@danshipper: YouTube: https://youtube.com/watch?v=OCEVqy8kl7Q&feature=youtu.be… Spotify: https://open.spotify.com/episode/62NJTryUh6…

X AI KOLs Following News

Summary

In an interview, GitHub COO Kyle Dagel discussed the impact of AI coding agents: expansion of the customer base to non-developers, 17 million agent-created PRs per month, surging pressure on open-source maintainers, and the evolution of business models from fixed licenses to usage-based pricing.

YouTube: https://youtube.com/watch?v=OCEVqy8kl7Q&feature=youtu.be… Spotify: https://open.spotify.com/episode/62NJTryUh6D8idheRZJm0e…
Original Article
View Cached Full Text

Cached at: 06/17/26, 07:59 PM

YouTube: https://youtube.com/watch?v=OCEVqy8kl7Q&feature=youtu.be… Spotify: https://open.spotify.com/episode/62NJTryUh6D8idheRZJm0e…


TL;DR

GitHub COO Kyle Dagel discusses with Every’s Mike Taylor the impact of coding agents: the customer base expanding beyond developers, an explosion in PR volume (17 million agent-created PRs per month), surging pressure on open source maintainers, and the business model shifting from fixed licenses to usage-based pricing.

The Customer Base Is Growing: Not Just Traditional Developers

Kyle points out that GitHub has always defined “developer” broadly. He himself initially wrote code just to pay for art school tuition before becoming a professional. Today, beyond large enterprise teams, GitHub Copilot App’s daily active users include many “knowledge workers”—legal, finance professionals, and hobbyists. They use AI to build small apps or assets. GitHub’s goal is to make it easier for more people to try coding and ensure there’s always an entry point into the world of software development.

Pressure on Open Source Maintainers: The PR Flood and Control Tools

Mike mentions that many open source maintainers say they’re “drowning.” Kyle believes better review tools are needed. GitHub is building an agent-capable Copilot code review that can find novel vulnerabilities and allow users to comment directly to let the agent make changes. Additionally, “agent merge” functionality lets a PR automatically wait for CI and policy checks before merging after review.

For open source-specific issues—not being able to control who submits what—GitHub’s strategy is to provide basic control modules to maintainers without presetting standards. For example, the guarantee system Mitchell Hashimoto shared: GitHub won’t force everyone to use it, but allows the community to evolve standards organically before locking them in.

PR Volume Explosion: 17 Million Agent-Created PRs per Month

Kyle reveals that activity on GitHub is far beyond previous levels. In all of 2023, GitHub saw 1 billion commits. If growth were linear, 2024 would hit 14 billion commits, but actually growth is nonlinear. In March 2024 alone, 17 million PRs were created by agents. These aren’t “junk code” but the result of climbing back up after early adoption. “When not just Kyle is building, but Kyle plus N agents, what can we build?” Kyle believes all code ultimately flows to GitHub, so the company must invest heavily to support everyone’s “agent moment.”

Business Model Evolution: From Licenses to Usage

Kyle admits the final model isn’t clear yet. Currently, users have a free tier and API rate limits. But agents can work while users sleep, so a fixed license doesn’t make much sense. The goal is to accommodate scenarios like “150 agents running simultaneously” while maintaining the core experience. GitHub historically expanded from free public repositories to free private repositories, and will adjust based on community needs. He cites Peter Steinberger’s example (running 150 agents simultaneously) and says they’ll empower users to do more, but need a baseline of agent usage.

GitHub and Microsoft Integration: Dual Role and Developer First

Kyle has been at GitHub for 13 years and now also serves as Microsoft’s Chief Marketing Officer for Developers. He emphasizes that GitHub always “builds tools for developers, enterprises buy them”—but 100% designed for developers. As COO and CMO, he is responsible for integrating all of Microsoft’s developer tools into a unified experience. For example, the 2024 Build conference invited external community members (like Peter Steinberger and someone from Swixs) to the main stage for the first time—intentionally so. “Software development is a team sport; no single company can answer all questions.”

Competition and Differentiation: Standing for Developer Choice

Facing intense competition, GitHub’s foundation is “caring about developer choice.” Kyle believes the industry is shifting from the open API era to “walled gardens” and “mousetraps.” GitHub always wants to help developers use other tools and simplify processes with all partners. Its superpower is covering the entire software building process—from code generation to collaborative review. Therefore, GitHub invests in its own models (like the new Microsoft AI model) while continuing to partner with Anthropic, OpenAI, and Google. “Choice is core; we will never give that up.”

Internal Decision-Making: Eat Your Own Dog Food, But Embrace External Tools

Kyle says the entire GitHub team uses various tools, otherwise they’d lose touch. Decisions require using the company’s own products (like the Copilot desktop app) while also allowing developers to try other tools. This balance is key to staying competitive.


Source: 《@danshipper: YouTube: https://youtube.com/watch?v=OCEVqy8kl7Q&feature=youtu.be… Spotify: https://open.spotify.com/episode/62NJTryUh6…》

Similar Articles

@danshipper: GitHub has a front-row seat to how code is changing now that everyone—and their army of agents—can ship code. In March …

X AI KOLs Following

GitHub COO Kyle Daigle discusses the surge in AI agent-created pull requests (17 million in March) and the projected 14 billion commits this year, emphasizing open-source maintainer control, usage-based pricing over per-seat licensing, and the collapsing developer/non-developer distinction as AI tools enable non-developers to build apps.

Win through AI powered products: Conor Spicer, Solutions Engineer, OpenAI

YouTube AI Channels

OpenAI Solutions Engineer Conor Spicer introduced Codeex, an AI coding agent that automates the entire software development lifecycle, significantly improving code delivery speed and security review capabilities, and demonstrated an end-to-end workflow from requirements to deployment.

- YouTube

YouTube AI Channels

The article argues that AI enables companies to evolve from rigid hierarchies into self-improving recursive loops, where token usage replaces headcount and humans handle only high-risk, high-touch interactions.

@hnshah: https://x.com/hnshah/status/2066761276945211442

X AI KOLs Timeline

Hiten Shah reflects on how AI is transforming GitHub from a repository of evidence into an environment where non-coders can apply product judgment directly to software workflows, bridging the gap between customer understanding and code.