Perplexity.ai started by bootstrapping a database by scraping Twitter and powering search over that, which became BirdSQL. The company also uses an existing search index that pulls content from the web and organizes it. Perplexity operates an additional layer of abstraction on top of this content, which they synthesize and organize even further. Srinivas emphasizes the importance of user experience, aiming to make the product as simple and intuitive as possible. He also highlights the need to earn users' trust and grow the user base by having a truly excellent product.
Key takeaways:
- Aravind Srinivas and his team at Perplexity.ai are building an "answer engine" powered by large language models (LLMs), which provides users with direct, accurate answers to their questions, backed by a curated set of sources.
- Perplexity.ai's model differs from traditional search engines as it places users at its center, not advertisers, potentially transforming the structure of the internet and the way we discover and consume knowledge online.
- Perplexity.ai ensures accuracy by using a modern version of PageRank to build a trust map of the web, using heuristics and data-driven learning from past queries to improve results.
- Aravind advises fellow founders to start with what they love, be patient, stay focused, and have a high bias for action, iterating on their idea in public and gathering feedback from users.