The Entry Point Moved

The Class of 2026 is graduating into something their advisors did not fully prepare them for. Not a recession, not a tight market in the traditional sense. Something more structurally permanent: the entry point to white-collar work has moved.

This week, a few pieces of news landed in quick succession that, taken together, tell a remarkably coherent story. Dario Amodei of Anthropic told the New York Times that up to half of all entry-level white-collar jobs could dissolve within five years. Separately, Yale’s School of Management published research showing that AI job destruction is hitting before careers can even start, with one striking finding buried in the piece: computer science majors are now having more difficulty finding jobs than humanities majors. Meanwhile, Business Insider tallied a growing list of major companies, including IBM and Coinbase, that have publicly attributed layoffs to AI-driven efficiency gains. And the U.S. Department of Labor, apparently reading the same headlines the rest of us are, launched a new initiative to build AI skills through registered apprenticeship programs. Across town, CNBC reported that even small businesses, long a reliable on-ramp for new graduates, are seeing AI reduce entry-level headcount. Drexel University’s College of Business found an expected decline in small-business hiring specifically for the Class of 2026.

This is a lot to absorb. Let me try to make sense of it.

The Grind Was the Gateway

Here is the thing nobody wants to say plainly at a May graduation ceremony: for decades, the entry-level grind was not just a rite of passage. It was the actual mechanism by which people entered the professional world. You did the boring, repetitive, low-judgment work. In return, you got proximity to decisions, mentors, and problems. You learned the business by doing its least glamorous tasks. The grind was the bridge.

AI has automated the bridge.

The data entry, the research compilations, the first drafts of routine documents, the basic financial modeling, the report formatting: these are not glamorous, but they were how someone went from “new hire” to “trusted contributor.” When AI absorbs those tasks, the scaffolding disappears. You can’t climb a ladder that’s been removed.

This is why the Yale finding about CS majors is so counterintuitive on the surface, and so logical underneath it. If you trained for the tasks AI is best at doing, AI is your most direct competition. If you trained to think, argue, connect ideas, and navigate ambiguity, you’re in a different race entirely.

What Survives

The four things AI cannot reliably do, at least not in the way humans do them, are exactly what I call the 4Is in my book: interpretation, integration, interpersonal connection, and imagination.

Interpretation is the ability to take data or information and make it mean something in a specific context, for a specific audience, at a specific moment. AI can summarize. It cannot tell you what this particular client needs to hear on this particular Tuesday given everything you know about them and their organization. That judgment is human.

Integration is the capacity to combine knowledge across domains in ways that generate novel insight. A marketing student who can connect behavioral economics to supply chain disruption to geopolitical risk is not doing something AI does well. Synthesis across messy, poorly-defined domains remains stubbornly human.

Interpersonal connection is, frankly, the one nobody in business school takes seriously enough. The ability to build trust, read a room, repair a relationship, and make someone feel genuinely understood: this is not soft. It is the most durable professional skill there is. And it atrophies if you spend four years writing essays for an AI to check instead of defending your ideas in front of people who disagree with you.

Imagination covers the full range of creative and strategic thinking: reframing problems, seeing possibilities that aren’t obvious, designing things that haven’t existed before. AI is extraordinarily good at variations on known patterns. It is weak at genuine novelty.

These are not abstract virtues. They are job requirements for what comes after the grind disappears.

The Government Noticed. Sort Of.

The Department of Labor’s new AI apprenticeship initiative is encouraging in intent and probably insufficient in speed. Apprenticeships are valuable. Hands-on, employer-connected, skill-focused learning is exactly what the moment calls for. But the initiative is being built for the job market of 2024, not 2028. The roles it is preparing workers for, AI oversight, process redesign, data infrastructure, are real, but they too will evolve faster than any apprenticeship curriculum can track.

The Yale researchers made a related point about universities: faculty are building AI courses that become outdated before the semester ends, and in some cases, students already know more than their professors do. Symbolic curriculum overhauls are not the same as genuine preparation.

What would genuine preparation look like? Less emphasis on tools, more emphasis on the four skills that tools cannot replace. Less optimization for the tasks that fit neatly into a syllabus, more practice doing things that require judgment, relationships, and creativity under real conditions.

The Honest Conversation

The NYT opinion piece asked whether the AI job apocalypse would actually happen. The honest answer is: it already has, at the bottom of the career ladder. Whether it continues to climb is the open question. What is not open is the fact that the on-ramp has changed shape, and no amount of campus recruiting optimism will change that for the Class of 2026.

The students who will be fine are the ones who understand that the grind, as it existed for the previous three generations of white-collar workers, is not coming back. The skills that survive automation are not the ones most easily measured in a classroom. They are the ones most developed in friction: in conversation, in collaboration, in the discomfort of genuine creative and ethical judgment.

The entry point moved. The question is whether higher education is willing to acknowledge that clearly enough to actually help.

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.