Denis Yaratts, CTO of Perplexity, discusses his company's approach to AI-powered search engines and how it differs from traditional search models. He highlights that while Google is free, users pay with their time. In contrast, Perplexity aims to save time by offering faster and more accurate answers, even if it means charging for the service. Yaratts explains that although the cost per query is higher for Perplexity than for Google, the complexity of the tasks Perplexity handles justifies the cost. This trade-off between time and compute power is seen as the key to their value proposition.

Yaratts shares that initially, investors were hesitant about the idea of building a search company in the presence of a giant like Google. However, Perplexity strategically focused on simpler projects like structured search while developing their core idea of AI-driven unstructured search in the background. The breakthrough came when ChatGPT increased public interest in AI, enabling Perplexity to launch their product and quickly gain traction.

The company now focuses on making search more efficient by leveraging large language models (LLMs) to simplify tasks, helping users avoid the need to navigate and analyze multiple links themselves. Perplexity’s system integrates multiple models and gives users the flexibility to choose the model best suited to their needs, which contributes to the platform's adaptability and user-friendliness.

Yaratts emphasizes the importance of agility and quick decision-making in startup success. Small teams with high autonomy drive the company forward, as engineers are empowered to lead projects and make decisions without the heavy oversight found in larger organizations. This flat structure allows Perplexity to remain responsive to rapidly changing technology and user demands.

In terms of hiring, Perplexity has focused on finding curious, passionate individuals over those with extensive experience, favoring potential and drive. Initially, they employed a trial-based hiring process, where candidates worked on real projects to assess their fit with the company. Yaratts highlights that hiring the right people early on has helped build a strong culture, and this mindset continues to influence their hiring strategy.

On product development, Perplexity prioritizes delivering fast and accurate answers while keeping user experience simple and delightful. To ensure this, they have built extensive infrastructure for testing and evaluation, including human evaluators and even LLMs that help assess the quality of answers. This rigorous approach ensures the reliability of the product, a key factor in building user trust.

Looking to the future, Yaratts envisions a world where AI-powered agents can handle even more complex tasks, possibly performing actions based on search results. He also sees potential in vertical applications like personalized shopping experiences. Despite scaling and a possible funding round, the company intends to remain lean, focusing on quality over quantity and continuing to optimize their product for user satisfaction.

Yaratts believes that with the improvements in AI tooling, small teams or even one-person companies could potentially achieve great things. He concludes by reaffirming Perplexity's commitment to innovation, agility, and making life easier for users through time-saving, accurate AI solutions.

https://www.linkedin.com/in/denisyarats/