In 1776, Adam Smith opened his masterwork, An Inquiry into the Nature and Causes of the Wealth of Nations, with a pin factory, and his little calculation became famous.

He shows that where a lone worker can barely make twenty pins a day, ten specialized workers turn out forty-eight thousand. The division of labor had just laid the foundations of two and a half centuries of industrial capitalism.

AI does the exact opposite, and that is the paradox we are living through. For the first time since Smith, a major technology no longer fragments tasks but recomposes them inside a single person.

This person already has a name. It is the HI-C, the high-impact individual contributor Elena Verna describes at Lovable, or the Member of Technical Staff popularized by Anthropic, two names for one concept, an augmented individual who pulls back into one pair of hands what several teams coordinated yesterday. It is the craftsman resurrected, a white-collar worker in cybernetic form.

Behind this augmentation sits the compression of white-collar work. It does not target code alone but reaches design, marketing, analytics, legal, product, and finance in turn. Becoming an average contributor in almost any field adjacent to your own now costs close to nothing, and that average, added to the craft you actually excel at, is enough to ship an entire project.

The problem is that the entire managerial scaffolding was built to handle that fragmentation. So if the value chain now fits inside one person, it is fair to ask what the chain of command is still for.

That reminds us that the Peter principle and the Dilbert principle already described the same bug, since we traditionally promote good craftspeople into roles where they stop being good. The root cause goes back to Smith, because the division of labor made coordination so expensive that producing coordination became a job in itself, better paid than producing value, and our whole economy still runs on that arrangement.

Yet agentic AI breaks exactly that dynamic. The cost of human alignment, the PRDs, the committees, and the daily meetings become the new bottlenecks, and the cost of coordination will soon be the anomaly to fix, because it will far exceed the simple cost of production. The middle manager will no longer arbitrate anything useful and will only slow a decision the machine could already have executed, so people will route around them and that intermediate layer will go obsolete.

Elena Verna shows it without hedging, since her growth team at Lovable was dissolved, she went back to being an IC, and she now ships alone the projects that used to occupy a whole team, while still being paid like a VP.

So of course, in that kind of structure, all of this is obvious, because it is an AI-native scale-up with a nearly flat hierarchy and software at the center of the work, the ideal terrain. The real question is how a bank, a government agency, or a several-thousand-person manufacturer will host this profile, and I see three locks standing in front of these HI-Cs, more anthropological than organizational.

The first lock is the compensation grid, indexed for forty years on span of control. Paying an IC like a VP blows up the entire HR system, that famous department in charge of human “resources”.

The second is the immune system of middle management, which was never designed to coordinate but to exist, and which will reject the graft through a thousand quiet gatekeepings, whether by blocking data access, stacking validation committees, or refining approval processes.

The last, and the most stubborn, is status. The title is still a social currency, and going back to IC stays symbolically violent even at equal pay. Why? Because the pyramid is not only an org chart, it is also a collective story.

This is why the HI-C will not fight to transform the pyramid, they will leave it, burn out, or worse. Large organizations that fail to welcome these profiles may become the employer of last resort for craft talent, the way some public administrations did against the private sector, while those that adapt will gain a decade of lead.

The fork may already be underway. On one side, hyper-productive contributors, autonomous, well paid, and mobile; on the other, coordinators sliding into obsolescence, stuck in structures that cannot reform themselves.

But let’s go back to Adam Smith to close. He promised us abundance through specialization, AI promises us abundance through recomposition, and between the two promises stand two and a half centuries of institutions, careers, professional pride, and a middle class built on the org chart.

Managerial techniques will have had their share of misery, and I personally will feel no regret watching them challenged, or simply reinvented.

That, I believe, is where the real political problem of the decade is being decided.