Y Combinator (YC), the renowned startup accelerator that gave birth to giants like Airbnb, Dropbox, and Stripe, has opened applications for its next batch. If you’re sitting on an idea or even just an itch to build, now’s the time to apply. But what lies ahead in the startup world—especially in the era of AI? For that, there’s perhaps no better example than Perplexity’s journey.
In a recent fireside chat, Aravind Srinivas, co-founder and CEO of Perplexity AI, shared raw, honest insights into startup life, what it takes to compete with Google, and why the future of AI lies in browsers, not bots.
🚀 The Only Moat is Speed
Aravind’s philosophy is simple and fierce: “The only moat you have is speed.”
In a world where OpenAI, Google, and Anthropic can throw billions at any innovation, startups can’t compete on resources—but they can win on execution. He likens running a startup to a marathon at sprint speed: continuous innovation is the only survival strategy.
Perplexity’s early traction proved this. With limited funding, a clunky product, and a name no one could pronounce, it still processed nearly 700,000 queries on New Year’s Eve—evidence of real product-market fit.
🔍 Why Perplexity Chose the Browser
While the current race is focused on building better chatbots or adding search layers into existing products, Perplexity is betting big on a different future: the browser as the cognitive operating system.
“Comet,” their upcoming AI-powered browser, aims to be more than Chrome with a chatbot. It’s about turning your browser into a productivity hub—managing your calendar, replying to emails, researching real estate, booking hotels, and completing tasks asynchronously.
It’s not about replacing humans with agents. It’s about building an AI assistant that’s integrated with your daily life—something that lives in your side tab, remembers your context, and helps you think, act, and execute faster.
🛡 Competing with Giants: Google and OpenAI
It’s no secret that Perplexity’s model—answering questions with cited sources—has inspired many imitators, from ChatGPT’s web browsing mode to Google’s AI Overview.
But Aravind isn’t fazed. He points out that incumbents like Google face the innovator’s dilemma. They can’t shift business models quickly without damaging their core ad-based revenues. Even worse? Launching AI that cannibalizes search can send their stock spiraling—like Bard’s botched demo that wiped billions off Alphabet’s market cap.
Startups, on the other hand, can afford to make mistakes.
“Google has great engineers. It’s not about incompetence—it’s about incentives,” he says.
This freedom to experiment, pivot, and rapidly iterate is what gives startups an edge.
💡 Starting from Scratch: From Grad School to Scaling Perplexity
One of the most relatable parts of Aravind’s story? He didn’t have a fully-formed idea when starting Perplexity.
Instead, he and his co-founders (grad school peers) began with an experimental tool that converted natural language into SQL queries for Twitter data. It didn’t scale—but it taught them what could. They pivoted, kept experimenting, and ended up creating one of the fastest-growing AI products.
His advice to students and aspiring founders?
“Don’t change ideas every week. Start with something, test fast, and build something people use repeatedly.”
Also: don’t fake being passionate about something. Aravind focused on AI because that’s the only thing he was good at. That clarity helped him go deep.
🔥 Why Perplexity Survives While Others Fade
What’s keeping Perplexity alive in a world where OpenAI could copy every feature?
Three things:
- Brand & Narrative: People don’t just use Perplexity—they trust it. Its obsession with accuracy, speed, and presentation has created a distinct identity.
- Distribution & Product Quality: Fast answers with verified sources and a seamless user experience make it stickier than most alternatives.
- A Bigger Vision: They’re not just another AI chat tool—they’re building a platform (Comet) that abstracts and integrates all AI tasks into one browser layer.
💰 Business Model: Not Another Google
Unlike Google, Perplexity isn’t ad-funded (at least for now). Its revenue relies on:
- Subscriptions: Growing rapidly, potentially reaching billions annually.
- Usage-based pricing: For agents that complete tasks—priced against the cost of human equivalents.
- Transactions: Think commissions from bookings, shopping, etc.
They’ve already partnered with Shopify, TripAdvisor, Yelp, and Selfbook. And as AI agents become more capable, transactional revenue will likely explode.
Aravind admits: “We’ll probably never have margins like Google. But we don’t need to.”
🌐 The Web in 5 Years: Will AI Kill SEO?
When asked whether AI tools like Perplexity are cannibalizing traffic to websites, Aravind gave a pragmatic answer.
Yes, long-tail sites chasing SEO traffic may struggle. But strong brands, content creators, and trusted information sources will thrive in a new form—either through direct engagement or integrations with AI systems.
The goal? Help users get accurate answers fast, without flooding them with irrelevant links.
🧠 Final Advice to Founders
Aravind’s closing advice is not glamorous—but it’s real:
- Work harder than anyone else.
- Don’t fear competition—expect it and out-execute.
- Build fast. Iterate faster.
- Stay obsessed with users and details—even if it means fixing bugs as a CEO.
And above all: don’t give up.
“If you’re building something that matters, assume a model company will try to copy it. That fear? Embrace it.”
🎯 Takeaway: Apply to YC, But Build Like You Mean It
If you’ve ever dreamt of building something, now is the time. YC is accepting applications, and the only way to figure out if your idea will work is to build it, test it, and keep iterating.
The AI race is wide open—not just for coders, but for creatives, marketers, researchers, and anyone who can spot inefficiencies in the way we work.