Q&A: Why is Perplexity such a big deal?

Ask Kitty is where we answer your questions about AI. Today’s question was submitted by @iamjasonlevin and was voted on in last week’s newsletter.

Looking for information on Perplexity? You’re not the only one. Just look at Google results for “perplexity” over the last year.

If Perplexity isn’t familiar to you yet, don’t worry. The search startup was only founded in 2022 and its growth has been rapid in the last year. Perplexity raked in 52.4 million visits in March 2024 (across 10 million users), up 2.2 million visits from December 2022. The 55-person startup became a unicorn just last month (April 2024) after raising $63M, valuing the company at more than $1 billion. Recent backers include Daniel Gross (Founder/Investor), Garry Tan (YC), and Dylan Field (Figma) not to mention early investment from Jeff Bezos and Nvidia.

Another reason you might not have heard of Perplexity until recently is simply that its top competitors are household names, namely Google and ChatGPT. And it’s hard to stand out as a search tool when that’s the case. The former has a brand name that is literally equitable to search (“google it”) and the latter (OpenAI) is far more capitalized, even as a startup, with $11.3B in funding and Microsoft at its side.

So, Perplexity may be an underdog, but it’s playing the job well.

First of all, the founders have a lot of experience that combines know-how across various markets. CEO Aravind Srinivas worked as a research scientist at OpenAI, President Andy Konwinski previously co-founded Databricks, and CTO Denis Yarats was an engineer at Quora along with CSO Johnny Ho — and that’s only touching on the key points of each founder’s accomplishments.

The team has not shied away from taking on Google, with its bark or its bite (the product). Let’s tackle the product stuff first.

Perplexity positions itself as an “answer engine,” not a “search engine”. The term refers to how Perplexity responds to your question with detailed answers. Unlike Google, you don’t have to dig through links to find an answer to your query (and wonder if you’ve been manipulated by ads or optimized content) – you’re given an answer in easy-to-understand sentences. That sounds like ChatGPT, but unlike ChatGPT, you aren’t just left with a paragraph of text and no way to tell if the answer is legit. Perplexity provides sources with its answers. You can check the work or dig in more directly yourself, or you can chat back to the bot to keep getting more answers. This part, and the ability to revisit past queries, is also similar to ChatGPT.

Perplexity has a freemium model. Free accounts are based on a combination of its own model and GPT-3.5, while paid Pro accounts let you access more models, such as GPT-4 Turbo and Claude 3 Sonnet.

About that “bark” we mentioned… The Perplexity team is taking Google head-on and being vocal about how its features contrast with Google. Last week, days after Google announced a bunch of updates to its Search, CEO Aravind Srinivas threw punches at the tech giant. At a speaking event with Fast Company, Srinivas poked at Google’s addition of “AI Overviews” to certain search results, saying “…[Y]ou’re not exactly sure what’s going to happen when you type in a query on Google anymore, it actually makes the product worse.” He went on to say, “People don’t like it when it’s a cluttered UI…. There’s ads, there’s panels, there’s links, all sorts of things in one single UI; it’s pretty complicated.”

Despite Srinivas’s comments, Google’s announcements still caused plenty of people to speculate that Perplexity won’t be able to cut it. Despite any die-hard fans (like the Shopify CEO), many users either find themselves switching back and forth between engines for different queries or jumping from one platform to another as new features are released.

Most of the skepticism around Perplexity’s longevity points back to one overarching point: Google’s got data. From maps to reviews, Google’s ecosystem is so large now that it will be hard to beat. For example, I searched “best gluten-free pizza by me” and Google’s results gave me everything from Reddit searches to map searches, including the results of a gluten-free app that uses Google’s tech. While messy, the end results were pretty good – I recognized the top-notch joints. Even when I gave Perplexity a zip code, I got sketchy Yelp results with places I’ve never heard of. Given the thought that Google’s AI updates and filters will presumably make this experience even better, I’ll keep going to Google for searches like this for the time being, but will be trying out Perplexity for less “brainstormy” or opinions-based questions.

You could say that Google is between a rock and a hard place thanks to AI. Google can’t be just an answer engine like Perplexity or even ChatGPT. Answers like that don’t generate ad revenue. Right now, it feels like Google’s only play is that eclectic array of search results that Srinivas was making fun of.

Personally, as a content person, I think it’s still possible for Google to find a good mix of search & answers that will satisfy most users. Still, it’s a tall order. I’ll leave you with this rundown of why that impressive roster of investors backed Perplexity in the first place, and you can decide for yourself.