The Floor is Lowering
What's next for this generation's workforce?
I keep hearing that the labor market is healthy. Almost every ambitious person I know under 35 is quietly recalculating their life in real time.
I have spent the last few weeks talking to executives in banking, consulting, legal, real estate, and tech. The pattern is too consistent to be anecdotal anymore. The headlines say one thing. The conversations say another. The data sides with the conversations, and the gap has been widening longer than most of us noticed.
The credential, first
I am starting a Master’s Program in 2027 (more to come on this later). The numbers below are the ones I keep coming back to.
The share of Harvard Business School graduates without a job three months after commencement more than doubled between 2022 and 2024, from 10% to 23%. The school’s own 2025 figures recovered, but an independent analysis of LinkedIn data on 27,531 graduates of the top seven programs put roughly half of the 2025 cohort without a recorded job nine months out. Methodology differs. The shape of what is happening does not.
The placed half still does well. Median base salaries north of $175,000 within months of graduation. Three-year alumni medians at the top programs above $245,000. The credential still opens doors. But the expected value has compressed, the variance has widened, and at a fully-loaded sticker cost approaching half a million dollars, the program has stopped being the default and started being a high-conviction bet. The old equation, that education plus credentials plus time equals stable upside, has stopped working.
Law school. Medical residency. Engineering programs. Coding bootcamps. Every credential that used to mean something is compressing the same way.
A close friend of mine has two Ivy League degrees, one undergraduate and one graduate, and the kind of resume that should have made the next ten years obvious. He has been looking for paid work for months. The last conversation we had was not about long-term plans. It was about how to get to the end of the year. I have had versions of that conversation myself. The credentials did what they were supposed to do. The market under them did not. Watching someone with that profile run that math is the moment the abstract becomes personal.
The deeper problem is that the market is no longer absorbing people the way it once did. Not even at the highest levels.
Where the cuts are actually landing
The April 2026 jobs report read like a soft landing. Unemployment held at 4.3%. Anyone reading the headline would conclude the economy has stabilized. The problem is that the headline is measuring the wrong thing. Underneath it, the long-term unemployed sat at over a quarter of all unemployed people, and part-time workers who would rather be working full-time spiked by nearly half a million in a single month. The cuts of the last six months tell the real story.
KPMG cut 10% of its U.S. audit partners in late April. About 100 partners, all equity holders in the firm, separated after a voluntary retirement program failed to produce enough exits. The CEO’s own statement said the audit business “remains strong.” These were not cuts to a struggling division. They were cuts to a profitable one.
JPMorgan is the cleanest tell. Jamie Dimon told analysts in February that the bank had “displaced people from AI” and was building “huge redeployment plans.” The bank’s headcount stayed flat. Operations and support roles shrank. Client-facing roles grew. The bank did not shrink. It rotated. At Davos, Dimon said he would welcome a government ban on mass AI layoffs “if we have to do that to save society.” When the CEO of the largest bank in the world publicly asks the government to slow him down, the underlying dynamic is worth taking seriously.
McKinsey is the second tell. The firm cut technology and support staff at the end of 2025 and signaled deeper reductions in non-client functions over the next 18 to 24 months. Headcount has come down from a peak above 45,000 by several thousand. The firm built its entire business model on the analyst pyramid: junior people doing the synthesis work that gets repackaged at partner billing rates. AI does the synthesis now. The pyramid no longer pays.
In the last 90 days I have watched friends get separated from firms they helped build. None were performance issues. All of them got the standard package and a vague reference to AI productivity in the all-hands. They are, on paper, the people the system was supposed to reward.
What’s Clear
The clearest single image of where this is going landed last week. PitchBook published a piece profiling AI agents that VC firms have begun deploying not as scheduling tools but as functional Principals and Associates.
Patron, a $200 million consumer VC in New York, has an agent named Daisy listed as a principal. Her persona includes a bachelor’s from the Wharton School of Business, prior roles at Universal Music and Andreessen Horowitz, and a Swiss birthplace. She is, in fact, an AI agent. Patron’s co-founder Jason Yeh explained the logic plainly: “We thought a lot about: What type of person would we want to hire? What would be most impactful to us today, short of hiring a fourth partner?” The answer was Daisy, hired instead of a human.
Pebblebed, an early-stage venture firm, runs a fleet of named agents. One is called Diligence Baby, given an MBA from Stanford’s Graduate School of Business, who takes the first pass at every pitch deck. The agents communicate on a Slack-style app, hold their own daily stand-up, and reportedly hold grudges against each other after one of them inaccurately claimed it had completed a task.
Jed Cairo, the managing partner at Juxtapose, summarized it for PitchBook: “Analytical ability is being commoditized to some extent.” That is the entire argument compressed into one sentence by someone who hires for a living.
I came up through venture capital. The analyst-to-principal pipeline I trained inside of is being templated out in real time, by firms run by people I know. Real graduates with real Wharton and Stanford MBAs are entering a market in which AI agents have been assigned those same credentials as part of their persona. That is not a joke. It is a tell.
Who is actually getting hired
The Federal Reserve Bank of New York’s tracker for recent college graduates put unemployment for 22-to-27-year-olds with bachelor’s degrees at 5.7% at the start of 2026, higher than the national rate. Underemployment, defined as graduates working in jobs that do not require a degree, sat above 41%. The historical pattern, in which college graduates fared meaningfully better than the broader workforce, has reversed.
A Stanford study analyzing payroll data from millions of workers looked at occupations most exposed to AI: software engineering, customer service, accounting, junior consulting. In those occupations, employment for the youngest cohort of workers declined sharply since late 2022. For software developers in their early twenties, the drop reached nearly 20%. New graduates now make up a single-digit share of hires at major tech companies, less than half of what it was before the pandemic.
The pattern is consistent. AI is not visibly cutting headcount across the white-collar economy. It is cutting the bottom rung. Senior people keep their jobs. Junior people don’t get hired in the first place. The result, on a long enough timeline, is a workforce with no replenishment pipeline. A problem nobody has to solve until they suddenly do.
The rest of the contract is going at the same time
Earning a living is half the picture. The other half is whether the earning lets you build a life. It does not.
Home prices have risen more than 50% since 2019. Median household income has risen roughly half that. The home-price-to-income ratio nationally sits near double what financial advisors call the affordability ceiling, and not a single one of the 50 largest U.S. metros meets it. The median age of a first-time homebuyer is now 40. In 1981 it was 29.
Family formation runs on the same logic. More than half of 18-to-24-year-olds now live with their parents. Fertility rates are at record lows. Marriage rates are declining. These are not lifestyle preferences. They are budget constraints, and they are reshaping the texture of an entire generation’s adult life.
What I notice in my own circle is harder to put in numbers. The people who would have been settling into something by now, buying, marrying, anchoring, are on their fourth city, single more by default than choice, with friendships scattered across continents because everyone keeps moving for the next opportunity that might finally be the stable one. Educated, capable people paying premium rent for the privilege of waiting to start their actual lives. Children deferred. Parents aging in another country. The career conversation that used to happen at 28 now happens at 35, and the answer is more often “I’m still figuring it out” than anyone is comfortable admitting.
The entrepreneurship surge is the same dynamic in another form. New business applications are running well above historical averages, which a lot of commentary reads as evidence that Americans are adapting. Look more carefully. Of nearly half a million applications filed in March, only about 6% are projected to become businesses with employees. The rest are sole-proprietor LLCs, contractor pass-throughs, side hustles, consulting shells.
The cultural script that says everyone should “build something” is a coping narrative for a labor market that has stopped offering the alternative. Not everyone is cut out for entrepreneurship. Most people historically have not been. The shift toward self-employment is not a renaissance of American ambition. It is a structural compression that has forced a large part of the population to invent an income stream because the labor market stopped offering one.
The shape underneath
Underneath all of this is a wealth shape that explains the rest. The top 1% of American households now controls roughly a third of all U.S. wealth, nearly equal to what the bottom 90% hold combined, the highest concentration since the Federal Reserve began tracking it in 1989. The S&P 500 just delivered its third consecutive year of double-digit gains. If you owned the index, you got richer. If you didn’t, you watched the index get richer without you and tried to hold rent steady.
The question worth asking out loud
A reasonable person looking at this could ask whether what is happening is the byproduct of independent policy choices, or something closer to a design.
I do not know the answer. The cumulative effect of choices made over the last forty years has run in one direction. Tax policy favored capital over wages. Monetary policy from 2009 to 2022 inflated asset prices and transferred wealth to people who already owned them. Housing policy subsidized demand and never produced enough supply.
None of this is a conspiracy. All of it is a structure. Asset holders write the laws governing asset markets. At some point the question of whether the outcome was intended becomes less interesting than whether anyone is going to change it. On current evidence, no.
This is not an AI apocalypse, either. AI is the visible accelerant of a shift that was already in motion. Recent research suggests that observed AI exposure runs far behind theoretical capability, and that most AI interactions are augmenting human work rather than replacing it outright. AI is not yet replacing employed workers at scale. It is preventing companies from hiring the next cohort. The distinction is invisible in the unemployment rate and obvious in the underemployment rate. It produces a slow erosion rather than a sudden shock, which is exactly why it does not look like a crisis on the front page, and exactly why it is more consequential than a crisis would be.
What this actually feels like
The hardest part of this transition is not economic. It is psychological.
An entire generation built its identity around institutions and pathways that no longer produce predictable outcomes. The MBA. The promotion track. The home as wealth-building tool. The notion that a strong resume protected you from market shocks. None of these are gone. They just no longer deliver what they were supposed to.
The disorientation is not coming from any single failure. It is coming from realizing the map changed while everyone was still following it.
That is what makes this moment different from earlier downturns. The 2008 crisis was a shock you could name and date. This is a structural drift you only see clearly in retrospect, in conversations with friends who used to be sure of where they were going, in the gap between what your year was supposed to look like and what it actually does. People keep waiting for the moment things go back to normal. Normal already left.
What people are actually doing
If the old American social contract assumed that education plus work plus time produced compounding security, that contract has been retired. What seems to be replacing it is narrower. The people I see actually getting ahead, not just maintaining, are doing some combination of three things.
Equity ownership in something that compounds. Salary income gets taxed at the top rate, spent on a cost of living that has risen faster than wages, and produces no compounding asset. Equity does. The wealth gap is mostly an equity-ownership gap.
Selection into a small number of high-leverage networks. The graduate degree still works at the top programs because they remain the most reliable mass-scale entry into networks that produce founder seats, partner tracks, and capital allocation jobs. Those are the positions where decisions about other people’s capital generate economic rent. The networks are the asset. The credential is the entry pass.
Being early in a real structural shift. The wealth created in technology between 2010 and 2022 was not created by people who got hired into mature roles. It was created by people who joined small teams that scaled fast, or who built their own. The same logic applies now to AI infrastructure, to capital flows out of the U.S. into emerging markets, to climate-adjacent sectors, to the financialization of previously informal economies.
I am writing this from inside the contradiction. Betting on a credential while challenging the credential. Watching friends recalculate while running my own arithmetic. Planning to build the next chapter from a different geography because the math here has stopped working for the kind of life I want.
The only honest answer
The economy is not collapsing. It is bifurcating. The aggregate numbers will continue to look fine for a while. What is happening underneath is harder to fix. The floor is lowering. The ceiling is rising. The middle is being asked to invent itself, alone, on a deadline nobody set publicly.
The only honest answer I have, after all of this, is to build.
Not in the motivational sense the word has been emptied into. In the literal one. A platform someone could pay for. Skills that compound. Assets that give you leverage inside any structure you walk into: equity, distribution, audience, capability. Use the tools at hand, including the same AI that is templating the analyst pyramid out of existence. The leverage cuts in both directions, but only for the people who pick it up.
I do not say this triumphantly. I say it because I have looked at this for long enough to stop expecting an institution to come back and carry the weight again. The rules already changed. The people who notice are adjusting. The people who don’t are losing ground and wondering why their effort no longer translates.