Perplexity, an AI-powered search engine, is backed by tech VIPs like Jeff Bezos and counts billionaires like Nvidia CEO Jensen Huang amongst its regular users. Its early traction now puts it on a collision course with the search giant.
From Rashi SrivastavaForbes contributor
In
August 2022, Aravind Srinivas and Denis Yarats waited outside Meta AI boss Yann LeCun’s office in Lower Manhattan for five hours, skipping lunch to provide the NYU professor a demo of their AI program. At one point, they showed him how their model could search platforms like LinkedIn, GitHub, and Twitter and look at content on Google
GOOG
He was certainly one of several tech VIPs, including Google chief scientist Jeff Dean, former GitHub CEO Nat Friedman, OpenAI co-founder Andrej Karpathy and outstanding angel investor Elad Gil, who received similar personalized demos – and three.1 million between them US dollars invested in a seed round in September 2022. “It was very relatable for her to do research on her own Twitter,” Srinivas said Forbes.
The money gave Srinivas, Yarats and co-founders Andrew Konwinski and Johnny Ho the chance to search out an inventive method to incorporate AI into search. After trying out several ideas, they landed on Perplexity, an AI-powered conversational search engine utilized by roughly 15 million people to source and aggregate information on every topic on the web – from the perfect date night restaurants to White Elephant Gift Exchange Ideas for the Cheapest Sneakers to Sell Online. Featured on ForbesIn the 2024 AI 50 list, Perplexity delivers concise four-to-five sentence answers, in addition to quotes and links to sources, by submitting hundreds of thousands of inquiries to a combination of huge language models, including Anthropic’s Claude, OpenAI’s GPT3.5 and GPT-4, as well Open source models reminiscent of Meta’s Llama and Mistral’s Mixtral.
“It’s almost as if Wikipedia and ChatGPT had a child,” said Srinivas, 29, who serves as Perplexity’s CEO.
In lower than two years, the buzzy AI startup has raised $102 million in enterprise capital from a number of the biggest names in tech, including Amazon
AMZN
BUSINESS
MSFT
DELL
“It creates snowball effects,” Srinivas said Forbes. “People take you more seriously.”
“You can never recreate the entire index like Google. It’s too late…It’s like trying to navigate a maze where you start at a huge disadvantage.”
Since its launch in December 2022, Perplexity usage has seen a steep upward trend. About 100,000 users pay a monthly subscription fee of $20 to access advanced features of the platform, reminiscent of: B. Browsing their very own uploaded files and creating images and text from scratch. Perplexity users may also create their very own “AI profile” by adding information reminiscent of career, location, likes and dislikes to receive personalized answers and suggestions. Additionally, users can limit their searches to specific databases, including trade magazines, YouTube and Reddit. The startup, which has annual recurring revenue of around $20 million, is considering integrating native ads into the product by letting brands influence a few of its suggestions on “related questions.”
But it still faces the trillion-dollar behemoth Google – the primary stop for billions of individuals on the lookout for information. There are still many questions that Google answers higher as AI serps like Perplexity, which, for instance, suggest which shows or movies it’s best to watch or precisely answer questions on a current football game. Additionally, the tech giant has a two-decade lead in indexing and scraping the online. “You can never recreate the entire index like Google. It’s too late… It’s like trying to navigate a maze where you start at a huge disadvantage,” Srinivas said.
But Srinivas is optimistic because he doesn’t see Perplexity as a direct competitor to Google. Instead, he hopes more people turn to Perplexity to search out nugget-sized information to make quick decisions, moderately than being served with 10 rows of blue links. Perplexity is particularly necessary with regards to retrieving information buried deep in various web sites, reminiscent of: B. Instructions for tasks reminiscent of “How to”. renew a passportor summarizing long passages of text reminiscent of news articles.
“Our success does not depend at all on Google’s failure. People can use Google and Perplexity at the same time,” Srinivas said.
Shipping faster
Srinivas and Yarats first met via email in 2020 after publishing nearly similar research papers on AI training methods two days apart at UC Berkeley and NYU, respectively. The two stayed in contact over time, updating one another on advances in AI, while Yarats worked as an AI researcher at Facebook and Srinivas went on what he calls a “full world tour” and at established firms like Google DeepMind and OpenAI worked.
It was July 2022 and the AI was within the pre-ChatGPT era. Srinivas and Yarats decided to team up with former Quora engineer Johnny Ho and Databricks co-founder and colleague Andrew Konwinski from UC Berkeley to develop an application that leverages large language models to create a brand new search experience. But the technology had not yet reached the stage where it could independently generate sophisticated and relevant answers. The co-founders hired contractors from countries like India to write down templates with questions and answers to coach the AI.
Their first product, called Bird SQL, used OpenAI’s code generation tool Codex to convert natural language prompts into code and search databases like Twitter. Launched publicly in December 2022, the tool caught the eye of luminaries just like the Twitter co-founder Jack Dorsey and will bring to light information reminiscent of “People Sam Altman interacts with the most” and “Top ChatGPT Contributors.”
When Twitter ended free access to its API in February 2023, they were forced to try a special idea: an enterprise AI search product that might search an organization’s Salesforce
CRM
HUBS
“You mention an idea to them and a day later they’ve built something.”
At this point, ChatGPT had exploded. The co-founders saw a chance to create an application that counteracts the tendency to create factually inaccurate information by adding quotes. “Citations are a great way to connect search and LLMs,” said Srinivas Forbes. Now every fact Perplexity spits out is attributed to a source that may provide more context and help confirm whether the reply is true or unsuitable.
The founders were confident that this recent idea would work, and investor Friedman emphasized that that they had little to lose. Friedman remembers telling Srinivas, “Look, your product is kind of irrelevant. The worst thing that can happen is that you’re still irrelevant.”
The final product, a search engine that scoured the whole web for answers, looked completely different than the initial Twitter demos presented to potential investors. Google’s Jeff Dean even brought this to Srinivas’s attention while congratulating him, Srinivas recalls. “I don’t think Jeff would have invested” if the co-founders had began looking into AI-powered search and attempting to tackle Google, he said.
The founders’ ability to quickly prototype has translated into rapid delivery of latest Perplexity features, investors said Forbes. “You mention an idea to them and a day later they built something,” said Elad Gil, who wrote the primary and largest check in Perplexity’s pre-seed round in August 2022. Friedman also praised the corporate’s watch Speed. “They sent me demos twice a day,” he said of the startup’s early days.
The team, made up of 45 employees, has also deployed its capital properly by consciously selecting to not develop a big base model in-house that requires costly computing power and infrastructure. Instead, Perplexity is counting on AI models from vendors like OpenAI and Anthropic, whose applications are prone to turn into cheaper as they turn into more efficient, Srinivas said. The company is concentrated on expanding Perplexity’s search capabilities through partnerships with consumer hardware startups reminiscent of Rabbits R1, a wearable AI assistant device and smartphone company Nothing, which said it sold 100,000 of its latest phones on its first day. The strategy “won’t get you a billion users,” investor Friedman said, but such partnerships may help the corporate grow.
As it competes with generative AI search capabilities from established tech giants like Google and Microsoft, Perplexity has one other factor working in its favor: novelty, Friedman said.
“I think a lot of people are fixated on Perplexity because it represents the new player, the new paradigm, the new product,” he said Forbes. And if its rapid growth and recognition amongst a few of tech’s most high-profile people is any indication, it looks like this recent addition has some endurance.
Subscribe to our recent newsletter The request for all the pieces AI-related – the largest news, the most popular firms and the boldest breakthroughs.
MORE FROM FORBES AI 50