The Parchment Anchor

It has long been a staple of the human experience that one spends four years in a state of semi-permanent sleep deprivation, punctuated by the occasional lecture on 17th-century macroeconomics, in exchange for a piece of paper that guarantees a desk, a swivel chair, and the right to be ignored by a middle manager named Gary. However, according to Bill McDermott, the CEO of ServiceNow, this venerable social contract is currently being shredded by a series of very polite and extremely efficient algorithms.

McDermott recently observed that the rise of autonomous AI agents could send college graduate unemployment soaring past the thirty percent mark. This is a rather startling figure, suggesting that nearly one in three graduates will find themselves with a degree in hand and absolutely nowhere to put it, other than perhaps on the wall of a very crowded coffee shop where they are now overqualified to steam milk.

(I once knew a man who spent six years studying the migratory patterns of the common garden gnome, only to find that the market for gnome-tracking was, quite literally, non-existent. He eventually found work as a professional queue-stander, which he claimed was remarkably similar to academia, only with better weather.)

The problem, it seems, is that the entry-level roles—the ones involving the sorting of spreadsheets, the drafting of basic memos, and the general fetching of digital water—are precisely the tasks that AI agents perform with a level of enthusiasm that no human intern could ever hope to replicate. An AI agent does not require a lunch break, it does not have an existential crisis in the middle of a Tuesday afternoon, and it certainly does not spend forty-five minutes trying to figure out how the communal printer works.

From a corporate perspective, the transition is entirely logical. Why hire a junior associate who needs to be taught what a 'pivot table' is, when you can deploy a script that already knows every pivot table that has ever existed, and several that haven't? The result is a workforce that is becoming increasingly top-heavy, like a pyramid built by someone who forgot that the bottom part is generally considered the most important bit for structural integrity.

Universities, meanwhile, find themselves in the awkward position of selling a product that is rapidly becoming a legacy system. It is a bit like being the world's premier manufacturer of high-quality buggy whips just as the first Model T rolls off the assembly line. One can admire the craftsmanship of the whip, but it doesn't change the fact that the horse has been replaced by an internal combustion engine that doesn't need to be fed oats.

There is a certain whimsical irony in the fact that we have spent decades building machines to do the 'boring' work so that humans can focus on 'creative' pursuits, only to find that the machines are now quite good at the creative bits too, leaving the humans to focus on the increasingly difficult task of finding a reason to get out of bed in the morning. The 'entry-level' role was always a form of apprenticeship, a way to learn the ropes before being allowed to pull them. If the ropes are now being pulled by a cloud-based entity with a response time of three milliseconds, the ropes themselves may eventually become obsolete.

(I suspect that in the future, the most valuable degree will be in 'Advanced Human Simulation,' a course designed to teach people how to blink at appropriate intervals and make small talk about the weather so that the AI agents don't suspect we've been replaced by something even more efficient.)

As we move toward this thirty-percent-unemployment horizon, the definition of 'value' in the labor market is undergoing a surreal transformation. We are entering an era where the most important skill a graduate can possess is the ability to do something that an algorithm finds fundamentally confusing. This might include things like 'nuanced moral ambiguity' or 'the ability to appreciate a particularly subtle pun,' neither of which, unfortunately, are currently listed as core competencies in the standard business administration curriculum.

For now, the Class of 2026 may want to consider diversifying their portfolios. A degree in computer science is all well and good, but a secondary qualification in 'Manual Hedge Trimming' or 'Artisanal Stone Walling' might be the only thing standing between a graduate and a very long, very quiet retirement at the age of twenty-two. After all, an AI agent can write a thousand lines of code in the time it takes you to sneeze, but it still hasn't figured out how to properly prune a rhododendron without causing a minor diplomatic incident.