Skip to content

Technological change is a bitch

The other day, I got yet another job offer to “train AI”. The company had the audacity to list $15 an hour as a “perk” of the job.

I’ve spent much of my 25-year writing career as a freelancer. I was always excited to scoop a story for a major publication, but I couldn’t have survived without the bread-and-butter jobs: web copy, corporate blogging and product descriptions.

Now I can’t survive. I’ve made $3,000 this year, which is only 40% of what the US government considers a poverty level income. I’m living off my credit card while crafting an endless flurry of cover letters as my career prospects grow ever more depressing.

Fifteen dollars an hour to train AI to steal my livelihood?

There were a couple of years just back there where I was doing very well indeed (no, I am not crying poverty now!) scribbling little pieces for stock market pages. $5k a month for three hours mornings Mon to Fri. Trivially easy stuff in fact. That’s now all – all – done by ChatGPT and the like. There aren’t even any jobs advertised in the sector, let alone the pay being like that.

Technological change just is a bitch. Ho Hum.

44 thoughts on “Technological change is a bitch”

  1. That’s why I gave up on translating. It wasn’t worth the paperwork for the tax or electricity for the computer.
    Machines and Romanians were doing the work and that was a dozen years ago.

  2. If I was starting over, I’d seriously consider becoming a plumber. AI isn’t going to replace them any time soon.

  3. People will only pay you to do what they want to pay you to do. Film at ten.

    Where on earth had this modern assumption, expectation, demand, come from that people will pay you to enjoy yourself doing what you’re skilled at? GROW UP! That was thrashed out of me four decades ago in the very first job I had after university. People will only pay you for what value they want to get out of you that is worth more than they are prepared to pay you. What you enjoy, what you are good at, is completely and utterly irrelevant.

    Nobody’s paid me to do software development for years, so I make ends meet by doing IT support. You get paid to do what people are prepared to pay you to do. That’s reality.

    Even then, my last contact was not even that, it was helpdesk, and that was 16 months ago; since then I’ve only had a dozen days Quite. Literally. moving furniture. “Oh, but it’s a computer desk, why are you complaining?” Just because it’s a ****ing computer desk doesn’t make it “IT”, and it ****ing well certainly isn’t software development.

    So, **** it, I’ve been researching how to sell my pension so I can afford to just say **** you to the employment market. If they won’t employ me, then why the **** should I bother to take part?

  4. I’m a skilled electrician and comptant plumber, but my knees and the trade guilds won’t let me do it in return for money.

  5. But I thought you said that the robots weren’t going to take our jobs?

    As jgh points out, jobs may be a cost not a benefit from one side of the transaction but to the employee, a job is everything.

    This is pretty neat. Two Worstall Fallacies debunked by a single blog post.

  6. Eh? I’ve always said robots will take our jobs. That’s the whole point of having robots – to destroy jobs.

    What I then go on to say – rightly – is that we’ll then do some other job. Which is the veryt reason that the robots make us all richer. We’ve now got what the robots are producing and also what we are – instead of as before, only what we were producing.

  7. As I waited in line, I looked at the blond cashier and wondered if she was the type of Republican who views poverty as a moral failing.

    She simply couldn’t resist a bit of casual bitching while ignoring the question of which party has been calling the shots since 2020, or equally pertinently the fact that Oregon has had a Democrat governor since 1987.

  8. The other day, I got yet another job offer to “train AI”. The company had the audacity to list $15 an hour as a “perk” of the job.

    I’ve spent much of my 25-year writing career as a freelancer. I was always excited to scoop a story for a major publication, but I couldn’t have survived without the bread-and-butter jobs: web copy, corporate blogging and product descriptions

    Person with no marketable IT skills is offered an entry level IT job, women and minorities hardest hit.


    Now I can’t survive. I’ve made $3,000 this year, which is only 40% of what the US government considers a poverty level income. I’m living off my credit card while crafting an endless flurry of cover letters as my career prospects grow ever more depressing

    Smart women have husbands, unhappy women have “careers”.

    That part of my brain is ignoring that I’ve been writing about food insecurity to international audiences for 14 years. A lot of that time, my fridge was empty. I guess the difference is that now it’s full. I just wish it were due to my decades of hard work.

    […]

    Capitalism is a rigged system the world over

    Guardianista makes poor life choices, therefore capitalism has failed.

  9. AI doesn’t need training to take her livelihood: it (or something else) already has, assuming she ever had one.

  10. I’m a skilled electrician and comptant plumber, but my knees and the trade guilds won’t let me do it in return for money.
    But not a competent speller?
    You don’t need to be certified to general plumbing work & getting certified so you sign off on electrics is a piece of piss if you actually know electrics.
    I’m a competent plumber, electrician, heating engineer, carpenter, roofer, plasterer (though will never admit to it- fekkin awful job), roofer, kitchen fitter & several other trades including plaster mouldings restorations to museum standards. I’m probably the only person you’ll meet or even hear of knows how to sling Lincrusta ceilings . Never been certified for any of them. Mostly its not necessary. Ones qualification are ones rep & that’s where ones work comes from. Personal recommendations. Never advertised in my life. Ran a company with a mate did refurbs, conversions, repairs & general maintenance. Where the talents come from. You can’t tell people what to do if you can’t do it yourself. Nor quote for things unless you know what’s involved. That never advertised either. Made enough out of it I can sit in the sun worry (but not very seriously) about the prices of coke & hookers (Both cheaper if you buy wholesale)
    I’m certainly more competent than anyone in this hopeless country. I could start up again here if I needed to. I’ve already done a complete house refurb for a mate & directed a very tricky structural repair reattaching a three storey wall to a building it should have been joined to had the locals watching it in disbelief. My problem is not getting involved in working. The only qualifications I have is a handful of O-levels & having passed the Stock Exchange exam. (So no doubt I could teach economics better than Spud. Although that wouldn’t be difficult.)
    So my advice to you would be to buy a pair of kneepads. Try Screwfix. If you want to earn & not afraid of working, you can earn very well. Just don’t expect it to be easy.
    I agree with Tim. As tech destroys jobs new opportunities are created. I have an idea for a new one about once a week. It’s finding people to put them into practice. So the more nonsense employment disappears the more people will be available to make the world a better place. It’s not only resources are infinite, so’s demand.
    I’d say I’m the result of never having been near the university system. Why I’m so opposed to it. It’s the last vestige of the Guild System. Like that did, it instils in its victims a sense of entitlement. That the world owes them a living. Take it from me, it doesn’t.

  11. Being able to weld and fix machinery meant that I have never been out of work for very long. I think that re-training to do something else seems quite daunting but not insurmountable if you are determined enough. Back when I had an acoustic piano I chatted with the piano tuner who had been a truck fitter. He had suffered an aneurism as was lucky not to have died. His doctor told him that he couldn’t truck fit any more so, in middle age, he re-trained to be a piano tuner. I also read about miners who were made redundant in the 1980s and never worked again. It wasn’t their fault that they were permanently unemployed, it was because Thatcha stole their jobs.

    “…wondered if she was the type of Republican who views poverty as a moral failing.”

    No evidence given for this accusation I notice. I would say that the truth is in between. Bad things can happen to anybody, there but for the grace of God go I as the saying goes. But how you play the hand that providence gives you is important as well. So you are responsible to some extent for your situation in life.

  12. Bear in mind this isn’t really AI. It’s just some really smart coders working on ML (Machine Learning) and calling it AI just as they label things SMART and Honest Politicians.

    Wait until it advances far enough and combines with the advances in robotics. We won’t be needed at all. Just a cost with no benefit. Bit like modern immigrants.

    That, imo, is why our leaders are trying to bump us off. So they can be the Eloi. They want to ensure that there are not any Morlocks around. Sadly for us they will continue down this path. Sadly for them they won’t get all of us.

  13. What I then go on to say – rightly – is that we’ll then do some other job.

    Really? Prior performance is no guarantee of future results so it is quite possible that there won’t be replacement activity this time around. Previous tech revolutions have provided their own information on what people will do next (buggy whip makers can see work in the car factories, etc) but this time the new tech isn’t designed to facilitate something new – it is specifically designed to make people redundant.

    Anyone have any inkling yet of what new thing/s people will do to earn a living (trade) once all the current jobs are torched by robots and “AI”?

  14. @Stonyground
    I’ve found that the more things one learns to do the more things one can learn to do. And then you start finding part of one thing fits neatly with another thing. So now you’ve got something nobody else can do.
    The list I gave above is by no means inclusive. In addition to being a stockbroker I’ve operated a women’s fashion boutique in Kensington, run a bar, run an electronic components business, built kit cars (yeah I can weld), run a whorehouse, a nearly endless list & occasionally, very occasionally, worked for other people. I learnt to do page layouts for a magazine publisher in Fleet Street one weekend. I also built the desktop I’m writing this at. So pretty well anything I have to tackle I probably already know half of it before I start.
    Problem with most people is they’re one trick ponies. That gig goes away & they’re lost. I could create my own employment out of my own talents. That’s all I’ve ever done. Find something people want. Work out to provide it for them. Flog it to them. Count the money.

  15. PJF,

    “Previous tech revolutions have provided their own information on what people will do next (buggy whip makers can see work in the car factories, etc) but this time the new tech isn’t designed to facilitate something new – it is specifically designed to make people redundant.”

    Lots of prior revolutions were about saving cost, and effectively, making people redundant. That’s the whole thing about the Ludd’s smashing up stocking frames.

    And there isn’t always a simple thing of people moving from X to Y. There wasn’t a natural evolution of all the people who bred horses, nor the farmworkers that used to pick crops before combine harvesters.

    But people find things to do. Mostly, the new people find a new thing to do. The older people keep doing the thing during its decline.

  16. PJF – Anyone have any inkling yet of what new thing/s people will do to earn a living (trade) once all the current jobs are torched by robots and “AI”?

    AI is very stupid, and it’s not obvious the (yuge) inherent limitations of LLMs can be overcome. I wouldn’t worry too much about being replaced by ChatGPT.

  17. Wonderful. I’m watching an NHK special news reports, and I’m now competing with Sudanese software developers on $2 a day.

  18. A grumpy old boss once told me that there are only two reasons to hire a person: 1) you can make a profit off their labours, or 2) they make your life easier.

  19. Wonderful. I’m watching an NHK special news reports, and I’m now competing with Sudanese software developers on $2 a day.
    I can’t understand why surprised about that. It’s not as if it’s difficult. There were 12 y/o kids teaching themselves to do in their bedrooms. It’s like motor engineering. Sure it’s an an art setting up an 8 cylinder Ferrari F1 car. It’s grunt labour doing services on Ford Fiestas. There was a lot of money on coding when there weren’t many people coding. So entirely predictably, a lot of people learned coding. So there’s a surplus. The demand isn’t there. Could have predicted this 30 years ago.
    That’s how things work. Unless you’ve managed regulatory capture where people are compelled what your selling & you’ve knobbled the competition. Always been such. Everything has its day.
    Like I said, I blame it on getting involved with the universities. They’re falsely selling the concept of a job for life. The only people seem to have a job for life out of that are university bods.

  20. Bis
    “I’ve found that the more things one learns to do the more things one can learn to do.”

    I can’t boast a CV quite as varied as yours but certainly fixing things is transferable. My apprenticeship was on diesel engines and farm equipment. Since then I’ve rebuilt motorcycles for myself, fixed a reed organ and an old digital piano and worked with industrial power tools and computer controlled factory machinery. I fix things around the house, I did a pretty good job of tiling the kitchen and can do basic plumbing. I’m reasonable at making things out of wood.

  21. I can’t understand why surprised about that. It’s not as if it’s difficult. There were 12 y/o kids teaching themselves to do in their bedrooms. It’s like motor engineering. Sure it’s an an art setting up an 8 cylinder Ferrari F1 car. It’s grunt labour doing services on Ford Fiestas. There was a lot of money on coding when there weren’t many people coding. So entirely predictably, a lot of people learned coding. So there’s a surplus. The demand isn’t there. Could have predicted this 30 years ago.

    Your mechanic analogy works for coding as well. There’s the grunt work, and then there’s the art. It rarely takes me long to find a new contract after the previous one ends, but then I have the track record of producing value for money at the art end of the coding business.

  22. Yeah I got one of those Green Jobs™ I oil the snibbets on wind turbines. Bit slippy up here and it’s pretty difficult holding on at the same time as typi

  23. Problem in all these fields, BiW, there’s people thinking they should be paid your sort of money for doing grunt work. If jgh reckons there’s Sudenese doing what he would like to do on $2 a day then that’s what the work’s worth. Learn to do something else.

  24. I’m reckoning the outcome of AI will be the disappearance of all those jobs are really distributed rules based data processing where most of the “work” is the processing nodes communicating with each other (Yep. The meetings!) Which is what 90% of the desk jockey class has been doing. And when they go so will HR due to an absence of HR. And along with them the diversity coordinators.
    No doubt history will regard it as the Great Cull.
    I reckon it because I’ve done those sort of jobs at one time or another. Though I never took them too seriously. Despite all their performers’ airs & graces you could teach a monkey to do them, let alone AI. Like Tim says about financial reporting. And the broker client manager came round to talk to me about portfolios in his shiny suit. I’d have better off with the monkey.

  25. Whenever I’ve read about training A.I. been in the context of quite a skilled task. I wasn’t sure what it meant so with some irony I asked ChatGPT:

    “ In the context of the job offer you received, “trading AI” likely refers to artificial intelligence systems designed to assist or automate the process of trading financial assets such as stocks, commodities, or currencies. Training AI in this context involves teaching the AI models to analyze market data, recognize patterns, and make trading decisions based on that information. This training typically includes tasks like:

    1. **Data Preparation**: Collecting and cleaning historical market data.
    2. **Feature Engineering**: Identifying and creating relevant features from the raw data.
    3. **Model Development**: Designing and training machine learning models using the prepared data.
    4. **Backtesting**: Testing the models on historical data to evaluate their performance.
    5. **Optimization**: Fine-tuning models to improve their accuracy and robustness.
    6. **Deployment**: Implementing the trained models into a live trading environment.

    The job would likely involve working with data science tools and techniques to develop and improve these AI models.”

    You don’t get that job done by a freelance journalist for $15 per hour.

    Maybe it was doing data entry or other support tasks, but whatever it was she wasn’t training A.I.

    And as a technology aside: we went in to a Rewe supermarket this morning and for all my travels in Germany this was the first time we’ve seen handheld self scanners.

  26. Further reflection.

    At that particular moment did it cross the writers mind that she could have applied for a cashiers job rather than demean herself by having to accept taxpayer funded SNAP payments?

    Or is such work too demeaning for Felisa and best left to stuck up blond bitches who she can accuse of being Republicans (with zero justification other than the fact that at least blondie has a work ethic and some degree of self-respect).

  27. It’s not Sudenese doing what I would like to do on $2 a day, it’s Sudenese doing what I am skilled and competent at on $2 a day. Bollocks to what I’d *like* to do, I need to do whatever I’m physically capable of doing that people are prepared to pay me for. I can’t eat “like”.

  28. 15/hr to train the AI should tell that author that their service training the AI isn’t very valuable and that the work product that the AI will then start to do isn’t very valuable either.

    Which should tell the author that they should train up for a more socially valuable job – AI isn’t taking over shovel jobs anytime soon and you’ll get more than 15/hr to do it!

  29. Jgh is right, and btw this is one of the reasons AI won’t steal your job.

    The product can’t be better than the cheerfully incompetent Indian computer jannies who made it, and that’s why everything Microsoft (for example) releases is complete garbage infested with hilariously cyber insecure spyware now.

    Google AI was telling people to leave their dogs in hot cars, and obsessively creating computer generated images of negroes nobody asked for, before the big G shut it down.

    The last thing the industry hyped as much as AI was the Metaverse. Do you see any companies recruiting Chief Metaverse Officers on LinkedIn these days? Nah, you don’t.

  30. That lefty fool Adam Tooze put up an interesting substack about a month ago

    “Roughly half of all occupations that existed in the data set in 2000 experienced job loss over the subsequent two decades, with work for switchboard operators, filing clerks, transcriptionists, machine setters and many others collapsing. … At the same time, a different kind of work flourished: Tasks that required face-to-face contact and interpersonal skills, like make-up artists and manicurists. Skincare and professional make-up jobs grew from about 14,000 at the beginning of the new millennium to nearly 70,000 in 2023. Manicurist jobs grew from less than 30,000 to nearly 150,000. ”

    Essentially his Substack could be done by Perplexity AI. Will Hutton probably gets it to write his articles. Polly would too, if she had heard of AI. Perhaps I should get it to write an article in the style of Spud

  31. OK here is an article in the style of Richard Murphy about the optimum rate of tax

    “The question of the optimum rate of tax is one that has long been debated by economists and policymakers alike. However, much of the mainstream economic thinking on this issue has been flawed, failing to consider the broader societal implications of taxation and the role it plays in shaping a more equitable society.
    At its core, taxation is not merely a means of raising revenue for government expenditure, but a powerful tool for redistributing wealth and addressing economic inequality. The notion that there exists a single “optimal” tax rate that maximizes revenue while minimizing distortions is a narrow and misguided view that ignores the fundamental purpose of taxation in a just society.
    The Fallacy of Revenue Maximization
    The idea of an optimal tax rate that maximizes revenue, as popularized by the Laffer curve, is a dangerous oversimplification. It treats taxation as a purely economic exercise, disregarding the social and ethical dimensions that should be at the heart of any tax policy discussion.
    Proponents of this view argue that beyond a certain tax rate, further increases will discourage economic activity and ultimately lead to a decline in tax revenue. However, this argument fails to consider the positive externalities of a more progressive tax system, such as increased investment in public goods, reduced inequality, and a more stable and cohesive society.”

    It continues for another 200 words but already it is more coherent and cogent than Captain Potato

  32. Part 2

    Richard Murphy argues that the way to reduce inflation is to cut the Bank of England’s base interest rate. His key points are:
    He believes the Bank of England’s policy of raising interest rates to combat inflation is misguided and counterproductive. Higher interest rates are making many people poorer by increasing borrowing costs, which in turn reduces consumer spending and economic activity, creating a recessionary spiral.
    While higher interest rates benefit wealthy savers, they do not spend enough to offset the reduction in spending by those on lower incomes facing higher borrowing costs. This creates a deflationary multiplier effect that Murphy argues is fueling inflation rather than reducing it.
    Instead of rate hikes, Murphy proposes cutting the base rate to around 2%, which would reduce borrowing costs and put more money in people’s pockets to spend, increasing demand without fueling a wage-price inflationary spiral.
    He argues that businesses, banks, and companies in sectors like food and energy are increasing profits excessively, contributing to inflation. Lowering interest rates would put pressure on them to be more competitive on pricing.
    Murphy believes the assumptions and models the Bank uses to forecast inflation impacts are flawed, as evidenced by them having to abandon their forecasting models. He argues their rate hike policy is making the inflation situation worse based on an incorrect understanding.
    In essence, Murphy advocates an alternative approach of expansionary monetary policy through lower interest rates to boost demand and spending power, rather than the contractionary policy of rate hikes currently pursued by the Bank of England.

  33. “Roughly half of all occupations that existed in the data set in 2000 experienced job loss over the subsequent two decades, with work for switchboard operators, filing clerks, transcriptionists, machine setters and many others collapsing. … At the same time, a different kind of work flourished: Tasks that required face-to-face contact and interpersonal skills, like make-up artists and manicurists. Skincare and professional make-up jobs grew from about 14,000 at the beginning of the new millennium to nearly 70,000 in 2023. Manicurist jobs grew from less than 30,000 to nearly 150,000.”

    And I bet that all those people who had to give up reasonably well paid and stable jobs that paid their mortgages are delighted to now work in the semi-gig work environment, where their earnings are variable and probably not much above minimum wage.

    What the ‘all the people in the past managed to find new employment’ brigade are forgetting is we don’t live in the 18th or 19th century anymore. In those days the government of the day was quite happy to send in the army if the natives got restless, and export the troublemakers to the antipodes, if they didn’t hang them.

    Thats not going to happen nowadays is it? If vast swathes of voters in a democracy get put out of work by AI, then within very few years they get to vote on what to do about it. And they aren’t going to vote for the ‘Let the Free Market solve the problem’ party are they?

  34. Dio – Horrifying, Candidly.

    PS – that “training an AI” job:


    Calling all aspiring novelists and poets: Your writing skills are needed to feed the machines. Firms that create training data for AI models are now seeking to hire creative writers, as Rest of World first reported. It’s an opportunity that could one day spell the end of writers’ marketable skills. But in an age where reams of books are being fed to AI without permission, it’s at least a more honest way to get one’s hands on literary expertise.

    The job listings have all the trappings of a gig economy side-hustle. Companies are offering hourly pay to candidates with graduate degrees or professional experience who can also work remotely.

    The biggest AI companies right now
    Scale AI, a San Francisco–based AI training startup, posted the job “AI training for creative writers” in May for those fluent in English, Japanese, and Hindi. Offering an hourly wage of $25 to $50 based on educational attainment and experience, it specifically seeks candidates with master’s degrees and PhDs. Australian firm Appen also posted a job for “creative writing expert” and cast a net for writers with advanced degrees.

    Remotasks, a micro-task job site, has posted a list of AI creative writing training jobs in languages including Xhosa, Slovenian, and Malayalam. Hourly wages vary based on language (Turkish, for example, offers $3.24 to $4.51 per hour, while Catalan offers $15.50).

    Across the job listings, the work appears to be project-based and aimed at gaining human feedback on AI writing. Scale AI and Remotasks even offer the opportunity to write short stories on assigned topics.

  35. I notice that during the inflationary aftermath of the Napoleonic Wars the BoE base rate remained at 4% – all the way to about 1840 when it started zig-zagging almost yearly between 2.5% and 5%.

    Prior to 1820 the base rate had been 5% *forever*. (Well, a blip in the 1690s)

  36. The Meissen Bison

    Diogenes: …«but already it is more coherent and cogent than Captain Potato»

    The acid test of course will be to run the identical AI exercise tomorrow and see if it produces something quite at odds with what it produced today. If AI can do that it should perhaps be taken seriously.

  37. Scale AI, a San Francisco–based AI training startup, posted the job “AI training for creative writers” in May for those fluent in English, Japanese, and Hindi. Offering an hourly wage of $25 to $50 based on educational attainment and experience, it specifically seeks candidates with master’s degrees and PhDs.

    I’d bet Felisia Rogers has a (bullshit) masters or doctorate. She should be biting their hands off, especially if she was happy to accept “$75 for a 1,700-word column” from Salon. I don’t know how long it takes hacks to write a 1,700 word column but it can’t be a good hourly rate.

  38. Jim,

    “And I bet that all those people who had to give up reasonably well paid and stable jobs that paid their mortgages are delighted to now work in the semi-gig work environment, where their earnings are variable and probably not much above minimum wage.”

    The idea that a stable job is the natural order of things is a misplaced one. It’s almost entirely a 20th century thing. It suited massive offices and massive factories.

    “Thats not going to happen nowadays is it? If vast swathes of voters in a democracy get put out of work by AI, then within very few years they get to vote on what to do about it. And they aren’t going to vote for the ‘Let the Free Market solve the problem’ party are they?””

    There is a huge queue of people who want healthcare, almost every straight man would like a Ferrari, and every woman would like clothes made to their custom specifications. This tells us that we still aren’t rich enough. And if we somehow get the AI making our entertainment and driving our taxis, it means a bit more money in our pocket for high performance sports cars and couture.

    Those 150,000 people doing nails is because we cut things like the cost of flying. All those travel agents that took a 30% commission got wiped out, we get better nails.

  39. OK, I get that ‘jobs are a cost’ argument, but for two things:

    1. We live in a society that pays people not just to sit on their arses all day, but subsidises many of them to go out and commit crimes. Does it really cost more to get them to do something even marginally constructive instead of destructive? (i.e. could it end up costing less?)

    2. For most of us, having a job of some sort is how we afford to live. That job may be a cost to the employer, or even society at large, but from the perspective of the employee, the benefits exceed the costs – if they don’t, then you get another job.

  40. 1) If we didn’t subsidise them – or even if marginal tax rates were low enough that even if we did then they’d still work – then that’s solved.#

    2) Sure. But we want that working effort – to gain the income – to be devoted to the things that other people value most among the things that can be done with that effort. So, sure, the income is a benefit to hte worker, the production is a benefit to hte consumer, the job is the cost of each, the income and the production.

  41. “Google AI was telling people to leave their dogs in hot cars, and obsessively creating computer generated images of negroes nobody asked for, before the big G shut it down.”
    The lifecycle is this: An AI is developed; it is shown to the public; It says something that horrifies people that failed to become adults; a politically correct filter is placed on its output; nonsense and hilarity ensues.
    Behind the scenes, behind all the jokes and memes, there are almost certainly AIs that are not nerfed, and those AIs are telling a few people what they need to hear, not what is socially acceptable.

Leave a Reply

Your email address will not be published. Required fields are marked *