Perplexity AI was founded in August 2022 by Denis Yarats, Aravind Srinivas, Johnny Ho, and Andy Konwinski. On the company’s website, Perplexity describes itself as a “small interdisciplinary team determined to bring novel technology and useful products to the world.”
Perplexity AI’s core product is Perplexity Ask, a search interface and chat tool powered by OpenAI’s GPT 3.5 and Microsoft Bing.
How does Perplexity Ask work?
Unlike a traditional search engine that answers queries by serving links, Perplexity Ask summarizes the results to provide a direct answer and provides citations that enable the information to be verified.
Perplexity Ask also incorporates follow-up questions and related searches to help users find what they want. Each follow-up question creates a new search query and in the process, facilitates a conversational search experience.
While much of the recent focus has been on ChatGPT, Perplexity has several key advantages over the popular OpenAI product. The first advantage, as we’ve already discussed, is that Perplexity backs up its claims with citations.
The second advantage is that the search interface incorporates real-time information. While ChatGPT does not consider information past 2021, Perplexity knew the release date of the Samsung Galaxy S23 a week after it occurred in February 2023.
Perplexity’s access to recent data also makes it suited to other queries where information is updated regularly such as store opening times and weather forecasts.
In December 2022, Perplexity introduced Bird SQL – a Twitter search interface powered by the company’s structured search engine.
Bird SQL utilizes OpenAI Codex to translate natural language into SQL. This enables users with no prior experience in SQL to have queries answered from the information contained in Twitter’s vast database.
Bird SQL is not yet a commercial product and the output is only based on Twitter data. Nevertheless, it can visualize, summarize and present results with stats and graphs generated from natural language with embedded code. Bird SQL is also unique in the sense that its capabilities are beyond those of conventional search engines.
Perplexity Ask’s future role
Perplexity Ask functionality is reminiscent of how Google answers some search queries with a small explanation box above the paid and organic results. It is also similar to ChatGPT with the ability for users to refine requests and iterate as if they were speaking to a person.
While OpenAI is quick to point out that ChatGPT is not a search engine, the AI nonetheless has a tendency to produce inaccurate information and can be a very convincing liar.
The Perplexity product is one of several (such as Jasper Chat and You.com’s YouChat) that attempts to bridge the gap with a blend of conversational refinement and informational accuracy. Perplexity is not immune from providing inaccurate information, but users can at least verify output from the citations provided.
Whether Perplexity Ask and its competitors pose a serious threat to Google search as we know it remains to be seen. This question would not have been asked a mere twelve months ago, but one can safely assume that Google is aware of generative AI’s potential and how quickly its capabilities are growing.
- Perplexity AI was founded in August 2022 by Denis Yarats, Aravind Srinivas, Johnny Ho, and Andy Konwinski. The company’s core product is Perplexity Ask, a search interface powered by OpenAI’s GPT 3.5 and Microsoft Bing that answers user questions.
- Unlike a traditional search engine that answers queries by serving links, Perplexity Ask summarizes the results to provide a direct answer and provides citations that enable the information to be verified. The other most obvious feature is that the search interface incorporates real-time information.
- The Perplexity product is one of several that focuses on a conversational search engine experience that provides accurate (or at least verifiable) information.
Read Next: History of OpenAI, AI Business Models, AI Economy.
Connected Business Model Analyses
OpenAI Organizational Structure
Stability AI Ecosystem