A new survey of 120 economists has a blunt read on where the job market stands: already cooled, with more cooling ahead, and white-collar work bearing the brunt of an AI reshuffling that’s less about mass layoffs and more about which tasks still need a person. That’s the headline finding from Indeed Hiring Lab’s inaugural Labor Market Outlook Survey, fielded in July 2026. It lands the same week one of Silicon Valley’s most prominent venture capitalists published his own warning that AI enthusiasm has outrun investment discipline — and a real example, buried in his own portfolio, of exactly the kind of white-collar task AI is already replacing.
Put those two data points next to each other and a clearer picture forms than either gives alone: the excitement about AI is real, the caution about the excitement is also real, and the actual disruption to white-collar work is narrower and more specific than either the hype or the panic suggests.
What the economists actually forecast
Indeed Hiring Lab’s Q2 2026 survey drew responses from 120 economists and labor market experts across academia, financial institutions, and independent research organizations. The consensus isn’t a downturn. It’s a drift. Most panelists expect the Indeed Job Postings Index to fall by roughly 1.4% on average through June 2027, and unemployment to inch up from around 4.2% to roughly 4.4% by year-end — a gentle cooling, not a collapse. For context, the Bureau of Labor Statistics put the actual July 2026 unemployment rate at 4.1%, so the survey’s baseline and the most recent official print are telling a consistent, if slightly noisy, story: a labor market losing a little heat, not falling off a cliff.
The AI-specific findings are where the “reshuffling” framing comes from. A slim majority of economists, 52%, expect AI to be at least a mild drag on employment over the next year. Only 35% expect a net employment gain from AI, and 13% expect no effect at all. That’s a meaningfully split panel — there’s no consensus that AI is a net job creator right now, but there’s also no consensus that it’s a net job destroyer. What the panel agrees on more strongly is productivity: nearly all expect AI to raise output over the next three years, though the expectations are restrained. Seventy percent expect only a modest-to-moderate boost, and just 4% expect anything transformative.
The most pointed finding for anyone in a white-collar role: 57% of economists expect downward pressure on the wages of college-educated workers over the next year. That’s not a forecast about job counts. It’s a forecast about leverage — and it lines up with a market where AI tools are absorbing routine cognitive tasks faster than they’re absorbing judgment-heavy ones, which shifts bargaining power away from workers whose value was mostly in doing the routine part well.
A venture capitalist’s own letter shows the reshuffling in progress
The same week the survey made headlines, Thrive Capital founder Joshua Kushner published his firm’s first-ever investor letter, and it read less like standard VC boosterism and more like a warning shot aimed at his own industry. “It would also be a grave error in our minds to let excitement weaken our investment discipline,” he wrote, arguing that Silicon Valley “can become fixated on hyperincremental technological turns rather than where the technology ultimately leads.”
That’s notable coming from someone whose firm has $60 billion under management and has ridden AI bets — including early stakes in OpenAI and a stake in Cursor, the AI coding tool that just sold to SpaceX for $60 billion — to a reported 41% gross internal rate of return. Kushner isn’t arguing AI is overhyped in the sense of being fake. He’s arguing that the market’s excitement about AI has started to outpace the actual, differentiated value of any given bet, and that discipline matters more than ever precisely because so much capital is chasing the same handful of trends. It’s a useful corrective for job seekers too: the existence of AI hype doesn’t mean nothing real is happening underneath it, and vice versa.
Buried in the same letter is a concrete data point that matters more to job seekers than anything about fund returns. Thrive Holdings, the firm’s operating arm that buys and operationally overhauls companies, runs an accounting platform where AI agents now produce tax returns 30% faster with 98% accuracy. Its IT services business has agents independently resolving half of all help desk tickets. Those aren’t projections or pilot programs. They’re operating numbers, from a firm with every incentive to overstate its AI story, in two of the most classic white-collar back-office functions: accounting and IT support.
That’s the reshuffling the economists are forecasting, made concrete. It’s not “AI is coming for white-collar jobs” as an abstract threat. It’s specific, well-defined, high-volume tasks — routine tax prep, tier-one help desk tickets — getting absorbed first, in exactly the functions where the work is repeatable enough for an agent to handle reliably.
Why AI reshuffling white-collar jobs in 2026 means “reshuffling,” not “replacing”
The distinction matters for anyone trying to figure out what to actually do with this information. A 52%-of-economists-expect-a-drag statistic sounds alarming in isolation, but paired with the Thrive data, it points to something narrower: the tasks being automated first are the ones that were already candidates for automation — repeatable, rules-based, low-judgment work. Half of IT help desk tickets resolved independently by an agent doesn’t mean IT support as a discipline is disappearing. It means the entry-level, ticket-triage layer of that job is shrinking, while whatever’s left over — the escalations, the ambiguous cases, the actual troubleshooting judgment — becomes relatively more valuable, not less.
That’s consistent with the wage-pressure finding too. If AI is absorbing the routine slice of college-educated work faster than the judgment-heavy slice, the workers most exposed to downward wage pressure are the ones whose day-to-day work looked more like the routine slice. The workers positioned to benefit are the ones who can point to the judgment slice specifically — decisions made, ambiguity resolved, outcomes owned — rather than tasks completed.
None of this cancels out the macro cooling the survey also found. A softer job market and a reshuffling within white-collar work are compounding pressures, not competing explanations. Fewer total postings plus a shift in what makes a candidate valuable within those postings is a genuinely harder environment than either factor alone. It’s also worth noting what the survey didn’t find: no majority of economists expects a recession-style contraction. This is a market getting choosier, not one collapsing, and choosier markets reward specificity over volume in a job search.
What a cooling, AI-reshuffled market means for your resume
The practical effect shows up first in how resumes get read, before it shows up in headline layoff numbers. A resume built around task completion — “processed X returns,” “resolved Y tickets” — is describing exactly the layer of work that’s shrinking fastest, even if the person doing it also handled every escalation and edge case along the way. The resume format itself buries the part that’s actually still in demand.
Rewriting that story for a resume helps, but it only gets read if the resume reaches someone. The more durable fix is separating the routine-task framing from the judgment framing entirely, and leading with the latter whenever there’s a chance to make a direct case to a person — in a cover message, an outreach email, or a conversation — rather than compressing both into a bullet point that an ATS scans and a human skims for ten seconds.
What this means for how you look for work
In a market like this, the passive version of a job search — submitting an application and waiting for an ATS to surface it — gets worse on two fronts at once. There are fewer postings to apply to, per the economists’ own forecast, and even among the postings that exist, keyword-matching software has no way to detect whether a candidate does judgment-heavy work or routine work well. Both types of candidates can write “detail-oriented” and “process improvement” on a resume. Only one type is actually insulated from the reshuffling these economists are describing.
The alternative is making that distinction explicit, to a specific person, before the application even happens. If the routine slice of a role is what’s shrinking, the pitch to a hiring manager needs to lead with the judgment slice: the ambiguous case that got resolved, the escalation that got handled, the decision that a rules-based system couldn’t have made. That’s not a story an ATS can extract from a resume. It’s a story that has to be told directly, to the person who’d actually recognize its value — which is exactly the kind of message that gets read and gets a reply, precisely because almost nobody sends it.
None of this requires waiting for certainty about how the AI reshuffling plays out over the next few years. The economists themselves don’t agree on the size or direction of the employment effect — that’s the honest state of the forecast. What’s clear enough to act on now is narrower and more useful: the routine slice of white-collar work is under real pressure today, the judgment slice isn’t, and the job search process most people default to has no way to tell a hiring manager which slice a given candidate actually belongs to.
Economists forecasting a cooler market and a VC admitting his own industry’s excitement has outrun its discipline are both, in their own way, telling job seekers the same thing: generic positioning in a softening market is a losing bet. angld.AI automates the part of direct outreach that’s hardest to do at scale — identifying the actual hiring manager behind a role, researching their background, and drafting a personalized message — so the differentiation can come from the story being told, not from whether someone had the hours to do the research themselves.
