The web, as a platform, ecosystem, industry, and market, is in a very odd place.
- The platform is genuinely great. Bugs are inevitable – every non-trivial project will encounter an implementation bug in a major browser at some point – but the platform overall is in a good place with capabilities I wouldn’t have even dared to hope for a decade ago.
- The ecosystem is a disaster, a blighted nightmarish landscape of ill-conceived frameworks and packages, dominated by outdated ideas, bloated implementations, and a complete disconnect in multiple dimensions between outcomes, income, costs, and ROI – all built on broken or decaying systems. TypeScript just outright isn’t a good, sound, or stable programming language. Npm is a disaster waiting to happen and suffers from odd regressions (like search, currently). VS Code is a bloated surveillance system with coding tools bolted on. GitHub is just an LLM-slop injection vector at this point. It’s all fucking owned and controlled by Microsoft. You genuinely can’t trust any of your npm dependencies as any of them might turn on you in some way in a single update, some of them unintentionally through LLM-inspired malpractice.
- The industry is a mess. Layoffs continue. End-user outcomes and software quality in web apps have absolutely no effect on organisation decision-making. Web media is largely dependent on the ad market, which is almost certainly substantially driven by fraud or at least deceptive practices. Web media companies are also being starved of traffic by search engines switching wholesale to fabrication engines.
- The market is crap. Most Software-as-a-Service businesses are crap. They’re all dominated by the same disconnect between decisions and end-user outcomes. Oligopolies rule the roost. Want to start an independent ecommerce store? Shopify or WooCommerce are your only options, both run by people who outright seem irrational, each in their own way. Every product is adding extremely unreliable “AI” features but, what’s worse, is their existing core products are also deteriorating as the layoffs, inattention, and overuse of LLM-generated code take effect.
Already, this is looking quite shit, but it’s actually worse than you think – at least if you’re a regular coder in the industry.
Before the “AI” bubble hit, my work was generally divided into three groups:
- Freelance web dev. Mostly related to web-based reading, ebooks, or annotation and mostly from old contacts among those few who are still in business after the ebook bubble collapsed.
- Freelance training and education work. Much of it as a tech editor or reviewer. But also some writing and course design.
- Writing, such as sales for Out of the Software Crisis, and analysis.
My freelance web dev was always driven by the fact that not that many people have both expertise in ebook standards, reading-related issues, and annotation. as well as web development. I’ve never put much effort into getting other kinds of freelance web dev gigs from new contacts, so this part of my income has steadily been declining over the years.
The training and education work evaporated when the “AI” bubble began to inflate as the companies I had been working with either laid off all of my contacts, switched to “AI” solutions, or both. Ironically I didn’t notice initially because I was busy researching, writing, and publishing The Intelligence Illusion. I attempted to launch a web dev course of my own but that tanked because I was paying more attention to what interested me over what people in the industry needed. Also, my newsletter was entirely geared towards analysis and not web dev, so it wasn’t a great platform for launching anything resembling a web dev course.
Writing and analysis is what has occupied most of my attention over the past couple of years despite it generally not paying that well. Not nothing, but also not much, and definitely not proportional to the time I invested in the newsletter.
Even though the new US administration isn’t likely to allow effective anti-trust action to happen to Google, I still think that the funding environment for web browsers is likely to change.
The rationale for much of the existing funding for developing and maintaining web browsers – one of the more complex pieces of software today – is search engine income. Even if we assume that anti-trust penalties for Google end up only being a slap on the wrist, the “AI” bubble itself threatens that revenue stream.
- If the promises of the “AI” bubble pan out, search will get replaced by chatbots and the motivation to use “AI” income to fund browsers isn’t nearly as strong as that of search.
- If the promises don’t pan out, the fallout from the bubble will be so widespread and catastrophic that it’s likely the funding will peter out anyway.
Bad outcome either way.
The same dynamic applies to the web dev employment market.
One of the overt goals of the “AI” bubble is to replace software developers with automation:
- “Mark Zuckerberg says AI could soon do the work of Meta’s midlevel engineers”
- “AI agents mean Salesforce won’t hire software engineers in 2025, apparently”
That’s why we’re in the middle of the biggest wave of tech industry layoffs since the collapse of the dot-com bubble, despite there being a supersized funding bubble going on at the same time.
And again, we face bad outcomes either way:
- If the promises of automation pan out, at least in web development, “web developer” as a profession will shrink to a fraction of its current size.
- If they don’t pan out, the ensuing bubble pop is likely to be so catastrophic that the surviving companies will have to both lay off even more people and likely won’t have the resources to hire the people needed to “repair” their codebases after years of constant injections of LLM-garbage.
We’re damned if it succeeds; damned if it fails. Software developers are no longer a meaningful part of the upper echelons of the tech industry and no longer share in the upside.
The profession is about to reap the consequences of pretending to be superior to the cleaners and canteen staff, because it turns out employers really want to treat all workers equally badly.
My guess is that “big company web developer” as a target market will effectively dwindle into non-existence over the next few years. The only question is how fast. Whoever remains will be underfunded and underpaid.
Also, “web developer for web media companies”, seems similarly under threat because much of the web media industry is not going to survive the ongoing changes to search or even the mildest reckoning with ad industry dysfunction.
Even if neither happens, those industries are still likely to be precarious going forward. Uncertainty is likely to be constant.
All of which is to say that I’m unconvinced that putting much effort into building up my freelance web dev again will pay off as opportunities are fewer and the competition from tens of thousands of newly laid off devs will be intense.
It’s also the reason why, as Chris Ferdinandi points out, many web dev courses have outright priced themselves out of the market. We probably need different education and training options at different price points and business models.
The web itself has a few low-hanging fruit opportunities available, I think. Wordpress’s ongoing crisis, the ever more obvious dysfunction of using frameworks like next.js for web media sites, and the fact that many of the web’s recent platform additions remain underused means there’s an opportunity to do really interesting things for the web as a medium (as opposed to the web as a software platform). The web is still a major information medium in modern society.
(I really wish more browsers released support for showDirectoryPicker()
. Local-only web apps where the user is just working with their own files, not a cloud whatever, would be a compelling alternative to both cloud-based SaaS web apps and subscription-based native apps.)
Likewise, the web as a software platform has mostly been about building Software-as-a-Service web apps. Local-first or even local-only web apps remain in a minority and are mostly overcomplicated research projects or overcomplicated early-stage startup projects.
I don’t think there are going to be many direct “pay me to work on this” freelance opportunities on either side of the web coin (medium versus platform) but there might be opportunities in helping others work on it, through courses, ebooks, blog posts, videos, and the like.
And, I think the balance has tipped for making things. Before the path of less risk was to build up freelancing, then consulting and education (courses, ebooks, writing, etc.), and finally making and selling software being the most risky, I’m starting to think that it’s now the other way around. Given how bad most of these services are and how expensive the rest have become, making software might be a slightly easier sell than before.
As long as you aren’t selling a subscription-based SaaS. Those are in a lot of trouble, I think.
My options:
- Get more web dev freelancing. Not likely to work that well for reasons outlined above.
- Full time web dev job in Iceland (remote opportunities are now much fewer than they used to be).
- Figure out a way of doing web dev education that’s equitable and fair and works with the current state of the industry.
- Focus on web media in some way. (Vague thoughts, but not as vague as they sound here.)
- Focus on making local-first or local-only web apps that get sold under a more old-fashioned single payment model.
- Get a job outside of web dev and tech, preferably one where the skills I have would still be useful.
The first is unlikely to work, although I wouldn’t say no to gigs if they come up.
The second is not particularly desirable as Icelandic tech seems to be all-in on React.
The third and fourth could be interesting, especially if I find a good way to combine them.
The fifth is very tempting but could be a big time sink.
The sixth is also not particularly desirable because I actually enjoy web dev proper. Although, that might be an instance of letting your passion and interests compromise your career.
We’ll see.