← All articles

The Trust Chain

Every career decision runs on a chain of trust: education leads to employment, employment leads to stability, stability persists, the rules hold. All four links broke at once.

Bert Carroll ·
LinkedIn RSS

You tell a seventeen-year-old to pick a major. You tell them this decision shapes the next forty years of their life. You tell them to take it seriously.

Then the requirements change before they graduate.

This is not a story about entitled kids or a broken generation. This is a story about a structural failure in the assumptions that held American career planning together for fifty years. And almost everyone diagnosing it is looking at the wrong layer.

The Chain

Every major life decision runs on a chain of trust. Not trust in a person or an institution, but trust in a sequence: that one thing leads reliably to the next.

For most of the twentieth century, the career chain had four links:

  1. Education leads to employment.
  2. Employment leads to stability.
  3. Stability persists long enough to build a life.
  4. The rules stay roughly the same while you are playing.

None of these needed to be guaranteed. They needed to be probable enough to justify the investment. A college degree did not promise wealth. It promised a floor: you would be okay. Maybe not rich, but okay. House, family, retirement, dignity.

The floor was the product. And for decades, it held.

The first link degraded slowly. A computer science degree used to map cleanly to a software engineering career. Now it might mean software engineer, data analyst, product manager, or “we are not hiring juniors right now.” Gallup tracked confidence in higher education from 57% in 2015 to 36% in 2023, a historic low.1 That is not a PR problem for universities. That is a market signal.

The second link broke faster. You can do everything the system asks of you and still find that the median home costs 5.6 times the median household income, up from 3.2 times in the 1990s.2 You can land the right job in the right city and discover that “right” does not convert to a down payment. The output decoupled from the input.

The third link is the newest break, and the most corrosive. Even when someone wins, they cannot trust the win to hold. Median job tenure dropped to 3.9 years in 2024, the lowest since 2002.3 An entire career’s worth of stability now lasts less than a presidential term. And that was recorded before the current wave of AI-driven restructuring.

The fourth link is the one nobody talks about directly, because it sounds abstract until it is happening to you. The rules changed. Entry-level job postings declined 29 percentage points globally since January 2024.4 Thirty-five percent of positions labeled “entry-level” on LinkedIn require three or more years of experience.5 KPMG cut graduate intake by 29% between 2022 and 2024, citing AI handling the tasks those graduates used to do.6

The rules did not just change. They changed while people were mid-game, holding debt they took on under the old rules.

Walk into any freshman lecture hall and you will find two groups sitting next to each other with nothing in common.

The first picked college because they have seen the outlier stories. Tech founder at 25. Six figures out of a bootcamp. They are optimizing for the tail of the distribution because the median outcome no longer justifies the investment. If the safe path does not feel safe, you might as well swing.

The second group is there because someone told them to be. Parents, guidance counselors, the cultural script that says college is what comes after high school. They did not make a decision. They followed an instruction. And now they are $30,000 in with no thesis about why they are there or what they are building toward.

Neither group has a feedback mechanism. There is nobody at month six of freshman year showing them what employment rates actually look like for their major at their school in their region. No signal. No course correction. Just sunk cost and social pressure to finish what they started.

The Error Bar

I am a CTO who builds software for a living. I have been working in technology for over two decades. And my confidence interval on what work looks like in ten years is wider than it has ever been.

That is not false modesty. It is an honest read of the environment. The tools I use today did not exist two years ago. The roles I hire for are shifting underneath active job descriptions. I can make guesses about 2030. I cannot make promises.

If my error bar is that large, imagine the error bar for an eighteen-year-old with no baseline to calibrate against. They are being asked to commit to a forty-year architecture with requirements that might be obsolete by the time they finish the foundation.

In software, when your requirements are unstable, you do not commit to a monolith. You build for adaptability. You keep your options open. You make smaller, reversible decisions until the spec stabilizes.

The kids opting out of traditional career paths are not lost. They are architecting for uncertainty. They are doing exactly what a competent engineer would do when the spec keeps changing.

Rational Fear

When all four links break at the same time, people do not experience a “career crisis.” They experience fear. Specific, rational, well-founded fear.

Forty-seven percent of Gen Z report excessive, difficult-to-control anxiety.7 Only 31% feel financially secure.8 The American Psychological Association reports that adults 18 to 34 carry an average stress level of 6.0 out of 10, nearly double the rate of adults over 65.9

The standard reading is that young people are fragile. The structural reading is that they are doing the math correctly.

When the institutions that were supposed to manage risk have all retreated (no pension, no employer loyalty, no affordable housing on the horizon), the risk lands entirely on the individual. They have to be focused on themselves. The alternative is not selflessness. It is ruin. The fear is specific: I will not be able to own a home. I will not be able to support a family. I will not be able to build the life I watched my parents build.

The data supports every piece of that fear. The US fertility rate hit 1.62 in 2023, the lowest ever recorded.10 Median age at first marriage reached 30.6 for men and 28.7 for women, both historic highs.11 These are not cultural trends driven by preference. They are economic signals. People are not choosing to delay families. They are waiting for a floor that is not arriving.

And when individual risk management becomes unbearable at scale, people reach for collective risk management. Gallup found in 2018 that 51% of Americans aged 18 to 29 held a positive view of socialism, compared to 45% for capitalism, the first time socialism polled higher in that age group.12 You do not have to agree with their conclusions to understand the math that got them there.

The Pipeline They Are Destroying

There is a second structural failure running parallel to the first, and it is being committed by the same companies complaining about talent shortages.

When companies replace junior engineers with AI, they save payroll for two or three years. Then their mid-level engineers leave. Their senior engineers retire. And they discover that no one in the pipeline learned their systems, their domain, or the judgment calls that keep production running at three in the morning.

AI can generate code. It cannot generate the five years of context that turns a junior engineer into someone who knows which corners not to cut. Companies cutting entry-level roles are not optimizing. They are borrowing against their own future and calling it efficiency.

The Big Four accounting firms cut graduate hiring by double digits between 2022 and 2024.6 The tech sector cut junior postings 34% from peak while senior postings held steady.13 These are not corrections. They are structural decisions to stop investing in the pipeline that produces experienced professionals.

In ten years, these companies will compete for a dramatically smaller pool of mid-career talent, because the people who would have become that talent were never given the chance to start. The error bar on “we will figure it out with AI” is at least as wide as the one facing the eighteen-year-old. The difference is that the company has a balance sheet to absorb the mistake. The kid does not.

Who Is Thinking Clearly

Trade school enrollment rose 16% from 2022 to 2023.14 Apprenticeship programs grew 11% in 2024 alone.15 College undergraduate enrollment dropped 8% between 2019 and 2022.16

The standard narrative frames this as anti-intellectualism. As young people giving up on aspiration. The structural read is the opposite.

An electrician’s license has visible demand, a known path, and a predictable payoff. The feedback loop between investment and outcome is short and verifiable. A four-year degree in a field that might be restructured by AI before graduation does not have those properties.

The kids choosing trades are not settling. They are making the most architecturally sound decision available to them. They looked at the error bars on both options and picked the one where the range was narrowest. That is not retreat. That is engineering.

What Still Holds

The trust chain has not broken everywhere. For electricians, plumbers, and welders, the links still connect. Visible demand, known path, predictable payoff. That is precisely why enrollment is surging. Those kids did not give up on ambition. They found the last place where the deal is still legible.

It still holds, for now, for teachers and nurses. The demand is real, the credential path is clear, the outcome is stable. But both professions sit directly in AI’s path. Not replacement, but augmentation that compresses the number of humans needed to deliver the same outcome. The chain holds today. Whether it holds in ten years is exactly the kind of question nobody is helping people answer.

And the framework does not explain everyone. Some people went to college, picked a major, got a job, bought a house. The chain worked for them. But a system that works for some people some of the time is not a system. It is a lottery with good marketing.

What Honest Looks Like

No one can tell you what work will look like in twenty years. Anyone claiming otherwise is selling certainty they do not possess. My error bar is wide. Yours is too. And it is wider still for someone just starting out.

But uncertainty is not the same as helplessness. When an engineer faces unstable requirements, they do not guess and commit. They build visibility: monitoring, feedback loops, decision points that let you change course without starting over.

That is what is missing from career planning. Not better advice. Better instrumentation. Show an eighteen-year-old, before they enroll, what graduation rates, transfer pathways, and local job outcomes actually look like for their intended major at their specific school. Show a first-generation family how to turn AP, CLEP, and dual enrollment credits into arriving with sophomore standing instead of freshman debt. Show a mid-career professional which of their skills are appreciating and which are depreciating. Show a company what their talent pipeline looks like in five years if they keep cutting the entry level.

Full disclosure: I am the CTO of College Decoded, which is trying to build exactly this. That is not incidental to the argument. I wrote this article because I am doing the work, not because I read about it.

The information exists. It is scattered across BLS reports, LinkedIn data, university outcome studies, and employer hiring patterns that no individual has time to assemble. The problem is not that we lack data. The problem is that the people making the biggest decisions have the least access to it.

We do not need to rebuild the floor. We need to make the ground visible so people can build where it is solid and move when it shifts.

Sources

  1. Gallup, "Americans' Confidence in Higher Education Down Sharply," July 2023. Confirmed at 36% in 2024 follow-up; rose slightly to 42% in 2025, still well below the 57% recorded in 2015. news.gallup.com ↑
  2. Harvard Joint Center for Housing Studies, "Home Prices Surge to Five Times Median Income, Nearing Historic Highs," 2023. Peak ratio of 5.6x median household income in 2022, compared to approximately 3.2x in the 1990s. jchs.harvard.edu ↑
  3. Bureau of Labor Statistics, "Median Tenure With Current Employer Was 3.9 Years in January 2024," USDL-24-1936, September 2024. Down from 4.1 years in 2022. bls.gov ↑
  4. Randstad, "Gen Z Workplace Blueprint," 2025. Analysis of 126 million job postings globally. Entry-level postings (0-2 years experience) declined 29 percentage points since January 2024. randstad.com ↑
  5. George Anders, LinkedIn Economic Graph, "Hiring's New Red Line: Why Newcomers Can't Land 35% of Jobs," August 2021. Analysis of 3.8 million job postings (Dec 2017–Aug 2021). 35% of entry-level positions required 3+ years of experience; rate had climbed to 38.4% by publication. linkedin.com ↑
  6. City AM, "Big Four Slash Graduate Jobs as AI Takes Over Entry-Level Tasks," 2024. KPMG graduate intake reduced 29% (1,399 to 942), Deloitte 18%, EY 11%, PwC 6%, all between 2022 and 2024. cityam.com ↑
  7. EY, "2023 Gen Z Study: Financial Worry Is a Top Driver of Anxiety," September 2023. 47% of Gen Z respondents feel unable to stop worrying more than half the days or nearly every day. ey.com ↑
  8. Ibid. 31% of Gen Z respondents reported feeling financially secure; 52% reported being "very or extremely worried" about not having enough money. ey.com ↑
  9. American Psychological Association, "Stress in America: Generation Z and Millennials," 2023. Adults 18–34 reported average stress of 6.0/10 versus 3.4/10 for adults 65 and older. apa.org ↑
  10. CDC National Center for Health Statistics, National Vital Statistics Reports Vol. 74 No. 1, March 2025. Total fertility rate of 1.62 births per woman in 2023, the lowest rate since national tracking began. cdc.gov (PDF) ↑
  11. Bowling Green State University, National Center for Family and Marriage Research, "Crossover in Median Age at First Marriage and First Birth: Forty-Three Years of Change," FP-25-13. Median first marriage age of 30.6 (men) and 28.7 (women) as of 2023. bgsu.edu ↑
  12. Gallup, "Democrats More Positive About Socialism Than Capitalism," August 2018. Positive view of socialism at 51% versus 45% for capitalism among adults 18–29. Notable: when "free enterprise" is substituted for "capitalism," support in the same cohort rises to 83%. news.gallup.com ↑
  13. Indeed Hiring Lab, "Experience Requirements Have Tightened Amid the Tech Hiring Freeze," 2025. Junior tech postings down 34% from pandemic-era peak versus 19% for senior roles. hiringlab.org ↑
  14. National Student Clearinghouse Research Center, Current Term Enrollment Estimates, 2023. Vocational-focused community college enrollment rose 16% from 2022 to 2023 (~813,000 students). nscresearchcenter.org ↑
  15. U.S. Department of Labor, Office of Apprenticeship, national apprenticeship statistics, 2024. apprenticeship.gov ↑
  16. National Student Clearinghouse Research Center, Current Term Enrollment Estimates, 2019–2022. Undergraduate enrollment declined approximately 8% over the period. nscresearchcenter.org ↑