Recently, Hupao published an article titled “Are Young Americans Dropping Out of College to Become Electricians?”, which puts forward an impactful thesis: AI is shattering America’s “everyone goes to college” educational faith, and more and more members of Gen Z are fleeing college in favor of blue-collar jobs such as electricians, plumbers, and firefighters. The article spread widely on social media, but do its data, cases, and causal inferences hold up under scrutiny? This article will examine its core content and point out the biases and exaggerations worth watching out for.

Part One: The Article’s Core Arguments and Data

1. Core Claims

The article argues that AI’s replacement of entry-level white-collar roles (customer service, basic programming, data processing, insurance underwriting assistants, and so on) is causing the “first rung” of the traditional career ladder for young people to disappear. Faced with this uncertainty, young Americans are reassessing their career paths on a large scale: some people “outwardly” flee toward the blue-collar world, which requires physical intervention and on-the-ground judgment; a smaller portion “upwardly” embrace computing power, trying to become the manipulators of AI rather than the replaced.

2. Key Data Cited

• Stanford Digital Economy Lab research (based on ADP HR data, as of July 2025): among occupations highly exposed to AI, employment rates for workers aged 22–25 have declined notably since the end of 2022. Different sources give different specific figures: the Hupao original cites a decline of 13% (as high as 20% in some occupations), while Zhihu’s Machine Intelligence retelling puts it at about 6%, and the Xinzhiyuan reprint says approximately 16%.
• Harvard University survey: 59% of young Americans aged 18–29 view AI as a threat to their career prospects.
• Community college enrollment data: enrollment at vocation-oriented community colleges across the country has grown nearly 120% since 2020, reaching 871,000 people.

3. Representative Cases

The article uses several specific people to strengthen its narrative:

• Ryder Paredes, a 22-year-old computer science student, dropped out to attend a vocational school and earn an electrician’s license, on the grounds that “wires won’t be eaten by AI, but my code will.”
• Jackson Curtis, a 28-year-old claims-underwriting assistant, plans to switch to being a firefighter, believing that in moments of crisis people will always crave the compassion of a real human being.
• Vedant Vyas, a 21-year-old who took a leave of absence to found the AI education startup Opennote, securing over $4 million in funding.
• Jensen Huang’s remarks at the Davos forum are quoted: construction workers and utility technicians could see annual salaries exceed $100,000, and a comfortable life doesn’t require a computer science doctorate.

4. Structural Metaphor

The article likens the current U.S. workplace to a system evolving from a “pyramid” toward a “barbell”: at the top are the very few decision-makers who command computing power; at the bottom are workers who need solid physical skills and face-to-face communication; and the middle layer — the former “brain labor movers” — is quietly evaporating.

Part Two: The Article’s Value Points

1. The data supports the trend judgment

Citing Stanford’s research based on big ADP data and Harvard’s youth survey does provide quantitative backing for the macro judgment that “AI is putting pressure on entry-level positions.” This is more persuasive than pure emotional output.

2. The cases strengthen its reach

From dropping out to earn an electrician’s license to taking a leave to raise funding, these stories are highly dramatic, quickly grasp readers’ attention, and help ordinary people understand abstract labor-market changes.

3. It raises the hidden risk of “vanishing apprenticeships”

The article points out that what AI eats is often the “low-hanging fruit” that is “repetitive yet accumulates experience.” If these positions disappear, young people will lose the chance to evolve through practice — a structural problem worth watching over the long term.

Part Three: Inaccuracies and Oversimplifications to Be Wary Of

1. The core data figures contradict one another

The same Stanford study appears with three markedly different figures — 6%, 13%, and 16% — across different republishing channels. This is likely due to differences in statistical scope (for example, relative vs. absolute decline, whether specific occupations are restricted, or differences in time windows), but it also exposes a problem of figure distortion in Chinese-language media during secondary dissemination. If you need to cite this study, it’s advisable to check the original paper directly:

2. The causal attribution risks oversimplification

During the period from late 2022 to 2025, besides the rapid development of AI, the United States also experienced:

• large-scale layoffs in the tech industry caused by the Fed’s aggressive interest-rate hikes;
• the normalization of remote work after the pandemic, with companies cutting back on recruiting entry-level positions;
• cyclical adjustments in sectors such as software development itself.

Attributing the decline in youth employment primarily to AI may underestimate the role of the macroeconomic cycle and industry structural adjustment.

3. The representativeness of the cases is questionable

“Dropping out to be an electrician” and “taking a leave to raise $4 million” are highly dramatized extreme cases. They can illustrate individual choices, but they cannot prove a mainstream trend among U.S. Gen Z. The 120% growth in community college enrollment is a fact, but the article does not provide breakdowns of how many of those people “dropped out” and went blue-collar versus how many represent routine growth in vocational training or post-pandemic rebound.

4. The context of Jensen Huang’s remarks is downplayed

Jensen Huang’s prediction that “utility technicians’ annual salary will break $100,000” is based on a specific current cycle of massive data-center and AI infrastructure construction in the U.S. If the infrastructure boom cools in the coming years, this prediction may not be broadly valid in the long term. Using it as evidence that blue-collar income will keep surging carries a risk of inferring a long-term trend from a specific cycle.

5. “Declining employment rate” is not the same as “being eliminated by AI”

A more rigorous phrasing would be: among occupations highly exposed to AI, employment growth for young workers has been relatively sluggish. Common phrases in Chinese-media headlines such as “the first to be eliminated by AI” and “workplace massacre” reflect obvious emotional exaggeration and easily manufacture unnecessary panic.

6. The source of “2026 as the turning-point year of the no-jobs big bang” needs verification

The article mentions that “2026 is exactly the turning-point year of what Hinton called the ’no-jobs big bang.’” Geoffrey Hinton has indeed publicly warned many times about AI’s long-term impact on employment, but whether “2026” is his exact wording or the result of media embellishment through multiple retellings currently lacks confirmation from a first-hand source.

7. Information has passed through multiple retellings, and original qualifiers are omitted

Tracing the provenance, the Hupao and Xinzhiyuan articles partially reference the Wall Street Journal report and the Stanford paper mentioned above. After multiple retellings by Chinese-language media, subtle conditions in the original research — sample scope, confidence intervals, author-stated limitations, and the like — tend to be omitted, making the conclusions appear more absolute than the original.

Part Four: Summary

“Are Young Americans Dropping Out of College to Become Electricians?” is a highly readable trend observation that successfully captures the collective anxiety over structural changes in the labor market in the AI era. The Stanford research and the Harvard survey it cites have some reference value, and its “vanishing apprenticeships” and “barbell-shaped workplace” are propositions worth pondering.

However, in terms of specific figures, causal inference, and the representativeness of its cases, the article contains biases worth being wary of. Through dramatized individual examples and emotional headlines, it inflates “sluggish employment growth” into “large-scale elimination” and “people switching to trade skills” into “everyone quitting college to become electricians.”

For readers, rather than being swept up in the anxiety of “AI eliminating white-collar workers,” it’s better to see more calmly that:

• AI’s impact is structural and differential, not a one-size-fits-all “annihilation”;
• the pressure on entry-level positions is a fact, but the macroeconomic cycle is an equally important variable;
• lifelong learning and skill renewal hold more long-term value than simply “fleeing college” or “embracing college.”

In the age of information overload, staying sensitive to data sources and prudent about causal inference is in itself an important form of cognitive defense.

References

• Hupao original
• Zhihu · Machine Intelligence retelling
• Xinzhiyuan reprint
• Wall Street Journal report
• Stanford original paper PDF