Why we invested in Perplexity, the leader in the emerging information interpretation market
Large language model capabilities can augment traditional search engines, opening channels for new contenders in web search.
We invested in Perplexity because of its focus on AI-powered information interpretation that sources from fresh and up-to-date material. Perplexity serves as a personalized research assistant, interpreting the sources in context of the question asked by the user. By utilizing AI, Perplexity provides concise answers to users’ questions with relevant citations. This approach, along with suggestions for follow up questions, streamlines the research processes. Perplexity established itself as one of the leaders in this information interpretation space within a year of launch with a team of just over 50 people. Their traction leads among startup alternatives and their product outperforms incumbent solutions.
Most notable, Perplexity addresses critical challenges in the realm of early large language model (LLM)-based information interpretation. LLMs often struggle with accuracy and providing the most up to date information due to reliance on pre-training. Naive implementations of retrieval augmented generation (RAG) emerged as a potential solution to control hallucinations but could not address the responsiveness and scale required for a consumer grade search alternative. Perplexity’s solution, however, stands out by offering low-latency, accurately sourced answers through its core product, the answer engine. Perplexity combines public and proprietary models, data indexing infrastructure, and caching infrastructure to deliver near real-time results that consumers expect from search products. With the initial beachhead focus on knowledge workers, Perplexity’s approach leverages a dual moat strategy: an expanding index and cache for rapid responses to common queries, and human feedback-driven refinement of their proprietary model.
Perplexity AI is spearheaded by a founding team of four with diverse yet complementary skill sets. Aravind Srinivas, the CEO, brings a wealth of expertise in AI from his tenure at OpenAI, showcasing strong leadership and a strategic vision for the company. Andy Konwinski, a seasoned entrepreneur behind successful ventures like Databricks, offers invaluable experience in building and scaling tech companies. CTO Denis Yarats specializes in back-end systems, ensuring the smooth operation of Perplexity AI’s search engine infrastructure. CSO Johnny Ho previously served as a quantitative researcher at Tower Research Capital and brings the data analysis skills needed to measure and improve AI algorithms. Their collective experience spans AI, business development, engineering, and disruptive technologies, positioning them well to tackle the challenges of developing a next-generation search engine.
Looking ahead, there are exciting prospects for Perplexity, including integration with tools, domain specialization, agents/large action models, and potential integration with voice assistants. Perplexity can evolve to become the end to end interface that consumers interact with for any task that starts with a search, such as product research or trip planning. As a result, Perplexity presents a compelling opportunity for leaders in consumer electronics to diversify their search solutions on devices. In the evolving landscape of search engines, Perplexity’s user-centric approach and AI-driven capabilities position it as a promising challenger to established players.
Andy Duong is an investor at Samsung Next. Samsung Next’s investment strategy is limited to its own views and does not reflect the vision or strategy of any other Samsung business unit, including, but not limited to, Samsung Electronics.