Copy the formatted transcript to paste into ChatGPT or Claude for analysis
Okay. Hi everyone and welcome to the
2025 HVS Entrepreneurship
Summit. Um we're super excited to have
you here today, especially because this
is our reinaugural entrepreneurship
summit. Um there hasn't been an event
like this at HPS since precoid. So we're
really excited to have this relaunch
again and have you guys all here. Um so
if you haven't taken a seat yet, take a
seat, settle in. We're super excited for
a full day of programming um here with
you today. Um special shout out to our
co-heads of Summit, Lindsay and Anita.
Um you'll meet them in a
AI-generated overview
Aravind Srinivas, CEO of Perplexity AI, discusses his journey from academic researcher to leading a $9 billion AI company in a conversation at Harvard Business School's 2025 Entrepreneurship Summit. Srinivas traces his path from electrical engineering at IIT Madras through a PhD at UC Berkeley, where internships at OpenAI exposed him to the future of generative AI. He emphasizes how academic principles—particularly citation and iterative experimentation—directly influenced Perplexity's product design. The company, founded in August 2022, serves over 600 million monthly queries by combining large language models with cited sources, differentiating itself from traditional search engines. Srinivas argues that Perplexity doesn't compete directly with Google's navigational searches but rather transforms question-answering by providing synthesized, sourced responses. He advocates for open-source AI as essential for preventing monopolistic control, shares his philosophy of quarterly rather than multi-year planning in the fast-moving AI landscape, and reveals plans for vertical-specific structured answers, browser integration, and personalized search across user data.
Academic training in citation and peer review directly shaped Perplexity's core product principle of providing sourced AI answers, making AI responses verifiable rather than just plausible.
The optimal planning horizon for AI companies is quarterly, not multi-year, because the field moves too rapidly for longer-term predictions to remain valid.
Open-source AI models are the primary defense against monopolistic control of artificial intelligence, as they prevent any single entity from charging excessive prices or controlling access to advanced capabilities.
Perplexity differentiates from Google not by competing for navigational one-word searches, but by synthesizing multi-source answers to complex questions that traditional search engines handle poorly.