AI has fundamentally changed every facet of business and software development and continues to rapidly change our world. We recently held a roundtable with Sid Sijbrandij, OCV’s founder and General Partner, to address how AI impacts open source software and the building of an open core company.
The main takeaway: Every startup needs an AI strategy. “This is one of those times—like the internet, cloud, and mobile shifts—that's taking off fast,” said Sid. “AI traffic to shopping sites rose 12x in half a year. Things are changing.”
AI is a boon for startups that need to move fast and grow quickly. It can be applied to every area of a business, from engineering to sales, marketing, and legal. While every startup's AI strategy will be unique to them, our roundtable with Sid revealed several ways founders can leverage AI immediately.
Embrace vibe coding and rapid prototyping
“A year ago, AI was augmenting humans; a person wrote the code and AI added to it. Now, you can just ask AI to make the entire application. Of the last Y Combinator batch, 25% of the companies have more than 95% of their code made by AI. And these aren't ‘less technical' founders, they are extremely technical founders.”
“Vibe coding” is the shift from AI as auto-complete to AI writing the entire application. Startups leveraging vibe coding will be able to move faster than those that don’t, according to Sid. While AI-generated code can hit scaling issues due to code bloat and duplication, Sid advises that, "it doesn't hurt to start like that, and you can always rewrite what the AI has written. Speed is more important than anything."
Rapidly prototyping using vibe coding can completely change customer interactions, explained Sid. “If you’re on a customer call and they request a feature, can you vibe code it in an hour and then show the customer a working prototype of how it would look? That is a much better conversation than a PowerPoint. Send them a URL.”
For OCV companies working with an existing code base, using tools like Greptile and Momentic for code review and code testing is another way to speed up development with AI. “ You can look at AI codegen as adjacent supporting tools,” said Rich Aberman, OCV General Partner. “Use it to review a PR for bugs before you merge it. Or for code search to understand context and breaking changes before you submit.”
Hire engineers who already leverage AI
"If someone isn't using AI to help with their coding, that's a signal they're reluctant to use it. That's a problem.”
The profile of first hires is changing—you want people who are eager to embrace AI, and have already done so. Leveraging AI doesn't replace familiarity with code bases or passion for projects, but if engineers haven't experimented with agentic coding, that raises a red flag.
Senior engineers sometimes show reluctance to adopt AI tools, particularly because AI tends to work better with new or smaller codebases rather than large ones, which senior engineers typically focus on, explained Sid. However, companies can't afford team members who resist this technological shift. As Sid bluntly stated, "Using AI should be a requirement to join and stay at the company."
No AI Strategy = No Investment
“There's an enormous eagerness to invest in AI, but there are too few opportunities. If you have fast growth, can articulate why AI is changing the game, and how you're embracing it, you will have an easier time fundraising. If you cannot articulate an AI strategy, you are making it impossible for the investor to invest in you.”
Startup founders must include an AI strategy in their fundraising pitch or risk failing to raise. Not including an AI strategy is a clear missed opportunity at best and a red flag at worst, according to Rich. “It's a naivete that they haven't thought deeply about how either they're a beneficiary of this huge trend.”
Every market will be revolutionized by AI, and if you’re not bringing your customers into the future with AI, someone else will, according to Sid. “We’re in a revolution, and it’s changing everything. Investors expect more growth, customers expect features faster. Everyone expects more rapid innovation out of the project.”
Stick with the “Big 3” LLMs
“Try the top three—Open AI, Google’s Gemini, Anthropic—and see what works best for your case. At some point, there might be pricing pressure, and you’ll use something cheaper or migrate to open source. The time for that is when you're at scale, not today.”
While there's been significant interest in open source language models, there can only be so many LLM winners. With 5-6 companies dominating the space right now, it doesn't make sense for investors to invest in a hundred more. The focus should be on gaining access to larger context windows, such as Google's new Gemini 2.5 million-token context window, for most companies.
Sid's practical advice is to try the best, typically OpenAI, Google, and Anthropic, and see what works better for your case. While he appreciates open source models like Mistral, Meta, and DeepSeek, he advises that typically the proprietary models are between one and four months ahead of them. Choosing an open source LLM may make sense in the future, but startups should focus on the one that is going to help them move the fastest right now.
Open source has an AI advantage
“AI is evolving so fast. New models are coming out really fast, best practices are discovered, and the open source projects have an advantage because they can just merge whatever the users are suggesting. In an area of rapid innovation, open source and open core have an advantage."
Speed and rapid innovation are open source advantages—projects can typically merge changes much faster than a development team working on a proprietary product. Even “legacy” open source projects with a decades-old code base can use AI to improve quickly. “ AI tends to be really good at following instructions,” said Sid. “If you have a bunch of best practices written out, maybe you can use them to quickly clean up some of the crud in the code base. I'm curious about the ability of AI to rewrite according to best practices.”
Startups today need to fully embrace AI to succeed. It's not just a helpful tool, but a fundamental shift in how business is done. Those that quickly adapt and integrate AI into their operations, from coding to fundraising, will have a significant advantage. Ignoring AI means risking being overtaken by competitors who are moving faster and innovating more efficiently with its help.