Aravind Srinivas, CEO and cofounder of Perplexity AI, in conversation with Ishan Sharma on the “Limitless” podcast. It dives into AI search, startup building, risk‑taking, and practical advice for students and 20‑somethings.
Table of Contents
Who is Aravind Srinivas
Aravind is an IIT Madras graduate who later completed a PhD in computer science at UC Berkeley, focusing on machine learning and AI research. He then worked as a research scientist at OpenAI, Google, and DeepMind before founding Perplexity in 2022.
He grew up in a middle‑class Indian family, talks openly about financial constraints, and shares stories like living on one cheap meal a day during internships to save for a MacBook so he could code more seriously.
Why Perplexity and AI Search Matter
Aravind explains Perplexity as an “answer engine” that combines large language models with live web search to provide direct, source‑backed answers instead of just a list of links. He contrasts this with Google’s ad‑driven model, arguing that Google is structurally incentivized to send users to other sites for clicks, while Perplexity is optimized to give fast, reliable answers in one place. He also discusses competition with Google, Bing, and OpenAI’s SearchGPT, saying the rise of similar products proves that web‑grounded AI search is a real user need and not just a fad.
From Researcher to Startup Founder
Aravind describes how Perplexity did not start as “let’s beat Google,” but as a series of smaller, practical ideas: searching over databases, tables, documents, and then over Twitter data before expanding to the entire web.
He emphasizes that great products often grow organically: solving one hard problem gives the confidence and insight to tackle the next, eventually leading to a new category like AI‑first search.
He also shares how his research background and PhD training in going deep on one topic helped him develop the persistence and problem‑solving skills required to build a company.
Lessons for Students and 20‑Somethings
Aravind admits he entered IIT Madras with a “rat race” mindset focused on rankings, grades, and competing for the best internships.
His turning point came when poor grades and disappointment forced him to step back, deeply study his electrical engineering fundamentals and later machine learning, shifting from chasing marks to genuine curiosity.
For young people, he recommends:
- Go deep in at least one area instead of staying shallow in everything.
- Choose work where lectures, research papers, or coding feel more enjoyable than entertainment, not like a sacrifice.
- Treat depth as a transferable skill: if you can master one hard domain, you can later switch and learn another.
Mindset, Risk and the “Inversion” Technique
Aravind talks about the “inversion” technique for career decisions: instead of asking “Why should I quit?,” ask “What is the long‑term cost of not quitting and never trying my own thing?
He compares two kinds of “suffering”: a safer job you don’t love versus the uncertain but meaningful struggle of building your own product, and urges people to honestly choose which pain they are willing to live with.
He suggests that starting up before 30 often gives advantages like fewer family obligations, more energy, and flexibility with visas and finances, especially for immigrants.
Building an AI Business and Making Money
The conversation covers why many AI companies lose money today: training large models and serving AI queries are expensive, similar to how Google, Amazon, Uber and others burned cash for years before turning profitable.
Aravind distinguishes between infrastructure/model companies that train foundation models (like OpenAI or Anthropic) and application‑layer companies like Perplexity that pay for models and must design products that deliver more value than their compute costs.
He explains that Perplexity focuses on:
- Reducing cost per query through better models and optimization.
- Increasing value through accuracy, speed, and workflow integration.
- Monetizing via subscriptions and carefully designed ads in the future without harming answer quality.Subscriptions vs Ads in AI Products
Aravind notes that today ad‑based businesses like Google and Meta earn far more revenue per user than subscription AI products such as ChatGPT.
He says a pure subscription model can work for high‑end “AI coworker” or research assistant products, but for a broad, everyday answer engine, some form of advertising is likely necessary to keep it widely accessible.
He stresses that any ads Perplexity introduces must remain transparent, non‑intrusive, and never distort the factual accuracy or neutrality of answers.
Hiring, Skills and What Startups Need
At Perplexity, Aravind says the most important traits are the ability to handle ambiguity, move very fast, own outcomes, learn new skills quickly, and maintain high quality under pressure.
He cares less about brand‑name degrees and more about demonstrated impact, communication, and evidence that a candidate can thrive in a chaotic, high‑velocity environment.
He contrasts this with big‑tech hiring, which often emphasizes generic problem‑solving interviews or non‑role‑specific hiring, while startups must hire directly for impact and cultural fit.
Key Personal Inspirations and Role Models
Aravind shares that discovering Elon Musk’s story of repeated risk‑taking, near bankruptcy, and relentless execution as an immigrant founder reshaped his own sense of what was possible.
He also highlights MS Dhoni as a model of long‑term thinking under pressure accepting short‑term losses, backing people patiently, and focusing on building a winning team over many years.
These examples reinforce his main message: nothing can truly stop you if you refuse to give up and are willing to endure discomfort for long‑term goals
Frequently Asked Question
1. What is Perplexity AI?
Perplexity is an AI‑powered “answer engine” that combines large language models with live web search to give direct, cited answers instead of just showing a list of links like traditional search engines.
2. Why did Aravind leave OpenAI to start Perplexity?
He believed there was a huge opportunity to reimagine search with AI, wanted to own and execute on his own vision faster than possible inside big organizations, and was willing to accept the risk of failure to avoid long‑term regret.
3. How does Perplexity plan to make money?
Perplexity currently earns through subscriptions and plans to add carefully designed advertising, while continuously reducing cost per query and increasing the quality and usefulness of its answers.
4. Is Perplexity AI CEO Indian?
Aravind Srinivas, co-founder and CEO of Perplexity AI, has been named India’s youngest billionaire.
5. Who has invested in Perplexity AI?
The Key investors, including Jeff Bezos and Nvidia, recognized the company’s potential and played a role in its subsequent growth. By April 2024, Perplexity AI raised an additional $165 million, resulting in a valuation exceeding $1 billion
Leave a Reply