This editorial appeared in the June 5th, 2025, issue of the Topline newsletter.
Want all of the latest go-to-market insights without the wait? Subscribe to Topline and get the full newsletter, including bonus commentary and takeaways, delivered to your inbox every Thursday.
For the longest time, you could walk into any McKinsey office and you’d see the same thing: rows of fresh-faced analysts, hunched over laptops, building slide decks until their eyes burned. Torturous work that was justified by the prestige of being a McKinsey Consultant, and the career track it would activate.
Those decks are still being made. But those desks are empty.
McKinsey doesn't need armies of fresh grads anymore. As their CTO put it: "The technology can do that."
The technology — AI — has cratered the value of a new grad. The unemployment rate for recent university graduates just hit 5.8%, compared to 4.2% nationally.
Even the "lucky" ones who find work aren't so lucky. 41% are underemployed — pouring lattes or stocking shelves with their business degrees. Half of those who start their careers underemployed stay that way forever.
This is because AI does entry-level work better, faster, and cheaper than any 22-year-old ever could. Which raises the question every new grad should be asking: If there's no bottom rung on the ladder anymore, how the hell do I start climbing?
The answer starts with understanding how we got here.
The Deal is Broken
Universities teach you how things work. Companies teach you how to work. This is the system that used to work.
The first job was never glamorous — fetching coffee, formatting spreadsheets, fact-checking reports. But companies needed the grunt work done (and were willing to pay for it). That need created the bottom rung, and graduates grabbed it.
But AI does grunt work better than any human ever could. And every month, the list of "junior tasks" AI can handle gets longer. Every month, companies need fewer entry-level hires. As these models double in capability every 12-18 months, the trend only accelerates.
This year, 4 million students will graduate from U.S. universities. Half carry at least $29K in debt — a collective $58 billion bet on futures that might not exist. When 2 million debt-laden grads can't find real work, they can't buy homes or start families. Consumer spending drives 69% of GDP. Do the math.
The system that used to work, isn’t working anymore. Companies won't train juniors because AI is cheaper. Universities keep pumping out graduates who are no longer job ready. Students keep borrowing against salaries they'll never earn.
Fewer juniors hired today guarantees fewer seniors tomorrow. That shortage will drive up wages for experienced talent. Higher wages create more pressure to cut juniors. The cycle makes us all worse off.
When tractors replaced farmhands, we had 50 years to adjust. When computers killed secretarial pools, we had 20. AI's eating entry-level jobs in 2. Each disruption moves faster. Each one hurts more. The solution…that’s older than all of them.
The Answer is 200 Years Old
The usual tech optimists say "new jobs will emerge." Except this time is different. Past disruptions transformed entry-level work. AI eliminates the on-ramp.
Universities can't build a new one. They're designed to teach theory, not practice. Asking them to teach practical skills is like asking a library to teach swimming.
So where do we look? To 19th-century Germany.
When industrialization threatened the craft guilds, Germans created apprenticeships — young workers learned by doing, earned while learning, and emerged as masters.
IBM gets this. While other tech giants slash entry-level roles, IBM runs paid apprenticeships. 12-24 months. Software engineering, cybersecurity, data science. No degree required. Their retention rate? 96%.
But that's not all. If AI has made new grads less valuable, companies won't pay them what they used to. The math has to work.
Germany's done this for decades: apprentices start at 50% of standard wages, rise to 85%, then jump to full pay. The progression matches the productivity.
But four years of university, $29K in debt, then apprentice wages? Nobody's signing up for that.
They might not have to. See, AI will add at least $300 billion in annual U.S. tax revenue by 2035. That same AI can compress four-year degrees into two with personalized learning. The problem can be the solution.
The new system would be two years of AI-powered education (subsidized), then paid apprenticeships. Minimal debt, real skills, good wages by year three.
Remember those 2 million grads who can't buy houses or start families? This fixes that. They enter the workforce with skills, not debt. They spend money. They build lives. The economy thrives.
Companies get affordable talent. Young people get actual careers. The government's AI windfall funds the whole thing. Everyone wins — including the birth rate.
The blueprint exists. IBM proved apprenticeships work in tech. Germany proved they work at scale. We've got the money coming. We've got the technology ready.
We just need to admit the old system is dying, and build something better from its ashes.
