if everything goes well, a short run in English next year.

And, once I sort out the details, it’ll probably mean starting a new newsletter at some point in the coming weeks.

Previous entry

29 October 2025