Open Source, AI Agents, and the Power of Showing Up: Brian Douglas from Continue
This week on Vermilion Cliff Notes, I sat down with Brian Douglas (aka B Dougie) to talk about open source, the new AI-powered dev stack, and what it takes to build a company when the market is moving fast and funding is tight. Brian’s current role is Head of Developer Experience at Continue, but his path there wasn’t typical, and that’s what made this conversation so insightful.
Brian didn’t start in engineering. He was an MBA student who Googled how to build an app, started blogging what he learned, and got a DM that changed his life. Since then, he’s been an early engineer at Netlify, led DevRel at GitHub, and founded OpenSauced, an open source insights platform backed by some of the best investors in the devtool world.
He’s got one of the most thoughtful perspectives out there on how open source is evolving, how to actually use LLM agents in production, and what it takes to build momentum in today’s ever changing market.
Below are a few highlights from the episode. Full interview below:
“You don’t need permission to start - I didn’t know this world existed until I blogged into it”
Brian’s startup story started with curiosity and consistency:
“I had an idea while getting my MBA. Googled how to build an app. Found Rails. Started a blog to track what I was learning. I had like 50 followers , but I tagged it with #ruby, and people actually read it. One of them said, ‘You should interview for a job in San Francisco.’ That’s how I got into startups.”
It’s a reminder that even small acts of public learning can unlock huge opportunities, especially in technical communities.
“Open source isn’t just code anymore, it’s execution.”
When it comes to monetizing open source, Brian doesn’t see a binary. He sees a shift.
“Your moat used to be the core offering. But now, any LLM can write decent code. So your real edge is how well you execute—how you onboard, how you guide users, how you turn community momentum into product adoption.”
He shared an example of a YC startup that built a massive OSS community over four years, and is now raising a strong round, not because the code is novel, but because the execution is undeniable.
“They’ve integrated every tool their users already rely on, and built something sticky. That’s where the value is now.”
“The (fundraise) story was already being told”
Brian raised $3.5M across a pre-seed and seed round. The seed took just four weeks and it wasn’t because he ran a some specific process.
“The story was already out there. I blogged, I did podcasts, I sent regular investor updates. So when we went out to raise, people didn’t need the pitch. They already understood the arc.”
His advice for founders?
“Do as much as you can before you raise, even if that’s just putting $500/month toward expanding your bandwidth. Build a prototype. Talk to users. Get evidence. It changes the conversation.”
In today’s market, that kind of prep isn’t optional, it’s what gets you a seat at the table.
Final Thoughts
Brian’s perspective is shaped by over a decade of hands-on work across open source, developer platforms, and now AI tooling. He’s opinionated, tactical, and full of lessons for anyone building in devtools or trying to navigate the new AI stack.
If you’re working on infrastructure, agents, or open source commercialization, this one’s worth a listen. → 🎧 Full interview here