You Can Generate Code Without Juniors. You Cannot Generate the People Who Can Tell When It Is Wrong.

In 1999 I wanted nothing to do with COBOL. Green screens, mainframes, and a remediation deadline that could not move and had no upside past January. Plenty of experienced developers felt differently, because the Y2K money was extraordinary, and they left the work they were doing to go chase it. I stayed where I was and watched the senior people around me walk out the door toward it.

What that produced for me was a seat. I joined AEI as a consultant on a global shipping application, on a development team where everyone else was senior, and I was treated as one of them from the start. Making that jump doubled my market value before I had proven anything. Then the project manager exited, and I ended up speaking for the team, because I was the only one willing to tell the client the project was not as far along as they thought.

I understood the work. What I did not understand was what the work meant: that I was operating at a level nobody had certified me for, and that the market was about to price me accordingly. I never once asked where the seat came from. I did what needed to be done, and I was in awe of the money I was depositing in the bank. Nobody in that building ever filed a line item for manufacturing my judgment, and I only understood it looking backward, years later.

The industry is now running that same experiment deliberately, from the other end. The math is not stupid: a junior engineer costs a salary, consumes senior review time, and produces output a model produces for free, so the requisition converts to a senior req or disappears. Every technical leader who has run that arithmetic reached a defensible answer, and not one of them thinks of themselves as the person who ended the pipeline. Underneath it sits a belief nobody states out loud: seniority is procurable. It is not. The senior engineers you plan to hire instead were built somewhere else, over about a decade, on somebody else's budget.

Here is what the ROI math misprices. The junior seat's primary output was never the code. In my case it was not close: the code I wrote at AEI mattered far less than the fact that I sat on a team of seniors, held to their standard, when the seat above me came open. The seat was not a classroom. It was a position, and cutting it removes the surface a capable person stands on when a vacuum opens above them.

Vacuums always open, because people leave, get promoted, and retire on a schedule nobody controls. Nothing here is failing; the structure is doing exactly what it was built to do. None of it shows up this quarter either, which is what makes it durable. Throughput holds, delivery holds, and the metrics confirm the decision, which is what makes the next one easier to sign. This is Creep: no face, no single bad call, just accumulation, where every concession looks reasonable in isolation and the damage is legible only as a photograph from a year ago held up next to today.

The bill arrives the first time a senior engineer departs and nobody is standing close enough to move into the work. A LeadershipOS™ Train Plan makes that person's reasoning survive their exit, and it cannot manufacture somebody to hand it to. Continuity with no successor is a document with no reader. This industry ran the experiment once already, and the result is not in dispute: employers stopped hiring COBOL developers at the junior level years before 1999, then discovered they needed senior COBOL developers against a deadline that would not move, and paid whatever the remaining supply cared to name.

The leaders who avoid repeating it will not be smarter or better funded than you. They are the ones who stopped pricing a junior seat as delivery throughput, which is the one thing about that seat that was never the point. Price it as succession surface instead. This is a Boundary Bearing, not a hiring preference: your delivery economics run on a quarterly clock, the industry's supply of judgment runs on a ten-year clock, both are legitimate, and the far side of that boundary does not report to you and will not change its tempo because you need it to. Your own side is the part you design.

Right now most organizations reading this are net consumers of seniority. They hire senior engineers they did not train, which has been effectively free, because somebody else paid to build them. I collected that same subsidy in 1999. I was given a seat, not a result, and what I did in it was mine, but I do not get to pretend I built the seat, and neither do you.

Run this against your own team today. Name the last person who operated above their title before anyone certified them for it. Not who you promoted; who was already doing the work when the paperwork caught up. If you cannot name one from the last twelve months, you are not manufacturing seniority, you are purchasing it, from a supply somebody else already decided to stop funding. Then name your floor out loud, in a room where it can be defended when the requisition math comes for it, because a floor nobody has heard is not a floor.

The Edge Case walks through how to map a Boundary Bearing before you cut the seat that runs through it: http://TheEdgeCaseBook.com

---

I write about structural leadership for technical leaders in high-stakes operating environments. If this way of thinking resonates, it runs deeper in The Edge Case: http://TheEdgeCaseBook.com