Flat Engineering Teams Aren't an Org Chart Decision. They're a Maturity Filter

Meta built an AI engineering team with 50 engineers per manager. The Wall Street Journal called it "ultra-flat." André Spicer, from Bayes Business School, was more direct: "it's going to end in tragedy." Both are wrong for the same reason: they treat flat team structure as an org chart decision.

Flat team structure is not an org chart decision. It's a cultural consequence.

And most companies cutting middle management right now aren't going flat. They're going broken with extra steps.

What changed since 2022

Pre-2024, "flat" was a cultural manifesto. Valve published a handbook in 2012 describing in detail how to operate without hierarchy. 37signals (Basecamp) published "It Doesn't Have to Be Crazy at Work" in 2018, defending calm and autonomy as competitive advantages. GitHub grew for years under a radically flat structure - until it hit a ceiling and had to add layers to scale. It was a philosophical position, with skin in the game and people publishing exactly why it was worth defending.

In 2026, "flat" turned into a spreadsheet line. Gallup reported that the average number of direct reports per manager in the US went up 50% since 2013, with most of the jump in the last 18 months. That's not a coincidence. It's the same window in which CFOs figured out that cutting middle management makes next quarter's numbers look good.

Meta at 50:1 isn't pioneering. It's the extreme end of a curve being pushed in the same direction for the wrong reason.

How I know it works

At Buser there is no middle management inside the engineering team. ICs report directly to me. I'm CTO and CPO at the same time, hands-on, making technical and product trade-off decisions in the same meeting. The team ships consistently. Operating margin is growing. Efficiency improves quarter over quarter.

It works. But it works because I'm paying a price probably nobody at Meta is paying: a very specific contract about who joins the team.

5 things that make flat teams possible (and most companies don't have)

1. A team that understands it's made of adults. I don't explain priority three times. I don't review PRs to make sure someone thought about the edge case. I don't remind people to test before shipping. If I have to do that, the flat team dies in the first week - not because the structure is bad, but because the team isn't ready. Maturity is not seniority. It's the ability to operate without anyone looking over your shoulder.

2. Explicit priority contract, not implicit. Every week the team knows what matters and why. There is no "let me figure out what Avelino wants." There is "this moves the number, this can wait, this is debt we'll pay down in two sprints." Without that, a flat team becomes noise - everyone deciding local priority without knowing if it ladders up to the global one.

3. Decisions without permission. If ICs need approval to choose architecture, library, feature flag strategy, database schema - the flat team is theatre. At Buser, technical decisions belong to the IC. If they got it wrong, we talk, adjust, move on. If they got it right, nobody hands out trophies - that's the job. A manager existing to "approve technical decisions" is the number one symptom of an organization pretending to be flat.

4. Distributed accountability, not just distributed autonomy. This is the point that catches people trying to copy flat. Autonomy without accountability is chaos. At Buser, the IC doesn't just decide - the IC also owns what they decided. Shipped a change that broke things? They go figure out why, fix it, write the postmortem. There is no "manager takes the heat." There is "you own what you shipped." This filters out a lot of people who say they want flat but actually want decisions without consequences.

5. Hiring criteria that filter for maturity, not years of experience. I have people with 5 years in the industry who outperform staff engineers from big tech. I have people with 15 years who wouldn't survive 30 days in a structure where no manager chases them. Maturity doesn't come from time. It comes from having operated without a safety net before. This is the hardest filter to apply and the most expensive to get wrong.

Where Meta will fail

Meta has a huge team, a vague mission (superintelligence is a project, not a scope), and a 50:1 ratio without any of the 5 things above figured out. There's no explicit priority contract - it's exploratory AI. There's no distributed accountability - it's big tech, accountability turns into politics. There's no maturity filter - they hired in volume.

It's going to play out exactly like Spicer said: tragedy. The loudest people will consume the manager's time. Juniors will disappear from the map. The best ICs will leave within 12 months because nobody saw their work. Then Meta will rehire managers and call it a "post-learning adjustment," and the cycle restarts.

Falsifiable prediction: more than 30% turnover on this specific team by March 2027.

The ethical point nobody talks about

A flat team without maturity isn't just inefficient. It's abusive.

You're taking people who aren't ready to operate without direction and dropping them into an environment where nobody will give direction. They will feel like they failed. They'll leave thinking they were the problem. They aren't. The problem is the company that sold autonomy to someone who needed mentorship.

That's why a flat team is a hard call, not a progressive one. You're telling a lot of good people to their face: "this isn't your place right now." That's honest. It's not kind. And anyone treating it as kindness is lying.

Where this leaves me in 2026

Buser is growing, hiring, and the filter stays the same. I'm not looking for senior. I'm looking for mature. People who operate without me having to play technical babysitter. Who ship without asking permission. Who own what they shipped.

It's a rare profile. And it's getting rarer as big tech trains an entire generation to operate with manager-coach-mentor-skip-level-staff-engineer-distinguished-VP between them and any actual decision. People who leave that model feel the void. Most run back. The ones who stay and thrive are exactly the profile Buser needs.

I've written before that the CTO's job changed but the chair didn't. This post is the complement: what changed for everyone on the team, not just the CTO. Cost discipline matters, but cultural discipline matters more. A company that cuts managers to show margin without first building maturity is taking on structural debt more expensive than any cloud line item.

Honest question before you close this tab: if your company needs a manager to function, is it the manager that's missing - or is it the team that isn't ready yet?

The answer says more about who you hired than about who you promoted.