Things you didn't know about GitHub - with CEO Thomas Dohmke

Episode Overview

• This episode chronicles the journey of GitHub, from its early developer-first principles to its acquisition by Microsoft and subsequent "refounding" on AI with the launch of Copilot.
• GitHub CEO Thomas Dohmke shares the origin story of Copilot, detailing how early access to GPT-3 led to a product that achieved explosive growth and fundamentally changed how developers write code.
• The conversation explores GitHub's internal strategies, including its unique approach to security, its counter-cyclical decision to hire junior developers, and its focus on community engagement as a key hiring metric.
• Dohmke provides an inside look at the Microsoft acquisition, explaining the cultural shift that made it possible, the principles that ensured GitHub's independence, and the future vision of AI as the next layer of abstraction in software development.

Key Concepts

• Origin of GitHub Copilot: GitHub's partnership with Microsoft and OpenAI provided early access to GPT-3, revealing its unexpected ability to write functional code. The initial vision included text-to-code, code-to-text, and conversational AI, but the "text-to-code" autocomplete proved most viable and successful.
• Developer-First Philosophy: A core principle driving GitHub's success from its inception, focusing on solving developer pain points and creating a platform for collaboration, exemplified by innovations like the pull request.
• Hiring Strategy: GitHub values demonstrated passion and community involvement (e.g., the contribution graph) over traditional resumes. They actively hire junior developers for their fresh perspectives and native AI skills.
• Proactive Security: GitHub leverages its holistic view across millions of repositories and its CodeQL tool to proactively find and help fix vulnerabilities in open-source projects, enhancing ecosystem security.
• The Microsoft Acquisition: Enabled by Satya Nadella's "cloud-first, mobile-first" cultural shift, the acquisition was guided by three principles: keep developers first, accelerate GitHub's growth, and allow GitHub to accelerate Microsoft. This led to revenue growth from ~$200M to over $2B ARR.
• "Refounding on Copilot": AI is positioned as the new foundational technology for GitHub's future, just as Git was its original foundation. This strategic shift is reflected in decisions like open-sourcing the Copilot VS Code extension.
• AI as Abstraction: AI agents are viewed not as autonomous developer replacements but as the next level of abstraction in software development, similar to how compilers abstracted away assembly code, empowering engineers to tackle more complex problems.

Quotes

• At 0:16 - "it was able to write decent code in different programming languages and would not mix up the syntax between Python, Ruby on Rails and JavaScript." - Dohmke describes the initial breakthrough realization about GPT-3's coding capabilities.
• At 0:32 - "text to code, code to text as in describing code, and conversational coding." - Dohmke outlines the three initial ideas GitHub and OpenAI explored for an AI coding tool.
• At 1:12 - "And within a few months, it had gone to a million users." - Dohmke highlights the explosive growth of Copilot shortly after its public preview in June 2021.
• At 1:20 - "but it actually is good enough that I don't want to work without it anymore." - Dohmke summarizes the sentiment of initially skeptical developers who became dependent on Copilot after trying it.
• At 23:40 - "We also have a security lab where we have a team of researchers that are hunting for security vulnerabilities using our own security products like CodeQL." - Detailing GitHub's proactive investment in security, where they use their own tools to find and responsibly disclose vulnerabilities.
• At 29:01 - "I think that's backwards in many ways... folks that go to high school now or to college or even kids earlier in the education, they get to use AI much faster." - Arguing that younger talent is more native to AI and brings valuable new skills, countering the idea that AI replaces junior developers.
• At 33:48 - "If you don't have a green contribution graph on your GitHub profile... that I think matters more to us than whether you have five years at one company and five years at another company." - On what GitHub values in hiring, emphasizing demonstrated passion and community engagement over a traditional resume.
• At 36:54 - "What always made GitHub GitHub is that they put the developers first." - Identifying the core principle that has driven GitHub's success from its earliest days.
• At 40:43 - "GitHub, I think, is the irony of software development tools... you have a decentralized system like Git and then you have a central hub that the world depends on today." - Reflecting on how a tool built on a decentralized version control system became the central hub for software development.
• At 58:28 - "Developers first, making great developer products. Accelerate GitHub. And then eventually, GitHub, you know, accelerating Microsoft." - Thomas Dohmke outlines the three guiding principles established for the acquisition to ensure GitHub's success and independence.
• At 59:02 - "In 2017... they were at about 200 million run rate... and we announced last July... that revenue's grown to over 2 billion annual recurring revenue run rate." - Dohmke provides specific figures illustrating GitHub's massive revenue growth since being acquired by Microsoft.
• At 1:04:39 - "'Just as GitHub was founded on Git, today we are refounded on Copilot.'" - Quoting his own blog post, Dohmke articulates the profound strategic shift positioning AI as the new foundation for GitHub's future.
• At 1:07:49 - "It didn't take long until people reverse-engineered the VS Code extension... At the end of the day, it's JavaScript." - Dohmke explains that because the community had already figured out how the Copilot extension worked, open-sourcing it was a natural step.
• At 1:09:44 - "I'm sure when compilers and languages like BASIC and others were introduced, there were also skeptics saying, 'I'm losing control over the instruction set.'" - Dohmke draws a parallel between the historical adoption of compilers and the current shift towards using AI agents.

Takeaways

• When innovating with new technology, focus first on the simplest, most effective application to build momentum before tackling a more complex, long-term vision.
• Internal adoption ("dogfooding") is one of the strongest indicators of product-market fit; if your own team is eager to use a tool, it's a powerful signal of its value.
• Prioritize a developer-first product philosophy to build a durable competitive advantage and a deeply loyal user base.
• In technical hiring, look beyond traditional resumes to value demonstrated passion and real-world engagement with the community, as these are often better predictors of success.
• Consider hiring junior talent as a strategic advantage, as they often bring native skills in emerging technologies like AI that can accelerate innovation.
• A successful acquisition depends on establishing clear principles that preserve the acquired company's culture and mission, allowing it to operate with independence.
• Treat foundational technological shifts like AI not as incremental features, but as an opportunity to "refound" your company's entire strategy and product roadmap.
• Frame new, powerful tools like AI agents as the next layer of abstraction that empowers professionals to solve bigger problems, rather than as a threat that replaces them.