Sundar Pichai, the CEO of Google, is credited with allowing Aravind Srinivas to consume eggs.
Srinivas recalls the incident seven years ago when a Pichai meeting appeared in his YouTube channel. His vegetarians culture in India had excluded egg, as it had for many in the state, but then, in his early thirties, Srinivas wanted to start eating more protein. These was Pichai, a hero to many budding entrepreneurs in India, lightly describing his night: waking up, reading magazines, drinking tea—and eating an omelet.
Srinivas ‘ family and I shared the picture. Sure, she said: You can eat egg.
Pichai’s effect reaches way beyond Srinivas ‘ diet. One of the most highly anticipated applications of the conceptual AI period, Perplexity AI, also has him as CEO. Srinivas is also taking signals from Pichai, the president of the nation’s largest search engine, but his enthusiasm is more complicated.
” It’s kind of a conflict nowadays”, Srinivas says. ” It’s odd”.
Srinivas and Pichai both grew up in Chennai, India, in the southern Indian state of Tamil Nadu—though the two were born 22 years off. Pichai had become Google’s main professional at the time Srinivas was pursuing his PhD in computer science at UC Berkeley.
For his first analysis apprenticeship, Srinivas worked at Google- owned DeepMind in London. Additionally, Pichai was promoted to CEO of Alphabet and Google in the same year. Srinivas found the job at DeepMind exhilarating, but he was dismayed to discover that the level he had rented look unknown was a disaster—a” terrible apartment, with animals”, he says—so he often slept in DeepMind’s offices.
In the Plex, a book Steven Levy, a Designed director at large, was written for him in the business libraries. Srinivas read it over and over, deepening his recognition of Google and its improvements. ” Larry and Sergey became my innovative soldiers”, Srinivas says. ( He offered to list the chapters in the Plex and cite memory passages, but WIRED refused. )
Soon after, in 2020, Srinivas ended up working at Google’s office in Mountain View, California, as a research volunteer working on machine learning for computer perspective. Srinivas was gradually navigating the Google universe and utilizing some of his AI research efforts.
Finally, in 2022, Srinivas and three cofounders—Denis Yarats, Johnny Ho, and Andy Konwinski—teamed up to try and create a new strategy to search using AI. They initially worked on algorithms to convert natural language to SQL, but they soon realized this was too narrow ( or nerdy ). Instead, they switched to a product that combines a comparatively recent power of large language versions with a standard search index. They called it Perplexity.
Because of the method it uses AI language technology to describe results, Perplexity is occasionally referred to as an “answer” motor rather than a search engine. New requests create verbal” threads” on a particular subject. Type in a keyword, and Perplexity listens with follow up issues, asking you to develop your ask. It favors text-based or visual answers that do n’t require you to go to another location to get information, and it steers clear of direct links.
For instance, if you ask Perplexity for a recommendation of a local restaurant, it may initially respond by asking if you’re looking for a particular food. After you’ve selected an solution, it follows up by asking for a price range. The emphasis is on its AI-generated reply, which unfolds below that. Perplexity shows its resources in small boxes at the top of benefits. Firms relying on search visitors are made to feel uneasy; hello, digital media; and growth-focused investors froth at the mouth. Perplexity has a free version, and a$ 20 per month” Pro” version that supposedly offers smarter responses.
Perplexity has raised$ 100 million in funding to date, the company says. In a Series B round of funding led by venture capitalist Institutional Venture Partners ( IVP), more than$ 70 million of that was announced in January. Its first- stage funders are a real who’s who of tech investors and entities, including Jeff Bezos, past YouTube CEO Susan Wojcicki, Google general scientist Jeff Dean, past GitHub CEO Nat Friedman, Meta chief AI scientist Yann LeCun, former Tesla AI director Andrej Karpathy, and investor Balaji Srinivasan, as well as the software company Databricks and GPU maker Nvidia. When WIRED just asked Jensen Huang, CEO of leading AI device service Nvidia, about his usage of AI bots, he eagerly responded that he uses Perplexity.
Elad Gil, who led Perplexity’s primary money round, said he was immediately impressed by Aravind. He claims,” We would brainstorm over a concept or have, and he would create it in a few hours or days.” Gil, a former Googler and lifelong businessman who has also invested in or advised AirBnB, Anduril, Figma, Instacart, Square, and Stripe. ” He is truly good”, he says of Srinivas.
Google, of course, has its own conceptual search application, the less snappily named Search Generative Experience. It’s powered by Google’s latest, domestic machine- learning concept, Gemini, released in soon 2022. Although Pichai has greater resources at his disposal, he arguably has less freedom to experiment than Srinivas, because he must also protect Google’s search ad revenue, which last year totaled around$ 175 billion.
Google’s initial experiments with conceptual AI search cross the ancient world of data-hoovering, advertising-based search, and a new conceptual search product that is ad-free for the time being. ( Google also developed the critical transformer model architecture, the” T” in ChatGPT, that led to the modern neural network behind these chatty new search tools. This means that Perplexity is one of the numerous businesses attempting to engage with Google using technology that Google created.
Compared to Google’s supremacy, Perplexity’s footprint is relatively modest. The company claims that its users center has grown to 15 million active users, a 50 % increase over the 10 million users it had just two months ago. However, Srinivas and his investors believe that Google Search is ready to be disrupted.
” We’re in a unique moment in history”, Srinivas says. The best technology has always resided within Google, and Google built its entire business around its 10 blue links. But now there’s Anthropic, there’s OpenAI, there ‘s]Meta’s ] Llama. Outside of Google, there are now AI tools used to create this answer engine.
” Throughout the AI industry, we’ll see businesses that create their own models, and others that build features and tools around already existing models,” Gil says. ” I believe that both kinds of businesses will succeed over time.”
Srinivas, echoing countless startup founders before him trying to take on an entrenched corporation, says just getting people to even notice or try out Perplexity is the company’s biggest challenge. The majority of people are still “used to” typing in two or three keywords into Google, clicking on them, and reading them. So our competition is n’t even Google, it’s user awareness”, he says.
It seems like his startup could eventually become more like Google as it matures, according to Srinivas about the business model of his startup. While Perplexity’s revenue comes primarily from subscriptions right now, Srinivas says he is n’t fundamentally against adding advertising on Perplexity. His pitch? A search app that allows advertisers to bid to place ads on what they believe to be the most “high-value traffic” out of the firehose of millions or billions of searches.
” Ads are not evil”, Srinivas says, an ironic twist on one of Google’s earliest mottos. ” When ads are done right it’s amazing, and generative AI is going to help us build even better targeting”.
Perplexity will need to win the trust of many millions more people to pull that off. A significant adoption hurdle for brands adopting AI search is the inherent trustworthiness of text generators, according to Richard Yao, a writer for IPG Media Lab, an advertising agency and think tank. ” At the moment, most of the AI ‘ answer machines’ are not very good at citing their sources”, Yao wrote.
And not all attempts at new AI search engines have demonstrated Perplexity’s potential. After failing to gain traction and funding constraints, cloud computing company Snowflake purchased the search startup Neeva, which was also founded by a former Googler.
Srinivas claims that he does n’t need to persuade the entire world to use Perplexity in his search for meaning. Or even to destroy Google. If just 1 percent of the searches people run on Google instead flow to his app, that would be significant, he says. And he would tell him he still is a big fan if he had the opportunity to chat with Google’s Pichai right away. ” I would say to him,’ Keep going,'” Srinivias says.
