The Vanishing First Rung

The data this week is hard to look away from.

Tech companies eliminated nearly 60,000 jobs in the first three months of 2026. Amazon, Block, and Atlassian led the wave. Block’s CEO Jack Dorsey was notably candid in his memo to employees: “This is not driven by financial difficulty, but by the growing capability of AI tools to perform a wider range of tasks.” Amazon posted $716.9 billion in revenue last year, a record, and still cut 16,000 jobs. Atlassian’s cloud revenue grew 26 percent year over year, and still eliminated 1,600 positions.

Let that sink in: record profits, record cuts. The grind is not being outsourced to cheaper labor markets this time. It is being absorbed by machines.

Goldman Sachs economists estimated that AI was responsible for 5,000 to 10,000 net job losses per month in the most exposed U.S. industries last year alone. A survey by Challenger, Gray and Christmas tied AI directly to 7 percent of all planned U.S. layoffs in January. These are not projections. This is the present tense.

And then there is what BlackRock CEO Larry Fink said at his firm’s Infrastructure Summit this week. He is worried about the class of 2026. Not because the economy is collapsing, but because it isn’t. The economy is doing fine. The entry-level job market, however, is a different story entirely. The unemployment rate for recent graduates aged 22 to 27 stands at 5.6 percent, near decade highs outside the pandemic. Job postings targeting students on Handshake fell 16 percent between 2024 and 2025. Applications per role rose 26 percent. Fewer seats, more contestants.

The structural explanation is not complicated. For decades, organizations were shaped like pyramids. A wide base of recent graduates handled the routine cognitive work: summarizing data, preparing presentations, answering customer queries, writing first drafts. Generative AI has quietly dismantled that base. The pyramid is becoming a diamond. Thinner at the bottom, wider in the middle, still narrow at the top. The entry point is disappearing.

This is the part where most commentators offer one of two takes. Take one: AI is just another wave of automation and new jobs will emerge, as they always have. Take two: this time is different and we are all doomed. I think both miss the more important point.

The grind was never the goal. It was the price of admission.

Young professionals did not spend two years summarizing research reports because summarizing research reports was a valuable end in itself. They did it because that is how you learned what good output looked like. That is how you built a mental model of the industry. That is how you earned enough credibility to eventually sit in the room where the real decisions happened. The grind was the apprenticeship structure, dressed up in corporate clothing.

AI has now taken the grind. Which means the apprenticeship structure is broken. And no one has fully replaced it yet.

Here is what I think higher education and students need to understand clearly: the solution is not to fight the trend. The solution is to rebuild the apprenticeship structure around what AI cannot do.

This is where the 4I model from my book becomes urgent rather than theoretical. AI is exceptional at handling interpretation of data once the questions are defined, at integration of existing information at scale, at executing tasks with clear parameters. What it cannot replace is the human judgment that decides which questions to ask in the first place. What it cannot replicate is the interpersonal trust built between a mentor and a junior colleague across a hundred small interactions. What it cannot manufacture is the imagination that sees a market nobody else has noticed, or the creative leap that reframes a problem in a way no training dataset would ever produce.

The entry-level jobs that are disappearing were never really teaching those skills anyway. They were building prerequisites: stamina, attention to detail, familiarity with industry vocabulary, the ability to follow instructions with care. Useful prerequisites. But prerequisites, not destinations.

The question universities and businesses both need to answer is this: if the grind is gone, how do we build those prerequisites in a different way? How do we give young professionals the pattern recognition and contextual knowledge they need, without sending them through two years of work that a language model can now do in seconds?

Some of the answers are already emerging. Simulation-based learning. Project-based curricula with real stakeholders. AI-assisted work where the student is responsible for judgment and quality control, not just execution. Mentorship that focuses explicitly on the interpretive and interpersonal layers of professional work, not just the technical ones.

BlackRock’s Larry Fink, to his credit, is not just warning. He is also investing: $100 million toward skilled-trade training, recognizing that certain kinds of hands-on, judgment-intensive work remains stubbornly human. That is a useful signal, even if it lands oddly coming from the world’s largest asset manager.

The 59,000 tech jobs eliminated this quarter are real losses for real people, and the pain in those numbers should not be abstracted away. But they also represent something that needed to happen eventually: the surfacing of a structural truth that was always there. The grind was never where the value lived. It was just where we made people prove they wanted it badly enough.

Now that the grind is gone, we have to figure out how to teach the rest.

That is the actual work. And there is no algorithm for it yet.

Andrew Perkins is the author of After the Grind: Rethinking Your Business Career in the Age of Artificial Intelligence and Robotics and Chair of the Department of Marketing and International Business at Washington State University’s Carson College of Business.