The management ladder was always a workaround

We built career paths that rewarded coordination over craft. AI is making that a question worth asking again, and Founders are the ones who get to answer it first.

On March 31, Jack Dorsey and Sequoia’s Roelof Botha published an essay that caused a quiet stir among those who pay attention to such things. It was called “From Hierarchy to Intelligence,” and its central argument was blunt: the corporate org chart is a 2,000-year-old information-routing system inherited from the Roman Army, and AI has made the messenger redundant. Dorsey had cut roughly 40% of Block’s workforce weeks earlier and framed it not as downsizing but as a permanent restructuring. The stock went up 20%.

The essay is worth reading in full, but the part that stayed with me wasn’t the grand architectural argument, it was the three roles Dorsey and Botha said would remain:

• individual contributors who do the deep specialist work

• directly responsible individuals who own specific outcomes

• and player-coaches who develop people while staying close to the work themselves.

Note the absence of a pure management layer, or someone whose job is purely coordination.

I had a version of this problem twenty years before they named it. The difference is that back then, I lost.

At an agency a decade ago, I had someone who ran campaign trafficking. If you haven’t worked in digital advertising, trafficking is the precise technical work that sits between a campaign being sold and a campaign going live: ad tags, placements, specs, deadlines, the plumbing that nobody notices until it breaks. Most people find it invisible or tedious, but he found it genuinely interesting, and the gap between him and the next-best person at it was not small.

He was the kind of specialist you can quietly anchor a team around. When something was broken at 11pm before a major launch, he was the call. When a new format arrived that nobody had seen before, he had it mapped before the brief had finished circulating. By any honest measure, he was doing senior-level work and deserved to be recognized as a senior.

I could not promote him without giving him a team.

That was the rule, real and enforced: seniority meant management and title meant headcount. The only lever available for rewarding exceptional individual output was to move that person away from the individual output that made them exceptional.

I tried to work around it, but I couldn’t. He eventually left after I did, moved through several agencies that were willing to think differently, and became a genuine reference point in his field, the person other specialists called when something was genuinely hard. He got there by being allowed to keep doing the work he was exceptional at, rather than being asked to stop doing it as a reward.

Then at Chango, building the revenue team that reached $60M ARR in 3 years, I ran into the same wall with different people. We had serious, well-respected investors on the cap table, and the norm was clear: promotion means management. I didn’t have the conviction or the institutional air cover to argue otherwise, and even if I had, the pushback would have been real. At the time this wasn’t irrationality on anyone’s part, it was the model that had worked at dozens of other companies, backed by people who had watched it work, and suggesting an alternative path for senior ICs was the kind of structural experiment that required both evidence and permission.

Neither existed in the form I needed them, and I lacked the personal gravity to make it so.

So some of the best individual performers on that team got teams. A couple were fine at managing, though noticeably less fine than they had been at the individual work that made them worth promoting in the first place. One left the industry. The thing that made them exceptional was the thing we quietly asked them to step back from.

I want to be careful about the retrospective criticism, because it is too easy from where I’m sitting now. The management ladder wasn’t irrational or even incorrect because coordination was genuinely expensive before AI, and the manager as information router, aggregating context upward and relaying decisions downward, was a real multiplier. The title reflected something real.

The problem was that it became the only path, applied universally regardless of whether the work was a coordination problem or a craft problem. The trafficking specialist’s work was a craft problem. Some of my Chango salespeople’s work was a craft problem. There was no lever for craft, only for coordination, and so we converted craft people into coordinators and called it a promotion.

What has actually changed

The Dorsey/Botha essay matters not primarily because of what it argues, but because of who is arguing it and where. A founder who has built two S&P 500 companies, co-signing with Sequoia, publishing a structural manifesto weeks after a major workforce cut that the market rewarded. That is not a contrarian take anymore, it is a signal that the conversation has moved.

The norm was enforced by the people holding the capital. The reason founders didn’t fight harder for IC-first structures in 2012 or 2016 was less about the idea was wrong and more about the institutional consensus was against it, and the evidence base for an alternative didn’t exist in a form investors could pattern-match on. When your board has backed a model that worked at a dozen other companies, the burden of proof for doing something different is very high, and most founders reasonably decided the fight wasn’t worth it.

What AI has done is shift that burden. The coordination problem that made management layers necessary is now at least partially solvable by other means. The companies flattening their structures aren’t doing it as an ideological experiment; they’re doing it because the tools finally make it viable, and investors are treating it as innovation rather than deviance. The experiments that would have cost you credibility in a board meeting five years ago are now the experiments your board is asking you to run.

That is a genuinely different environment to build in, and it means the question of how you reward your best people, whether the management ladder is the only available path, is now actually askable in a room full of investors without the conversation dying immediately.

The questions worth sitting with

Dorsey’s prescription, eliminating the management layer wholesale and replacing it with AI-driven coordination, is more radical than what most early-stage Founders need to act on right now, but the underlying question is one every Founder building a team should sit with.

Roughly 70% of companies still have no clear advancement path for individual contributors beyond a certain level. That gap is where the best craft people quietly start looking elsewhere, not because they’re focused on titles, but because the signal they receive is that their kind of contribution has a ceiling, and the only way through it is to become something they’re not.

Who on your team are you about to promote because you don’t have another lever? Is that person a coordinator by nature or a craftsperson? And if they’re the latter, what are you actually getting by giving them a team, beyond a title change that makes sense on an org chart?

What does seniority mean in your company if management isn’t the only signal? Output per dollar of human time is one frame. The leverage ratio, how much a single person produces relative to what they cost, especially when that person is directing agents rather than supervising humans, is another. Neither of these is how most companies currently measure senior contribution, but both map more honestly onto what an AI-era team actually looks like in practice.

The person who can run a complex workflow, quality-check agent output, raise the bar on what comes out the other side, and compress what used to take a small team into something one person controls is doing genuinely senior work. Whether that role fits the inherited ladder is a separate question from whether it’s worth a great deal.

The management ladder solved for coordination at a time when coordination was the expensive problem. AI is solving for coordination in ways that are faster and cheaper than a human middle layer. That doesn’t make managers obsolete, but it does mean the calculus of when you need a coordinator versus when you need a craftsperson is shifting, and the old default of promotion-means-management deserves to be examined rather than inherited.

I think about the trafficking specialist from that agency sometimes. He became what he was always going to become, just at a different company, inside a structure that let him. The question for any Founder building a team right now is whether you’re building the structure that keeps your version of him, or the one that eventually loses him to someone who will.

Dax is the Co-Founder & CEO @ FOMO.ai, and the author of 84Futures.com.

👉 Founder or operator? You can now schedule a 1:1 session with me.