Your Entry-Level Pipeline Is About to Get a Lot Thinner, and Faster Than You Think
Sign Up For Daily Job Market News
Bill Gates published a long essay this week, and the section on jobs is the part I’d stop and read twice. His central claim is that AI can replace, and in some cases exceed, human cognition for the first time. He thinks that single fact makes every comparison to past technology shifts, the PC, the shift from farm work to office work, misleading.
Here’s his math on speed. The PC took 20 years to meaningfully change how people worked, because software had to get built, prices had to fall, and workers had to learn new tools.
AI runs on devices we already own and speaks plain language back to us. In Gates’ words, “We don’t have to adapt to it because it can adapt to us.” He expects law, customer service, medicine, software, and manufacturing to feel the disruption over roughly a decade, not the several generations it took agriculture to give way to office work.
His job math is just as direct: “There will be some new jobs, but without the right policies there will be far fewer than exist today.” The roles he names first in line are sales, customer support, software engineering, and paralegal work, with loan assessment, data analysis, and patient triage following behind. Software engineering gets a partial exception in his view, since falling costs there should generate new demand even as some tasks get automated.
The part that should worry every hiring manager reading this: Gates says the jobs at the most risk sit at the entry and mid-level, while the new jobs being created mostly require skills that take years to build. He points to data showing that after generative AI adoption took hold, employment fell significantly among young workers in the jobs most exposed to replacement, but not among their older colleagues in the same roles.
I’d put it this way for anyone managing early-career hiring: Gates is describing a broken ladder, not a broken job market. Employers still hire experienced people at the rate they always have. The rung where a 22-year-old used to get their start is the one quietly disappearing from requisition lists.
Gates also proposes “Human Reserved” jobs, a category of work set aside for people by choice rather than by capability. He compares it to a nature reserve, land we could develop and choose not to. His example is the caregiving team that looked after his father through late-stage Alzheimer’s, work he calls irreplaceably human.
He’s honest that he can’t fully answer who decides which jobs qualify or how you’d stop a company from cutting corners on the idea. My addition to his list: if you reserve a job for people, you have to be willing to pay for it like protected work. Reserving a role without raising its wage just reserves a shortage.
On the money side, Gates points to a real asymmetry. Hire a person and you pay payroll taxes on their wages. Buy a robot and you can usually write it off immediately as a business expense.
He wants to tax AI tokens and robots to slow that pull a little and fund retraining and the safety net. He’s proposed a version of this before, got called strange for it, and hasn’t backed off. I doubt the robot tax goes anywhere politically, but the tax code’s current preference for machines over people is real, and it’s worth knowing which way the incentive points before you build next year’s headcount plan around cost per hire alone.
He reaches for 1933 to make one final point, when U.S. unemployment ran roughly 25% and stayed in double digits for most of the following decade. He isn’t predicting AI gets anywhere near that level. His point is that this kind of shift doesn’t fix itself with a normal economic cycle the way a recession does.
Gates also discloses his own financial ties to Microsoft and other AI companies through his foundation work, and says the profits from those investments go to the Gates Foundation. He leaves it to readers to weigh that against his conclusions, and I’d apply the same standard to the next story.
Staffing your team doesn’t have to be hard.
Reach out and see how we can help.
The Labor Department Just Admitted It Can’t See What AI Is Doing to Jobs
Acting Labor Secretary Keith Sonderling said essentially that this week, and his fix is to ask the companies selling AI to hand over the data the government doesn’t have. The Labor Department has signed memorandums of understanding with OpenAI, Google, Meta, and Amazon, among other firms, to get a better read on how businesses are actually using AI. “The government does not have the data,” Sonderling said. “The large tech companies and the large Fortune 500 companies… are really going to be the most impacted by this.”
I understand the reasoning. Companies like OpenAI and Amazon hold usage and hiring data no federal survey can match for speed or detail. But the incentive problem sits right in the open: every one of those companies has a revenue reason to describe AI’s effect on jobs as augmentation rather than replacement.
Sonderling told Axios he expects “more augmentation in jobs” and “new jobs being created,” in the same conversation where he says the government doesn’t have the data to back that up. I’d hold him to publishing the raw inputs and letting people draw their own conclusions.
The stakes here reach past labor economics. BLS employment and inflation numbers feed directly into Federal Reserve interest rate decisions, and Fed Chair Kevin Warsh started a task force this summer looking for alternative real-time data sources for policymakers.
Part of the problem is structural: BLS survey response rates have been falling for years, which widens the margin of error and makes revisions bigger, feeding the argument that the numbers can’t be trusted in the first place. Fixing survey participation would do more for BLS credibility than any communications plan.
Trust is the other half of this. Sonderling pointed to the post-election revisions that showed job growth weaker than first reported, an episode that hurt the agency’s standing, and said flatly that “a lot of people don’t trust BLS anymore.”
The leadership turnover is part of that story too. The prior administration fired BLS Commissioner Erika McEntarfer without evidence that data had been manipulated, drawing warnings from economists across the political spectrum. The Senate has since confirmed economist Brett Matsumoto as commissioner, and he’s said decisions should be “driven by science rather than politics.”
My practical advice for anyone making hiring calls right now: stop waiting for a clean national number on AI and job loss. It isn’t coming this year, and it may not come next year either. Watch your own requisition volume, your own time-to-fill, and which roles your managers quietly stop asking you to backfill. Those numbers are your leading indicator, and they’re available to you today, not on a government reporting schedule.
Sonderling also mentioned 530,000 active registered apprentices, which he called about halfway toward the administration’s apprenticeship goal, though the article didn’t state the actual target. Apprenticeship works well in the trades and has never scaled the same way in white-collar work, which happens to be exactly where AI is hitting entry-level roles hardest. Halfway to a goal doesn’t tell you whether those programs sit in the fields where the displacement is actually happening.
Construction Owners Keep Hiring Even With Earnings Sliding
NFIB’s quarterly industry survey, taken in July, shows optimism improved across all four sectors it tracks after a rough April. The overall Optimism Index landed at 99.8. Manufacturing had the best quarter by a wide margin, with its Optimism Index jumping 6.3 points to 104.3, the largest gain and the highest reading of any industry NFIB tracks.
Construction is where the real hiring story sits. 53% of construction owners reported unfilled job openings, up 7 points from April and the highest of any industry, running 17 points above the level for all firms combined. Even with that many roles sitting empty, construction owners want to hire more: their plans to increase employment rose 9 points to a net 28%, the second highest of the four industries. At the same time, their earnings trends fell 8 points to a net negative 19%, wiping out the improvement they’d built between January and April.
Hiring plans rising while earnings fall is the combination I’d watch. Owners hiring into shrinking margins are usually betting on committed backlog rather than current profitability, and if that work slips, hiring plans get cut fast.
If you’re staffing for the trades right now, the data backs up that word of mouth alone is already losing the race for trade talent, and a tighter labor pool only raises the stakes on how you source. Working with a staffing partner that already has trades candidates in the pipeline closes that gap faster than posting and waiting.
Retail was the weakest of the four industries. Its Optimism Index rose just 0.7 points to 94.8, below its 96.1 historical average, and expectations for real sales fell 16 points to a net negative 7%, even as retail job openings climbed 11 points to 37%. Services had the biggest jump in expected business conditions, up 13 points, but it’s the only industry where hiring plans moved backward, falling 8 points to a net 10%. Since services employs the largest share of the private workforce, that pullback matters more than the headline optimism number suggests.
Across all four industries, 63% of owners reported supply chain disruptions, down just a point from April. The number has barely moved in months, which tells me owners have stopped treating disruption as temporary and started pricing and staffing around it as the new normal. “Small business optimism improved in all industries after April’s slump,” said Holly Wade, executive director of the NFIB Research Center. “While most metrics improved, many owners are reporting supply chain disruptions and difficulty finding qualified employees. Despite some challenges, Main Street is feeling more confident about economic conditions and the health of their business.”
If AI is already reshaping which entry-level roles you’re posting, or if construction and skilled-trade openings are sitting empty longer than they should, 4 Corner Resources can help you find the people who fill them. And if you’re the one navigating this shift as a candidate, especially early in your career, browse our current openings and let’s find the role that gets you in the door.
