Perplexity's unique feature is its ability to summarize uploaded files and maintain contextual awareness for follow-up questions. However, Roose found that it inaccurately summarized a paper he uploaded. Perplexity CEO, Aravind Srinivas, defended the engine's performance, highlighting the opportunity in the asymmetry of user expectations between established search engines like Google and newcomers like Perplexity. Despite its potential, Roose concluded that all AI search engines, including Perplexity, still have significant issues with producing accurate results.
Key takeaways:
- Perplexity, a new AI-powered search engine, has been reviewed by tech columnist Kevin Roose, who found it performed well on complex searches but struggled with simple factual queries.
- The search engine allows users to upload files which it can then summarize and use to maintain contextual awareness when responding to follow-up questions.
- Perplexity CEO, Aravind Srinivas, believes the search engine's ability to perform better than Google on certain queries presents an opportunity for the company.
- Despite its newness and lack of ads, Perplexity, like other AI search engines, still struggles with producing consistently accurate results.