---
title: "The Org Chart Math Behind AI-Native Speed"
description: "Boris Cherny ships 20-30 PRs per day. A traditional engineer ships 3 per week. The difference isn't talent — it's communication overhead. Here's why AI-enabled startups are pulling ahead."
categories: ["startups","management","AI"]
date: 2026-03-10
lastmod: 2026-07-23
canonical_url: https://www.tomtunguz.com/communication-tax-small-orgs/
author: "Tomasz Tunguz"
---


> "Since last November, 100% of my code has been written by Claude Code. I have not manually edited a single line, shipping 10 to 30 PRs per day."

Boris Cherny, creator of Claude Code, ships 20-30 pull requests per day. Major code changes, not typo fixes. He runs five parallel AI instances, each on a separate branch.[^1]

Compare that to a traditional engineer : 3 PRs per week.[^2] Cherny isn't 10% more productive. He's 30x more productive.

That productivity gap compounds at the company level. Anthropic generates ~$5 million per employee.[^3] Cursor, $3.3 million. Midjourney, $2 million.[^4] Traditional SaaS considers $200-300k strong. A 10-20x difference.

One explanation : communication overhead. The math follows Metcalfe's Law.[^5] Each new team member adds n-1 new connections. Coordination drag doesn't grow linearly. It explodes.

Now consider what AI does to this equation.

A traditional 150-person organization runs four layers deep. The org chart creates 11,175 potential communication channels. Meetings multiply. Alignment decays.

An AI-enabled team producing equivalent output might need 30 people. Communication channels drop to 435. A 96% reduction.

{{< email_image src="s0pmjkgfnwo7ktgkvknh" alt="Communication channels in 150-person org vs 30-person AI-enabled org" width="540" height="377" >}}

This is one reason AI-native startups are pulling ahead, and why building AI companies feels fun. The advantage comes from organizational structure. Fewer humans, fewer channels, faster iteration, compounding speed.[^6]

R&D adopts this fastest. AI writes the code. Human communication becomes the bottleneck. The span of control debate shifts from "how many people can one manager oversee?" to "[how many AI agents can one human orchestrate?](https://tomtunguz.com/how-many-agents-can-you-manage/)"

Small teams have always paid less coordination tax. AI cuts it further.

[^1]: Cherny, Boris. [Claude Code creator landed 259 PRs in 30 days](https://news.ycombinator.com/item?id=46407967), Hacker News, 2025.
[^2]: Seporaitis, Julius. [What Can 75,000 Pull Requests Tell?](https://www.seporaitis.net/posts/2021/07/19/what-can-75000-pull-requests-tell/), 2021. Median developer opens 3 PRs per week; consistent with Google's internal data.
[^3]: Estimated from [Anthropic's ~$20B revenue run rate](https://www.bloomberg.com/news/articles/2026-03-03/anthropic-nears-20-billion-revenue-run-rate-amid-pentagon-feud) (Bloomberg, March 2026) divided by ~4,300 employees (LinkedIn).
[^4]: Dealroom estimates. [AI startups revenue per employee](https://x.com/dealroomco/status/1914264599505018989) : Cursor $3.3M, Midjourney $2M, OpenAI $1.5M per employee.
[^5]: [Metcalfe's Law](https://en.wikipedia.org/wiki/Metcalfe%27s_law), Wikipedia.
[^6]: [How to start a Lean, AI-Native Startup in 2025](https://henrythe9th.substack.com/p/how-to-start-a-lean-ai-native-startup), Henry the 9th, 2025.