.webp)
Generated by Master Biographer | Source for LinkedIn Content
Two developers meet at a San Francisco sports bar after a Ruby meetup. One of them has a side project. They agree to spend their weekends hacking on it together. Three months later they have a product. Four years later — without a single outside investor — they're running the world's largest code collaboration platform. Then one of them turns down $300,000 from Microsoft to keep going.
That's the public story. Here's everything they don't tell you.
The founding mythology says GitHub was built "in a weekend." The reality is more interesting.
In October 2007, Tom Preston-Werner was nursing a problem. He'd built a Ruby library called Grit for accessing Git repositories, but sharing and collaborating on code with Git was still painful — there was no central place to host it, discover it, or contribute to it. He pitched the idea of a hosting platform to Chris Wanstrath at a bar in San Francisco after a local Ruby meetup called "I Can Has Ruby."
The first repository commit happened the very next day — October 19, 2007.
What followed wasn't a single weekend sprint. It was three months of every Saturday. Preston-Werner handled design and the underlying Ruby library. Wanstrath built the Rails application. They met in person every Saturday to make decisions, sketch wireframes, and argue about pricing. PJ Hyett joined as a third co-founder. Scott Chacon came aboard to handle Git expertise.
GitHub launched in public beta on January 1, 2008. The official launch followed on April 10, 2008.
The "weekend hack" story compresses three months of unglamorous weekend work into a cleaner myth. The real story is actually better: four people, no funding, no office, meeting on Saturdays at a bar because they genuinely believed the world's developers needed a better way to build things together.
The detail nobody mentions: GitHub's first revenue model came from customers asking to be charged. They made public repositories free and charged for private ones. As Wanstrath later explained, they "charge the people asking to be charged" — private repo customers showed up wanting to pay before GitHub had even fully designed the billing system.
From April 2008 to July 2012 — four full years — GitHub took zero outside investment.
This wasn't for lack of offers. Investors came knocking repeatedly. The founders politely declined every time, pointing to a simple fact: the company had been profitable since its founding and didn't need the money.
The bootstrapping math was elegant. Startup costs were almost nothing: "a few hundred bucks" in legal fees, a domain name, a cheap Slicehost server slice, and stock artwork for the site. Engine Yard, a hosting company, provided free infrastructure in exchange for publicity — a deal that eliminated the single biggest early cost. The three co-founders started with small paychecks and gave themselves raises each month when they hit revenue targets, incrementally working up to full-time salaries.
By 2010 — just two years in — GitHub had hundreds of thousands of users, tens of thousands of paying customers, and almost a million repositories hosted.
Then came the moment that crystallized the entire philosophy.
In July 2008, when a company called Powerset was acquired by Microsoft, Tom Preston-Werner received an employment offer: salary plus $300,000 over three years. He was 29 years old, carrying debt, and GitHub had been public for three months. He turned it down.
His reasoning, in his own words: "When I'm old and dying, I plan to look back on my life and say 'wow, that was an adventure,' not 'wow, I sure felt safe.'"
He wrote a blog post about it on October 18, 2008. It remains one of the clearest articulations of the founder psychology that made GitHub possible.
When Andreessen Horowitz finally invested — $100 million in July 2012, the largest Series A check the firm had ever written at the time — it was GitHub's first outside dollar. GitHub was already profitable and fast-growing. They took the money not because they needed it, but because, as Preston-Werner said, "We want to be better. We want to build the best products. We want to solve harder problems."
The $100M investment eventually returned over $1 billion when Microsoft acquired GitHub for $7.5 billion in 2018. Andreessen Horowitz partner Peter Levine later explained the original conviction simply: "developers would drive the new economy, and GitHub was at the epicenter of that transformation."
On August 23, 2010, Tom Preston-Werner published a blog post called "Readme Driven Development." It is one of the most quietly influential pieces of developer philosophy written in the last twenty years.
The argument: before writing a single line of code, write the README. The whole thing. What the software does, how to use it, what problems it solves.
His central logic: "A perfect implementation of the wrong specification is worthless. Until you've written about your software, you have no idea what you'll be coding."
He positioned it as a deliberate middle ground. Waterfall development produces exhaustive specification documents for the wrong solution. But the opposite — building with no documentation at all — produces brilliant libraries that nobody can use. The README is small enough to write in an afternoon, concrete enough to debate and iterate on, and honest enough to force you to actually think through what you're building before you build it.
The post was written by the co-founder of the world's largest code-sharing platform. The irony was not lost on readers: the man who built the place where all software lived believed most developers were building the wrong software in the wrong order.
It changed how a generation of open source maintainers thought about project documentation.
From 2008 to 2014, GitHub ran without traditional managers.
There were co-founders with loose titles. There were no reporting structures, no formal performance reviews, no one to tell you what to work on. Engineers self-selected into features based on genuine interest. If a project wasn't interesting enough for anyone to pick up voluntarily, it didn't get built.
The theory was elegant: hire smart, motivated people, create the conditions for good work, and trust the culture to handle the rest. The reality produced genuine innovation — GitHub shipped fast, the product was developer-first, and the culture attracted the best engineers in the industry who wanted autonomy over bureaucracy.
By 2014, the company had grown to 600+ employees. The flat structure that worked at 20 people was starting to show cracks at 600. Employees didn't know who to bring conflicts to. Performance conversations had no formal home. People didn't know what was expected of them or whether they were growing.
Then came the crisis that forced the issue.
In March 2014, engineer Julie Ann Horvath publicly described her reasons for leaving GitHub: gender bias, intimidation, a hostile environment. Her allegations included being verbally pressured by co-founder Tom Preston-Werner's wife and feeling systematically marginalized in a culture that had no formal mechanisms for addressing any of it.
An internal investigation followed. Preston-Werner ultimately resigned. And CEO Chris Wanstrath — who had stepped into the role that year — announced within weeks that the company would introduce formal management.
The software engineering department began assigning managers in spring 2014. VPs, HR functions, product managers, technical leads, and official performance reviews followed. The experiment that the tech industry had been watching as a model for "bossless companies" was over.
The lesson GitHub's leadership drew: "not all of those paths were the right paths." The flat structure hadn't just failed to prevent the Horvath incident — it had created the conditions for it. Without even a minimal management layer, there was no mechanism for employees to raise concerns, and no accountability structure for addressing them.
GitHub became, in time, a case study taught in business schools — not as a model to copy, but as evidence that structural idealism has structural consequences.
GitHub's mascot is the Octocat: a half-cat, half-octopus figure named Mona Lisa.
The origin is stranger than the character itself.
Simon Oxley is a British freelance illustrator who designed two of the most recognizable logos in tech history: the original Twitter bird and GitHub's Octocat. He designed them at roughly the same time, around 2006, as stock clip art for iStock. The character he called "Octopuss" — a reference to Git's "octopus merge" feature, which combines three or more branches simultaneously — was purchased by GitHub's founders for use on error pages.
When a journalist from Pando asked Oxley about his famous creation in 2013, he gave one of the most honest answers in tech history: "I don't remember drawing it."
GitHub later hired designer Cameron McEfee to develop the character into a full brand identity. McEfee built interactive error pages with JavaScript parallax effects — move your mouse and the Octocat reacts. The pages launched to significant Twitter buzz, reportedly generating half of GitHub's social media traffic that week.
McEfee went on to create the Octodex: a dedicated repository of pop-culture-themed Octocat variations. Rather than locking the character behind corporate IP, GitHub adopted a permissive policy: anyone could create their own Octocat variant as long as it linked back to a GitHub profile or project. The result was one of the most successful brand community-building strategies in tech history, executed almost entirely by accident.
The canonical character is named after a developer's daughter who wrote a fictional origin story about how a cat gained octopus legs.
When Microsoft announced it was acquiring GitHub for $7.5 billion in stock on June 4, 2018, developers did not celebrate.
Some did. The argument for optimism was real: Microsoft had genuinely changed its stance on open source, Satya Nadella had built credibility in the developer community, and Nat Friedman — named as incoming CEO — had spent two decades in open source and was respected across the Linux community.
But the instinct of the developer community was flight.
In the 24 hours after the acquisition was confirmed, GitLab reported importing over 100,000 repositories. Within a single hour on that Monday morning, more than 13,000 projects were migrated. GitLab saw a seven-fold increase in orders. On Twitter, the consensus was: "GitHub is dead."
At the time of acquisition, GitHub had:
- 28 million developers
- 85 million repositories — the largest source code host on earth
- 2.1 million organizations as customers
GitHub had also never turned a profit. The acquisition price implied roughly 30x annual recurring revenue — the highest revenue multiple Microsoft had ever paid for anything. Critics called it reckless. Defenders called it strategic.
Nat Friedman became CEO on October 29, 2018. He stepped down in November 2021, replaced by Thomas Dohmke.
The acqui-skeptics were, by most measures, wrong. GitHub grew dramatically post-acquisition. The most important thing Microsoft did wasn't integrate GitHub into the Microsoft stack — it was invest in what came next.
On June 29, 2021, GitHub announced Copilot.
An AI pair programmer, built on OpenAI's Codex (a fine-tuned version of GPT-3 trained specifically on public code), that suggested entire lines and functions as you typed. It launched as a technical preview inside Visual Studio Code.
On June 21, 2022, it became a paid product. The subscription model that had always driven GitHub's revenue — pay for private hosting — was now joined by something fundamentally different: pay to think faster.
The numbers became extraordinary:
- GitHub Copilot reached over $2 billion in annual recurring revenue — making it, as Satya Nadella noted in 2024, a larger business than all of GitHub was when Microsoft acquired it in 2018
- Copilot now accounts for roughly 40% of GitHub's overall revenue growth
- 90% of Fortune 100 companies use Copilot in their development workflows
- Over 20 million all-time users as of July 2025
- GitHub holds approximately 42% market share among paid AI coding tools
GitHub had started as a hosting business. It became a social network for code. After Microsoft, it became an AI company that happens to host code.
The $7.5 billion acquisition — which looked expensive in 2018 — looks different in 2025. Microsoft bought 28 million developers, their workflows, their repositories, and most importantly, their trust. Copilot monetized that trust in a way no one had predicted when the deal was signed.
1. The first commit happened the day after the pitch.
Tom Preston-Werner pitched GitHub to Chris Wanstrath at a bar on October 18, 2007. The first code was committed the next morning. The weekend build myth is a compression of three months of Saturdays.
2. They turned down $300,000 from Microsoft to keep going.
When Microsoft acquired Tom Preston-Werner's employer Powerset in 2008 and offered him $300K to stay, he wrote a blog post explaining why he said no. GitHub had been public for three months. He chose it anyway.
3. GitHub's mascot was designed by someone who doesn't remember designing it.
Simon Oxley — who also drew the original Twitter bird — sold the "Octopuss" as stock clip art in 2006. GitHub bought it for their error pages. When asked about it years later, Oxley said: "I don't remember drawing it."
4. The "no managers" culture lasted six years — and the scandal that ended it was partly its own fault.
GitHub ran without formal management from 2008 to 2014. The Julie Ann Horvath incident — which led to co-founder Tom Preston-Werner's resignation — exposed that a bossless culture has no mechanism for accountability. Management arrived within weeks.
5. When Microsoft bought GitHub, 100,000 repositories fled to GitLab in 24 hours.
GitLab — built in 2011 by a Ukrainian developer who wanted an open-source GitHub alternative — saw a seven-fold increase in orders the day the acquisition closed. Developer trust in Microsoft was so low that GitLab saw 13,000 migrations in a single hour that Monday morning.
6. GitHub Copilot is now bigger than all of GitHub was when Microsoft bought it.
In 2024, Satya Nadella said GitHub Copilot's revenue exceeded what the entire GitHub business generated when Microsoft paid $7.5 billion for it in 2018. The $7.5B bet wasn't really on code hosting. It was on owning the place where AI learns to write code.
| Date | Event |
|---|---|
| Oct 18, 2007 | Tom Preston-Werner pitches GitHub to Chris Wanstrath at a San Francisco bar |
| Oct 19, 2007 | First GitHub repository commit |
| Jan 1, 2008 | GitHub enters private beta |
| Apr 10, 2008 | GitHub public launch |
| Jul 2008 | Preston-Werner turns down $300K from Microsoft |
| Aug 23, 2010 | "Readme Driven Development" published |
| Jul 2012 | Andreessen Horowitz invests $100M — GitHub's first outside capital, 4 years in |
| Mar 2014 | Julie Ann Horvath goes public; Preston-Werner resigns |
| Spring 2014 | GitHub ends flat organization, hires formal managers |
| Jun 4, 2018 | Microsoft announces $7.5B acquisition |
| Oct 26, 2018 | Acquisition closes; Nat Friedman becomes CEO |
| Jun 29, 2021 | GitHub Copilot technical preview announced |
| Jun 21, 2022 | GitHub Copilot launches as paid subscription |
| 2024 | Copilot ARR exceeds full GitHub revenue at time of acquisition |
| Jul 2025 | GitHub Copilot crosses 20M all-time users |
Sources: Tom Preston-Werner's blog (tom.preston-werner.com), Signal v. Noise bootstrapped profile, GitHub Blog investment announcement, Fortune/a16z retrospective, TechCrunch acquisition coverage, Cameron McEfee Octocat design retrospective, Globe and Mail flat organization post-mortem, Wikipedia (GitHub, GitHub Copilot, Simon Oxley), Business of Apps Copilot statistics, GitLab founding accounts.