The Ladder is Burning

How Tech Killed Entry-Level Jobs and Economic Mobility

Photo by Lance Grandahl on Unsplash

In the introductory post to What We Lost, I talked about my career. I deliberately spent a lot of time explaining how I stumbled into tech journalism by accident, and the kind of work that allowed me to get a foot in the door.

For those who didn't read it — or for those who don't have the stomach to read nearly 4,500 words about yours truly — my first media job was MakeUseOf, where I was paid $60 (later $120) to write articles explaining trends in tech, or how to use the latest gadgets or websites.

MakeUseOf wasn't the sexiest of publications. The work wasn't the most exciting. The pay wasn't particularly great — although, for a college student living with his parents, it felt like a lot. But it gave me the experience and portfolio that allowed me to move forward with a career in tech media.

I'm incredibly grateful for my time at MakeUseOf, not just because it was instrumental in bringing the opportunities that came later, but because I can't imagine getting such an opportunity today.

Entry-level jobs in media — as with many industries — are rare, and many of those that exist come in the form of unpaid (or woefully-paid) internships that, by definition, exclude everyone that doesn't come from wealth. If you've ever wondered why certain publications in the UK are mostly staffed with people with very specific accents, and very specific names, and a "gap yah, daddy bought me a horse" vibe, that's why.

But there were opportunities for people who didn't fit the usual mould. Opportunities that didn't require you to live in London or New York, or that required the existence of parents that could subsidize your entire life. Not many, but they existed, and they were mostly scattered among the digital world.

An aspiring tech journalist could, with a bit of luck, find a site that'll take a punt on them and pay them a few bucks for news coverage and features. While these sites didn't offer much in terms of mentorship, or editorial oversight, or even a living wage, they were something. I've seen a lot of people turn that something into a career.

Fast-forward to today, and things look especially bleak. Those entry-level opportunities either no longer exist, or are so poorly paid that it's hard to imagine anyone doing them for the cash. These sites barely pay "beer money" for features and news coverage, unless, that is, you're buying your beer in Moldova or Vietnam.

The Decline of Digital Media

I'm not solely going to talk about writers and journalists in this piece, because the issue I described above — the evisceration of entry-level opportunities due to corporate greed and short-sightedness — applies to so many different sectors of the economy.

Independent repair. Graphics design. Software engineering. The list goes on, and on, and on, and I'll talk about them later.

But I'm starting by talking about media and content because it's a sector that I know best, and it's where the decline is most evident.

Sure, journalism has been in crisis since the Internet took over our lives. Faced with dwindling subscriber numbers, and a shift in ad spending from newspapers to tech giants like Meta and Google, publishers had no choice but to downsize. Although larger organizations adapted, the same wasn't true for smaller news outlets — particularly local newspapers.

These local newspapers — the ones that perform the unglamorous, but critical, job of covering school board elections and town council meetings — were the place where many reporters got their start. Today, they're either extinct, or echoing shells of their former selves, operating with a largely skeleton staff.

There are exceptions, of course. Reach Publishing, which dominates UK local media, is a thriving, profitable business — but that's largely come at the expense of readers. The website for my local newspaper, the Liverpool Echo, is virtually unreadable without an adblocker. The content, meanwhile, oscillates between shameless clickbait ("What day Met Office says Liverpool will hit 30C this week") to thinly-veiled advertorials ("Home Bargains shoppers beg 'never get rid of' 99p product").

But that's legacy media. What about digital media? Those companies that were born in the digital era, and thus, don’t carry the costly baggage of printing presses, offices, and pensions?

Well, that side of the industry is fucked, too.

A huge part of the problem is the largely-ignored consolidation of the tech and entertainment digital media ecosystem. These sites are relatively cheap to acquire, especially compared to their reach and status and, as a result, we've seen a flurry of once-independent sites become subsumed by a handful of deep-pocketed companies.

In a relatively short space of time, one company, Valnet, has acquired a dizzying number of titles, including three publications I once freelanced for — Android Police, How-To Geek, and MakeUseOf.

Valnet — a Canadian company whose two founders made their name, and their fortune, in adult media — also owns ScreenRant, The Gamer, XDA Developers, and Polygon, to name but a few sites from its fast-growing portfolio.

If you assumed that the acquisition of a fledgling, independent site by a company with deep pockets would be a good thing for writers, or indeed, for the quality of the content, think again. Valnet is notorious for the abysmal rates it pays, as a cursory glance of the Freelance Writers subreddit shows, with some citing pay as low as $40 for a feature-length (1,500 word) article. One site in its stable, GameRant, reportedly pays just $12 for a news article. Even if you produce these short-form articles at a prodigious rate (and that's saying nothing about their quality), it's hard to imagine anyone making a living from them.

You'd be forgiven for asking whether these rates are even legal, to which the answer is "we'll soon find out."

In February, a former writer for Movie Web, Daniel Quintiliano, filed suit against the company in California Superior Court alleging Valnet violated aspects of the California Labor Code. The violations cited in the suit include a failure to pay the minimum wage or overtime, to provide rest and meal breaks, and to reimburse business expenses.

That suit, which Quintialiano hopes to convert into a class-action, is ongoing as far as I can tell.

For what it's worth, Valnet relies on independent contractors for much of its output. Independent contractors lack the protections of salaried employees, or indeed, the legal entitlements that come with traditional employment. This presents a massive challenge to Quintiliano's lawsuit — although there have been cases where judges have reclassified independent contractors as employees, with Uber in the UK being a good example.

I digress. The problem with consolidation — not just in media, but in any industry — is that it naturally leads to really shitty outcomes for the people working in that industry, and arguably for the consumers of whatever that industry produces. This isn't just my opinion. It's the entire basis behind antitrust law.

On the employment side, it reduces the amount of leverage people have, particularly when it comes to negotiating rates or salaries. If your town has just one employer, it's not like you can threaten to quit if your boss doesn't give you a raise. Where are you going to go? What are you going to do? You're utterly beholden to that employer because you have no other options.

Consolidation can also lead to practices that, I would argue, are tantamount to abuse. Umberto Gonzalez, of The Wrap, wrote an in-depth examination of labor conditions at Valnet, and I highly encourage you to read it. The piece alleges that Valnet has a blacklist of nearly 400 freelancers that are banned from contributing to any of the company's sites for a myriad of violations, ranging from disparaging the company online, to complaining about the site's "abysmal" rates.

I fully expect my name will be added to that list after this post goes live.

One freelancer, according to documents obtained by Gonzalez, was reportedly blacklisted for "[suggesting] Valnet include payment information in job listings." Another met the same fate after posting ScreenRant's rates to Twitter.

I should note that Valnet (and company founder Hassan Youssef) filed suit against The Wrap and Gonzalez after the publication of the piece, claiming it "damaged the reputations of Valnet and Youssef, also a plaintiff, with inaccuracies and malicious intent that caused actual and anticipated financial losses." If you're curious, you can read Valnet's press release announcing the lawsuit here.

It also, according to The Wrap's own coverage of the lawsuit (well, technically two lawsuits), caused Youssef "tremendous emotional distress, affecting (among other things) his mood, temperament and his ability to sleep." Aww.

Valnet has also reportedly — according to The Wrap — issued legal threats towards YouTubers that cited Gonzalez's investigation in their own coverage.

Although I doubt that Valnet will be so moved to issue similar threats to a brand-new Substack newsletter, it nonetheless remains a possibility. It's probably prudent to say that I am, in fact, completely broke and have precisely zero assets. Therefore, any threats of financial devastation will be met with, at best, bored disinterest, and at worst, very public mockery.

For anyone curious about what the latter looks like, I recommend you watch FriendlyJordies's notorious response to a defamation threat from mining magnate Clive Palmer. Or his hour-long response to legal threats from former New South Wales premier John Barilaro. Or the footage of him being served with the papers for the aforementioned lawsuit.

You could argue that at least Valnet is providing an opportunity for people to get their foot in the door — even if that opportunity is, as noted, poorly paid. And while that's possibly true, I'd have to question how useful these articles are as portfolio pieces.

If you're being paid peanuts per article, you're unlikely to do much original reporting. Your ability to pay your rent is governed by the volume and velocity you produce content, and not by whether they’re actually good. Your research will be superficial. You won't bring in many — if any — outside sources to provide comment. You won’t file any FOIA requests, or dig through company records to form the basis of your story.

In fact, you’re more than likely to be re-writing the work of other journalists — a practice that’s known as “aggregation,” because it sounds a lot nicer than “copying.” And, to be clear, I don’t think aggregation is bad per se. It’s inevitable. If publications only printed their own original work, they wouldn’t publish much at all.

The problem, at least for the writers working at these sites and under these conditions, is that they’re only doing aggregation. And the work itself kinda sucks. It’s thinly-written, clickbait-y, and often written with search engines in mind, rather than actual human beings. You know exactly what I mean. Articles where the title is a question with a simple, definitive answer, and instead of answering the question in the first few paragraphs, the writers drag you through several paragraphs of exposition before giving you an answer.

That style of writing pisses off human beings, but for search engines — which often treat length as the main indicator of quality — it’s exactly what they’re looking for.

To be clear, I’m not trying to shit on any writer that works for any of these publications — or does this kind of work for a living. Nothing I’ve written should be considered as a personal attack on anyone, or a judgement about their ability. In fact, I’d argue that there are hundreds of talented writers and journalists working for Valnet — and other companies like it. The problem is that the nature of the work, and the pay and conditions they get, don’t allow these writers to demonstrate their abilities.

And so, it makes life a lot harder for anyone trying to find their next job. Ideally one that pays a salary, and a living wage, and provides health benefits, and where you can’t be blacklisted at a whim.

While the consolidation of the digital media industry is bad, it wouldn’t be fair to single it out for blame. I’m deeply concerned about the rise of generative AI, in part because it’s being used to produce the kinds of work that gave me my start in journalism. The unglamorous, poorly-paid job of writing explainers, listicles, and how-to guides.

To be clear, I don’t think that generative AI can do this type of writing well — or, for that matter, any kind of writing. The hallucination issue isn’t getting better — and it won’t get better, because the large language models (LLMs) that power ChatGPT, Gemini, and Claude don’t actually know anything.

The short-sighted companies that have tried to use generative AI to write this kind of content have, invariably, fallen on their face. Embarrassingly so. Like when MSN recommended a food bank as one of the must-visit culinary experiences of Ottawa. Or when GO Media — the company that bought the remnants of Gawker Media after its bankruptcy — wrote a chronological list of Star Wars movies that was not, in fact, chronological.

It’s not just the fact that AI-generated content is unreliable. It’s just, on a purely aesthetic basis, crap. Have a look at the above Star Wars listicle and tell me what strikes you first? Is it the repetitive style? The general vagueness of the summaries for each film that don’t actually tell you anything?

This is slop of the worst kind. It’s bad. And people hate it. People can tell when something is AI-generated, in part because it has a certain “smell.” Much as an AI model will try to impersonate a human writing style, it simply can’t, in part because these models aren’t human.

But does that matter? No, because no matter how terrible this AI generated content is, it has one advantage over human writers — namely, cost. Why would you pay someone $40 — or less — to write something when you can pay $25-a-month for a ChatGPT subscription, which can churn out content far faster than any human ever could?

No amount of public shaming or humiliation will make these idiot media executives — or the imbecile private equity ghouls that acquire publications — reconsider their use of AI, in part because they don’t care about the actual product, but rather the unit costs. They’d rather spend $0.50 to produce a dogshit, factually incorrect, bland, generic listicle than pay a human being to make something good.

In fairness, people hate SEO-centric content, whether it’s created by a person or an algorithm. And so, I recognize that this argument — that our SEO slop should be made by humans, not machines — is going to be a hard sell to most people, who are fatigued with their search results being dominated by articles and pages that exist purely to rank highly in the algorithm.

I’d ask people to put that visceral, instinctive dislike aside and think not about the form of that content, but rather its function.

These articles are where people get started in their media careers. They’re a stepping stone for better things, and they illustrate to other employers that you can, in fact, write well. For those laid off in newsroom cutbacks, they can act as “survival jobs” — a bit like how someone might get a job at a coffee shop while they look for their next role in the field they’ve trained for.

They were, in essence, the first rung on the media career ladder. And that rung is, right now, on fire.

The Ladder Is Burning

The title of this newsletter was inspired by a video from Louis Rossmann — a vocal right-to-repair advocate and YouTuber who runs his own electronics repair shop in Texas — where he reflected upon his own career.

Before Rossmann became who he is today — a man with his own business that employs several people, a man with 2.2 million subscribers on YouTube, and a man that is actively shaping right-to-repair legislation across state legislatures throughout the US — he was just a guy trying to make a living.

Electronics repair was, at the time, an incredibly accessible field. Someone like Rossmann wanting to do common jobs — like screen repairs and battery replacements — could simply buy the parts and start advertising their services to the public.

But, as Rossmann points out, that pathway no longer exists. Companies like Apple started to exert a tight grip on their supply chains, making it difficult for independent repair shops to access the parts they need without signing up to the vendor’s own program — which, in turn, imposes restrictions on the kinds of repairs that can be performed, and often requires the repair shop to accept random audits, sign NDAs, and to share customer data with the vendor.

Hardware manufacturers — with Apple, again, being the best example — started serializing components. In essence, this means that they created a software-based link between a device’s parts (like the screen, camera module, and battery) and its motherboard. As a result, you could no longer just, for example, switch out a broken screen on an iPhone with a working one from a donor phone. To get it to work properly, you’d need Apple’s own software tools.

Device manufacturers also began joining components that were once separate, individual things with other components, making repairs vastly more expensive. Apple, for example, started integrating the keyboards on its laptops into the entire top case, which also included the battery.

And so, if you had the misfortune of owning a MacBook Pro with a butterfly keyboard that started to malfunction, you’d also have to replace the battery, the trackpad, the palm rest, and the Touch Bar to fix it.

Not only does it make these common repairs far more expensive for owners, but it also makes them harder for independent repair shops to perform, as they have to access parts that are more expensive, and arguably easier for the vendor to control the supply of.

Even if you don’t care about right-to-repair — though you should — these changes should concern you, as they essentially make it impossible for someone to just enter the industry and start repairing devices. You can’t, as Rossmann did, just buy the parts and wait for the work to trickle in.

I should note that Apple has started to walk back some of its more aggressive anti-repair decisions, particularly when it comes to component serialization. That reversal is, however, too late for the countless repair shops that have closed over the past decade.

It’s also incomplete. Although Apple will now sell you a screen, or a battery, to fix your own devices, it still maintains a tight control over the supply of the various board-level components — the myriad chips that sit on your phone’s logic board and your MacBook’s motherboard.

This means that if a $1 chip on your laptop burns out, you can’t just pay someone to replace that chip (unless they happen to have a donor laptop they can extract the chip from). You have to replace the entire board, which can cost hundreds — if not thousands — of dollars.

And you still can’t upgrade the RAM or storage on your Mac. That one stings personally, because I made some decent beer money at university by performing upgrades for the people in my student halls.

To be clear, Apple — or any other device manufacturer that’s clamped down on independent repair — didn’t need to do any of these things. The argument that restricting independent repair is necessary to protect consumer safety has always rung hollow, especially when you consider the multiple examples of misconduct by those performing repairs on behalf of vendors, and the fact that there’s little-to-no risk from a screen replacement gone wrong.

Or, for that matter, the absolutely absurd arguments made by the electronics industry when lobbying against independent repair. Like how one person suggested that independent repair shops might surreptitiously install TikTok on their customer’s devices for… reasons?

It’s about greed and control. That’s it. Apple — and other tech companies — make more money when you buy a new logic board than when you buy a $1 chip. They make more money when you buy a new device rather than repair the one you already own. They’d prefer you pay them to fix your devices, rather than someone else.

Apple — a company that spent $100 billion on share buybacks earlier this year — would quite happily destroy an entire industry, and an entire pathway into the middle class, because apparently it needs the money more than some guy working out of his apartment.

The disturbing thing is that this isn’t just a consumer electronics problem. We’ve seen the same kind of anti-repair actions from other segments of the economy, from automotives to white goods, and especially farm equipment.

For a long time, a job working as a car mechanic was a sure-fire route into the middle class, promising a dignified life and economic stability. The rise of EVs — and, especially, EV manufacturers that act like technology companies — has put that entire pathway into question. You can’t just set up a business fixing Teslas, living and dying based on your reputation, as you would have done in the past. A mechanic can’t salvage parts from scrap vehicles.

Automotive repair, like consumer electronics repair, is increasingly controlled by a handful of totalitarian gatekeepers that not merely limit what kinds of repairs can be done, but who can do them, and who has access to the spare parts supply chain.

The effects of this will, obviously, trickle all the way through the career ladder, affecting those who are established automotive mechanics, as well as those who are just getting out, and I’d argue it’s the latter that are most impacted. It makes it harder — if not impossible — for those who can’t find a salaried position to start their own business, or their own automotive repair side-hustle.

The fact that vendors are trying to lock in mechanics by limiting access to the tools, parts, and diagnostic software mechanics means that you can’t, as in times past, have someone who can fix anything. Ford. Volkswagen. Saturn (remember them?).

People’s careers will be inextricably linked to the fortunes of whatever manufacturer they’ve hitched their cart to. Which sucks if the company ultimately goes bust — as has been the case with countless EV makers to date, like Fisker and Lordstown — or loses popularity, as is the case with Tesla right now.

Again, there’s no reason why Tesla — or any other automotive manufacturer — couldn’t open its tools and parts supply chain, or allow customers to perform repairs that it itself doesn’t recommend (like using salvaged parts from cars that have been written-off). It could make the case that caveat emptor — or, buyer beware. An uncertified mechanic might not be as good as one that we’ve personally certified, but we respect your right to make your own decisions, and to take risks that you are comfortable with.

When you consider that most countries — except, weirdly, many US states — have roadworthiness checks for vehicles, and cars that are in a state of disrepair are banned from the road until they’re fixed, the whole “driver safety” argument that Tesla (and others) have made feels especially hollow.

No, it’s greed. Plain and simple. Like Apple, the automotive industry is content to restrict independent repair — and, by extension, the independent repair career ladder — to make more money.

When Good Enough Is Actually Bad

I’ve barely touched on AI in this article, but it can’t be ignored, as it’s something that’s directly impacting entry-level jobs in the creative and knowledge industries. And not, as I mentioned, because AI is good at things like coding and graphics design. Rather, because it’s cheap, and even though its outputs are objectively and aesthetically terrible, it’s “good enough” for the besuited management class that tends to reduce everything to numbers on a spreadsheet.

Graphic design is especially interesting because, in many ways, people enter the industry in a way that’s similar to media and journalism. Before someone gets a salaried job — one with relative security, health benefits, and a stable monthly wage — they’ll build a portfolio, doing one-off jobs for clients.

These one-off jobs may not pay well. They often aren’t predictable, or a reliable source of steady income. But, crucially, they provide newcomers to the field with portfolio pieces they can show to other potential clients, and other potential employers.

The problem is that people who, at one point, would have paid someone $50 for a design are now turning to AI tools, which can deliver a serviceable image at a fraction of the price. I say “serviceable,” because I want to stress that these models still suck — they’re just good enough for people who care more about saving $50, or whatever.

These AI models still can’t properly generate text in images, with words appearing as fragments of random letters — or things that kinda look like letters, but aren’t. They still struggle with drawing hands and fingers. No matter how much OpenAI and Google inch towards photorealism, their models still produce images that look and feel artificial and weird.

I subscribe to the ArtistHate subreddit — which, despite what the name suggests, isn’t about hating artists, but rather a community of creatives that are lamenting the rise of AI, and how people with actual talent are being displaced by soulless, artless algorithms. There are no shortage of stories of creatives talking about how the market for freelance work has dried up, or how they’ve lost long-term clients to AI models — and, often, been asked to come back and fix the outputs of these models.

Probably the most heartbreaking thing you’ll see in that subreddit is the number of people that have concluded that AI has essentially made it impossible for them to have a career in art and design, with some experiencing mental health crises as a result.

Software engineering is a similarly dismal story — although, I don’t believe it’s possible to blame AI entirely for the decline.

It’s true that entry-level jobs in tech are more elusive than they’ve ever been, and the pandemic-era sugar rush — when tech companies hired in vast numbers, and offered huge salaries to attract the best talent — has since been followed by a massive slump in recruitment, as well as wave after wave of layoffs.

Blaming AI is a useful tool for people like Mark Zuckerberg, and Sam Altman, and Dario Amodei, because these ghouls are actually building large language models, and therefore, they benefit from perpetuating the myth that generative AI has the potential to replace human software engineers.

These predictions are not useful, nor are they grounded in reality. They are marketing, pure and simple, and the media is doing a disservice to the public by uncritically repeating them. Especially when you consider the fact that there’s plenty of evidence to the contrary, showing that large language models are pretty bad at coding.

The Snyk 2023 AI-Generated Code Security Report, which surveyed software developers about their experiences with using AI tools at work, showed that 91.6% of respondents reported having received insecure code suggestions, with 56.4% saying it was a common occurrence. Engineers don’t trust the outputs of these tools — and it’s debatable whether they actually improve performance, or whether the quality of the code is any good.

This has been a pretty heavy piece so far. Let’s have some levity. So, Microsoft has launched a Github CoPilot coding agent, where people can ask for changes or additions by — in simplified terms — making comments on a kind of forum thread.

Microsoft — a company that’s desperately trying to position itself as the leader in generative AI — has been using this product on its open-source projects, which means that the interactions between Microsoft employees and Github Copilot are visible for all to see. And you can watch these human engineers steadily lose their patience — and sanity — as the AI gets things wrong.

This is literally the best that Microsoft has. It’s a product trained on the combined contributions of millions of GitHub users. And it still sucks.

I genuinely don’t believe that AI is taking jobs in tech in any meaningful sense — nor do I believe that it will. I do, however, think it’s excellent cover for practices that the wider public might find objectionable.

Like, say, outsourcing. American coders are expensive. Engineers in other countries — and, especially, India — are much cheaper. If AI was genuinely taking over software engineering, it wouldn’t make sense that Cognizant — one of the big five Indian outsourcing companies — saw its revenue grow 7.46% in the last quarter. Or that Infosys — another “big five” Indian outsourcing firm — grew by 3.61% in the last quarter.

Look at IBM, perhaps the poster child for tech industry outsourcing. See how many roles are available in the US, and compare them to the number in Bangalore.

People have a visceral, negative reaction to when jobs go overseas. The hollowing-out of America’s industrial heartland was one factor — though not the sole factor — in delivering Donald Trump’s victories in the 2016 and 2020 presidential elections. It foments a palpable rage as lives are disrupted and communities hollowed out for the benefit of shareholders.

It’s striking to see how there’s no such rage for AI. When Dario Amodei predicts that AI will replace 50% of the entry-level workforce in the imminent future, there are no protests outside of Anthropic’s offices, or effigies of Patagonia-wearing engineers burned. We just accept it, because we either believe it to be untrue (smart), or because we’ve come to accept the inevitability of automation.

It’s bollocks, of course, but it does provide cover for the things that the technology industry actually does that would make people upset.

And I think we’re going to see more of it. Companies like Meta and Microsoft are coming to realize that the era of exponential growth is over. That’s a problem, because these companies are publicly-traded, and their CEOs are compelled to maximise shareholder value. Their share prices are set with the expectation that these stocks are growth stocks and will only continue getting more valuable.

How do you keep a stock ticking up when there are no more customers to acquire, or indeed, customers are fleeing your increasingly-degraded products, as is the case with Meta?

Well, you fire people and you cut back on hiring. Whenever someone asks why, you say that you’re trying to streamline your processes, or — as is the case with Meta — lie and say that the people being fired are “low performers.” That itself is an utterly reprehensible thing to do, as it makes it harder for those fired employees to find work in what’s already a tight job market.

Whatever you do, you don’t tell the truth. You don’t say that you’ve run out of ideas for new products, and that people hate the products you already have, or are totally disinterested in the products you’re working on — be they AI assistants or the metaverse or whatever.

It’s genuinely frightening to see how bad things have gotten for CompSci graduates. We went from a time when someone with a good resume, a few hobby projects on GitHub, and maybe an internship somewhere could easily land a new role right out of college. Now, they’re adrift.

And, again, it’s not because of AI. It’s because the only thing that these companies care about, and the people running them care about, is money and pleasing shareholders. They don’t care about building and retaining talent, or making things that people want. They don’t see graduates as the future of their workforce, worthy of mentoring and support, because they’ll be the people who write the next killer app or product feature.

Things Are Bleak

I can’t stress how much this terrifies me. On a basic level, I don’t think it’s good that there’s an entire generation that have no career prospects — or are resigned to accept the fact that they have no hope for a job in whatever it is that they’ve trained in, be that design or software engineering, automotive repair or journalism.

It’s not just that the first rungs of the career ladder are increasingly broken — or don’t exist. There are also, as Rossmann pointed out in his video, the first rungs on the economic ladder. They’re the first steps you climb when you want to buy a house, or move out of your childhood bedroom, or get married.

We’re witnessing the creation of yet another lost generation — and perhaps the third in two decades, with the first coming in the wake of the 2007 financial crisis, and the second coming of age during the pandemic. And whereas we could hope that the economy would recover after the late 2000s meltdown of the economic system, and we could hope that eventually Covid-19 would recede and life would return to normal, there’s no such hope here.

And that’s because what we’re witnessing isn’t the product of a biological phenomenon, or a handful of bad choices at banks that were “too big to fail,” but rather corporate avarice and short-sightedness on steroids. We’re witnessing the resurgence of a profoundly anti-social strain of thinking that crested during the time of Reagan and Jack Welch, where it’s okay to shred the aspirations of the young, and the fabric of society, if it delivers a certain financial goal.

Be that a spike in a share price — as what happened after Microsoft laid off 6,000 people earlier this year — or, as in the case of Sam Altman and Dario Amodei, the perpetuation of a belief that generative AI is better than it actually is, and it’s inevitable.

The problem is that these actions have consequences. The 2007 financial crisis led to stubbornly-high levels of youth unemployment across Europe — and, I’d argue, contributed to a resurgence in the kind of ugly, far-right nationalism that we haven’t seen since the 1940s.

When people lose hope, they go to dark places. Even if their anger isn’t directed to an outside group — like immigrants or ethnic minorities — it can be channelled internally, as we saw with the rise of mental ill-health during the financial crash and the pandemic.

What sets this crisis apart from the others is that it promises to transform society in a way that’s aesthetically ugly, and that strips people of autonomy.

I don’t want to live in a world where AI models write the stuff I read, or the art I look at. I don’t want to listen to AI music, or watch AI films. I want people to turn their passions into careers — be they writers, artists, software engineers, or mechanics. I want people to be able to start their own businesses without first having to get approval from some giant multinational tech company, like Tesla or Apple.

I want beauty. I want freedom. And I want hope!

These are the things that we’ve lost. And it’s a loss that, at least generationally, isn’t being distributed evenly. It’s the younger generation — those with the most to gain, and the most to prove — that are suffering the most.

I don’t know how this ends. I think that generative AI will eventually collapse. OpenAI will die. Anthropic will die. Meta will quietly back away from its AI offerings before pivoting to some other shit that nobody wants — just as it did previously with the metaverse or crypto. People are waking up to the fact that these products aren’t particularly good, or useful, and investors are questioning whether they’ll ever make money from them.

But, as I have argued throughout this, I don’t believe that AI was really taking people’s jobs — except, perhaps, in places like art and written content.

The death of AI doesn’t change the underlying cause behind the vanishing of entry-level opportunities, which is primarily an insatiable desire for money and control from a handful of large corporate entities.

I don’t know how this gets better. And that’s what scares me most of all.