AI's impact on employment is less directly replacing a human worker one to one, but suppressing hiring demand by providing potential employers with a plausible alternative to increasing headcount. This is disproportionately impacting early entry hiring as the ramp up time to productivity for inexperienced workers is today too long for companies to willingly invest for.
Never depend on a megacorp shit job for stability.
The way out is unionizing and workers coming together to form a worker-owned co-op business with other good people, a very low turnover rate, and a civilized environment. They don't need a megacorp and can find markets themselves.
I’m not making an argument, I’m offering a thought experiment: “what if workers are no longer needed at all, so that they have nothing to threaten to withhold?” It’s a bleak possibility.
Then you have masses of idle, desperate, poor, purposeless people who either find coping hobbies to fill their time while some inevitably revolt.
Unlike the fallacy of the "lump of labor fallacy", there is always a limited number of job vacancies and limited number of employees rather than infinite open jobs just waiting around to hire everyone. (The pie is always a countable, finite number but never static, it can only grow or shrink within the means of finite-sized, finite businesses affording finite workers; the fallacy is believing in an infinite, unbounded pie that just magically grows because people exist.) Demand will/is rapidly shrink(ing) in many fields for certain roles, and the remaining ones maybe paid less and be more competitive to the point of extreme desperation. There are too many people left without a means to achieve fulfilling or a comfortable life because a few have reordered society to give them far more than they need and have taken away their livelihoods; history shows this rarely ends well without constructive intervention.
It's interesting how stable the numbers are. The GPT-3 to 4 leap made a dent but it was followed by the pre-training plateau. I would have expected the huge leap with agentic AI since late 2025/early 2026 to be reflected in these numbers.
software engineering was one of the best ways for working- and middle-class kids to make a good living. I suspect it held a fairly large % of "good jobs available to kids without rich parents" (far larger than its % of overall jobs).
Unattractive nerdy kids being able to escape generational poverty thanks to programming skills is what broke law, business and politics caste members minds. It is unfathomable that somebody could be paid well for a job when they didn't even do Erasmus program or have an internship in EU or NATO office.
The Epstein class needs a desperate precariat living in favelas to scare the middle class into working more for less. $50T of wealth transfer upward since the 1971 Powell memo[0] and the introduction of so-called "trickle-down economics" of neoliberalism, the liquidation of the middle class has been the primary goal and wildly successful. No amount of voting for neoliberal Democrats or Republicans, two sides beholden to the same master, would ever change anything apart from the momentary blip that was Carter because there were very few politicians interested in We the People for 45 years. Corporate and capital gains tax cuts and non-enforcement of anti-trust laws led to perverse incentives that cut wages, drove offshoring, stock buybacks, and the rise of extreme wealth and inequality, and megacorps.
Young Americans have been fed a constant diet of “the Modern World is Awful” and “You Are Doomed” for quite some time now. This is just the next round, and as always it is overstated.
I do see a rebound in 2027 though. Several signals tell me things will not work out for many AI companies/startups and targets will not be met. People will have to be rehired.
AI's impact on employment is less directly replacing a human worker one to one, but suppressing hiring demand by providing potential employers with a plausible alternative to increasing headcount. This is disproportionately impacting early entry hiring as the ramp up time to productivity for inexperienced workers is today too long for companies to willingly invest for.