Refactoring Podcast

How Perplexity Works — with Denis Yarats 🤖

May 31, 2024
Denis Yarats, CTO at Perplexity, discusses AI products, company structure, and internal AI utilization. Topics include Perplexity's history, AI model costs, open vs proprietary source, engineering team organization, startup planning, 'secret sauce', metrics, and future plans.
Ask episode
AI Snips
Chapters
Transcript
Episode notes
ANECDOTE

Pivot to Web Search Amid ChatGPT Surge

  • Perplexity founders initially worked on structured search, translating queries to SQL. - The launch of ChatGPT gave them the perfect moment to shift focus to web search, which quickly gained traction.
INSIGHT

Perplexity's User Alignment Edge

  • Google diverged from pure user needs due to ad and ranking systems focused on revenue. - Perplexity provides trusted, fast answers that enhance productivity beyond traditional search.
INSIGHT

Time Savings Justify AI Costs

  • AI startup costs per query exceed Google's but enable solving complex tasks. - Saving users' time justifies higher price, creating a beneficial trade-off between compute cost and user time.
Get the Snipd Podcast app to discover more snips from this episode
Get the app