Jun 2, 2026 · 1h 24m · latent-space
GitHub’s Agent Era: 14x Commits, 200M Developers, Copilot’s Next Act — Kyle Daigle
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
GitHub COO and Microsoft CMO of Developer Kyle Daigle joins Sean Wang to discuss how autonomous agents, ambient enterprise context, and architectural overhauls are redefining software creation for over 200 million developers. The conversation covers scaling through 14x commit growth, the evolution of GitHub Copilot into a full-lifecycle agent, and modern developer security paradigms.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. How this is scored →
The hosts as informed peer 4.3 Guest teaching 4.7 Guest disagreement 1.4 The hosts pushing back 2.0
speaking balance: gold is the hosts, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 44:10 Pushing back against developer gatekeeping
Daigle firmly rejects Sean's skepticism regarding GitHub's 200M developer metric, stating that splitting hairs over who qualifies as a real developer is counterproductive and exclusionary.
Hardest push from the hosts ▶ 44:00 Questioning developer account inflationSean directly challenges GitHub's headline user statistics, arguing that holding a GitHub account does not equate to being a genuine software developer.
Biggest teaching moment ▶ 53:40 Detailed explanation of MySQL 1 and permissioning architecturesDaigle walks Sean through the underlying database architecture of MySQL 1, Vitesse sharding, and the unique challenges of synchronizing on-prem enterprise containers with cloud data residency.
The host holds their own ▶ 13:30 Synthesizing Postel's Law with AI skill architecturesSean demonstrates deep software architecture knowledge by applying Postel's Law to skill orchestration and proposing standard repository conventions for composable AI agents.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | The hosts as informed peer | Guest teaching | Guest disagreement | The hosts pushing back | Why |
|---|---|---|---|---|---|---|
| Kyle Daigle's Dual Executive Role at GitHub and Microsoft | 3 | 4 | 1 | 2 | Sean asks Daigle about balancing the dual COO and CMO roles given his developer origin. Daigle provides historical background on building early GitHub infrastructure like webhooks and transitioning into management. | |
| Analyzing the Surge in Executive GitHub Commit Activity | 2 | 5 | 1 | 1 | Sean notices Daigle's surging commit activity. Daigle explains how he builds retrospective AI workflows connecting Obsidian, Teams transcripts, and GitHub issues. | |
| Deploying Frictionless AI Workflows Across 3,000 Employees | 3 | 4 | 1 | 2 | Sean inquires how AI workflows scale across GitHub's 3,000 employees. Daigle explains their CLI distribution and reliance on ChatOps and Slack despite Microsoft's ownership. | |
| Managing AI Skill Bloat with Composable Micro-Skills | 6 | 3 | 1 | 2 | Sean invokes Postel's Law and proposes dedicated repository conventions for composable skills. Daigle agrees with moving away from monolithic skills toward atomic Lego-like micro-skills. | |
| Domain Context and Plain English Customization in Prompt Engineering | 4 | 4 | 1 | 1 | Sean touches on skill integration while Daigle explains how domain-specific prompt tuning in plain English adapts identical tools for analysts versus executive briefings. | |
| Former Developers in Leadership and the AI Executive Slide Deck | 4 | 4 | 2 | 3 | Sean probes whether developers make better executives and notes AI-generated artifacts have honest charm. Daigle recounts generating an entire executive budget slide deck with AI while intentionally tuning it to look handcrafted. | |
| GitHub Actions Origins, Early Webhooks, and Cloud Isolation | 5 | 6 | 2 | 2 | Sean asks about the security footprint of GitHub Actions. Daigle educates the host on legacy architecture, recounting how GitHub used to execute arbitrary Ruby scripts in webhooks without modern containerization. | |
| The NPM Acquisition and Navigating Ecosystem Security Tradeoffs | 4 | 5 | 1 | 1 | Sean asks about the history of the NPM acquisition. Daigle explains the engineering balancing act of tightening supply chain security policies without breaking long-standing ecosystem conventions. | |
| Open Source Vendoring, Supply Chain Security, and Platform RFCs | 5 | 5 | 2 | 2 | Sean asks if AI vendoring and eliminating NPM packages is viable. Daigle points out that agents are vulnerable to prompt manipulation and clarifies GitHub's community RFC adoption pattern. | |
| The Evolution of Pull Requests and Autonomous Code Trust | 4 | 5 | 2 | 2 | Sean brings up alternative PR frameworks like prompt requests. Daigle argues PRs are fundamentally social verification instruments codifying human trust rather than pure technical diffs. | |
| Rethinking Developer Signals Beyond Gamifiable Star Metrics | 4 | 4 | 1 | 2 | Sean questions the validity of gamifiable stars and contributor counts. Daigle highlights GitHub Sponsors as a hard financial trust signal and advocates for programmable agentic heuristics. | |
| Surpassing 200 Million Developers and Demystifying Coding | 5 | 5 | 3 | 4 | Sean pushes back on GitHub's 200 million developer figure, arguing many are merely account holders. Daigle rejects gatekeeping definitions, asserting anyone turning ideas into running code is a developer. | |
| GitHub Spark, Runtime Simplicity, and Visible Code | 3 | 4 | 1 | 1 | Sean notes GitHub's ability to retain beginner audiences. Daigle discusses GitHub Spark's balance of runtime hosting simplicity while ensuring underlying code is never concealed. | |
| Confronting 14x Commit Growth and System Scalability Pressures | 4 | 5 | 1 | 3 | Sean introduces spicy questions regarding system reliability under 14x commit volume growth. Daigle openly details novel permission scaling bottlenecks across the stack. | |
| Deep Dive into Infrastructure Bottlenecks, Actions, and MySQL 1 | 7 | 7 | 2 | 3 | Sean presses on why GitHub suffered downtime despite peers handling rapid growth. Daigle gives an extensive technical explanation covering CPU-bound CI runs, MySQL 1 sharding, and monorepo blob scaling. | |
| Long-Term Reliability Roadmap and Engineering Transparency | 6 | 5 | 1 | 3 | Sean references YouTube scaling with Vitesse and automated simulation practices. Daigle provides a roadmap for reliability improvements and commits to engineering transparency. | |
| GitHub Copilot's Transformation into a Full-Lifecycle Agent | 4 | 5 | 1 | 1 | Sean asks about the origin and evolution of Copilot. Daigle describes shifting from fine-tuned completion models to unified multi-agent SDK harnesses across the development lifecycle. | |
| Ambient AI Context and Expanding Beyond the Code Editor | 5 | 5 | 1 | 2 | Sean questions if software development tooling is at the end of history. Daigle argues tools suffer from hyper-myopic design and outlines the need for cross-platform ambient context. | |
| OpenClaw, Operating System Sandboxing, and Platform Primitives | 4 | 5 | 1 | 2 | Sean asks why Microsoft appointed a dedicated CVP for OpenClaw. Daigle explains the engineering necessity of OS-level sandboxing on Windows and providing open primitives for agent execution. | |
| Microsoft Build Preview: WorkIQ, FoundryIQ, and Enterprise Context Engines | 5 | 4 | 1 | 1 | Sean asks about upcoming Microsoft Build reveals and questions for Satya Nadella. Daigle previews WorkIQ and FoundryIQ context engines for enterprise workflows. |
Statements from this episode (25)
Daigle: LLMs are most valuable for retrospective workflow analysis
“I find AI in like what most of this like launch here is, is actually like less building forward. It's actually like. A recursive loop backwards. I'm always looking at what had happened first, like go back through the week and tell me what we did, what worked, …”
Daigle: GitHub distributes internal AI CLI skills across non-technical staff
“What we've actually ended up doing is
We've built like a set of you know, skills internally.
We have like, we each have our set of skills and we've just been distributing even to the non-technical folks, the CLI.
And then effectively we're just giving it acces…”
Daigle: GitHub uses Microsoft Teams for video, but keeps Slack for ChatOps
“So we use teams for video communication, like, but we don't use it for chat.
We get home for a long, long history, right?
We always talk about chat ops and like, everything is built into slack, like every command, every flow, even though you have been acquired…”
Daigle: The era of massive, monolithic AI skills is ending
“First is I think that we're ending the era of these like massive, beautiful, perfect skills that are just like not any of those things, you know?”
Daigle: Domain-specific prompt permutations are essential to avoid obvious AI output
“The interesting matrix problem when we go from like a developer set of concerns to all kinds of different professions is that what that word means to me is different than it means to you is different than it means to the, you know, analyst or the salesperson. …”
Daigle: Plain English modification of AI skills is a transformative paradigm
“There's just something more magical about getting a response back and being like, that's not right, and then you just crack the skill open, you just type English words, you know, and it's different. That, that, that building block is just I think very unique.”
Daigle stealthily generated GitHub's executive revenue deck using custom AI
“I built like a little app just to make it so I could look at some of the information in a SQLite database more easily. And I ultimately built this entire presentation without touching any of it. And I was like, okay, I'm just going to present this to our CRO, …”
Daigle: AI shifts chief of staff role to human connection
“It's not that the jobs, like the roles don't all go away. They just change, you know? And so, yeah, I don't have someone spending all their time building out slides for me in presentations. Cause I don't need that anymore, but now I need that person that is ab…”
GitHub uses Azure dev compute for fast, isolated AI agent VMs
“We have a new underlying agent containerization service. It's like through, we're using it under the hood. It's through Azure. They recently announced it. The Azure like dev compute, but it's like very fast very fast compute to be able to like spin up you know…”
Daigle: Every security enhancement made to NPM breaks many developers' workflows
“Every move that we make to make it more secure will break a lot of people.”
Daigle: AI code vendoring will not solve supply chain vulnerabilities
“And so I do think there's something true there where having like either taking only what you need or the dependencies just getting incredibly small over time, I think will help to some degree, but it's not going to solve the fundamental problem. I don't think …”
Daigle: Replacing pull requests is hard because code review codifies social trust
“I think the reason why there's not a single answer is ultimately we're trying to codify trust. We're trying to say like, okay, if Sean reviews this, I'm going to trust it because you're Sean or you're the senior dev or you're the whatever. And right now when w…”
Daigle: GitHub stars and commit counts remain fundamentally gamifiable despite anti-abuse efforts
“I think the thing that I don't love about stars or commit counts or anything else is like, ultimately, even with all of the various like abuse and de-spamming and deduplication work that we do or anti-abuse, you know, work that we do, these are all like not ac…”
Daigle: GitHub has surpassed 200 million registered developers
“Oh, we're over two hundred million now.”
Daigle: Using AI to build running apps qualifies you as a developer
“If you create code, if you have an idea and you create it into some way of like, I'm going to run it and use the app right now, you may still use AI in that moment, but that's okay. At some point you're going to do the next thing.”
Daigle: GitHub will never hide source code in low-code abstractions
“When we put a veneer on top of something, we still always show you the code that's kind of like a tenant. We're never gonna, like, hide the code from you ever because what, like, yeah, it's the whole point, you know.”
Daigle: GitHub Spark's real value is easy runtime, not simple packaging
“What we learned with things like Spark is that really the value of Spark for most devs is, like, easy runtime. And you may have a runtime or a host that you're going to use for that, or you just build something and run it. But like the package of making that l…”
Daigle: GitHub processes more activity in a month than all last year
“And now we're doing more in a month than we did in a year last year.”
Daigle: GitHub Actions expansion drives massive CPU demand across data centers
“Some of the growth issues which is why we're, I was talking about you know, pushing hard on more CPUs is in actions in particular more tools, more agents, more PRs mean more builds in more CPUs. And so we are expanding through not just our data center, but obv…”
Daigle: GitHub's permissioning bottlenecks stem from a legacy 'MySQL 1' database
“The place that we continue to have pain is in permissioning. And so, right now, many of our permissioning layers sit into a database that we, like, internally call MySQL one, and old Hubbers will know what I'm talking about.”
Daigle: GitHub Copilot tools and agents share a single unified SDK
“Also have now had, now have like a single underlying SDK and harness for our coding agent, you know, co-pilot ultimately the new CLI, the new desktop app, cloud agents that use the same SDK.”
Daigle: The real value of AI agents is full software lifecycle automation
“What we think is that it's not solely about the code generation. It's really about having the ability to use these you know, coding agent brained harnesses or runtimes across, not just. The coding experience where I'm going to like send a bunch of tasks out, o…”
Daigle: Running AI agents locally requires improved OS-level sandboxing
“One of the problems that we have, right. Is that our agents, if you install them, not on a Mac mini or not on a hosted device, you install them on a personal device or a work device. We need better sandboxing at the OS level.”
Daigle: Operating systems must evolve because autonomous AI agents now use them
“Operating systems need to look different than they looked five years ago because it's not just you using them anymore.”
GitHub deploys Microsoft's WorkIQ context engines to its internal developers
“These little, like oversimplifying it context engines are wild good, and, like, we've given them to our developers at GitHub. We've given them to employees at GitHub as we've used these tools to be able to just ask questions around
Everything that you have in …”