Tech Celebrities Are Big Fans of Perplexity.ai – Here’s Why

If there’s one artificial intelligence (AI) vendor that can’t stop winning in 2024, it’s Perplexity AI.

Since the start of the year, it has not only hit a valuation of $520 million but developed an impressive following of 10 million monthly users.

That includes a Who’s Who of tech A-listers, such as Nvidia CEO Jensen Huang, who says he uses Perplexity AI “almost every day,” Amazon executive chairman Jeff Bezos, and former YouTube CEO Susan Wojcicki.

But what is it about Perplexity that keeps users like Jensen Huang coming back for more “almost every day?” What’s so compelling about this chatbot that South Korea’s SK Telecom is giving 32.5 million users access to Perplexity Pro for free?

The answer is that it’s revamping the traditional search experience. In just a short amount of time since its launch in August 2022, ex-Google AI researchers have built a chatbot-style research assistant that users can ask questions to, which will respond with information taken from the internet in real-time, alongside the sources that it uses.

As it stands, Perplexity.ai is poised to become Google’s biggest threat in online search, offering a highly accessible and generative AI-driven alternative that, as we will come to, is not laden down by SEO promotions or allegations over a “woke bias.”

Table of Contents Table of Contents

1. Key Takeaways
2. Bringing the Intelligent Research Assistant to Life
3. How Perplexity.AI Differs from Search
4. Breaking Down the Next Generation of Search
5. The Bottom Line
6. References

Show Full Guide

Table of Contents

1. Key Takeaways
2. Bringing the Intelligent Research Assistant to Life
3. How Perplexity.AI Differs from Search
Show Full Guide
4. Breaking Down the Next Generation of Search
5. The Bottom Line
6. References

Bringing the Intelligent Research Assistant to Life

In recent years, there has been a growing dissatisfaction with the state of modern search engines like Google, as users complain about search results appearing to become less relevant and openly promotional.

There is some research to suggest these concerns have some validity, with researchers conducting a study of 7,392 product reviews across Google, Bing, and DuckDuckGo (PDF) and finding an overrepresentation of affiliate links.

In the paper, the researchers noted that “we find the majority of high-ranking product reviews in the result pages of commercial search engines (SERPs) use affiliate marketing, and significant amounts are outright SEO product review spam.”

How Perplexity.AI Differs from Search

Perplexity AI provides a compelling alternative to a search experience that has become dominated by SEO-driven content and affiliate spam. We can see this trend echoed in Gartner’s prediction that by 2026, traditional search engine volume will drop by 25%, as search marketing loses market share to AI chatbots and other virtual agents.

From this perspective, Perplexity AI is poised to be one of the key beneficiaries of this shift away from the traditional search experience, offering an alternative that is centered

around large language models (LLMs) with the ability to use natural language processing (NLP) and natural language generation to summarize complex text into easily digestible content.

While users may have looked to ChatGPT to fill this role following its release in November 2022, its 2021 knowledge cutoff and its tendency to hallucinate and share incorrect information left it poorly positioned to provide a search experience. In this sense, there is a gap in the market that Perplexity AI has been able to exploit.

Although Google’s rebrand of Gemini appeared extremely promising at the start of 2024, boasting impressive performance on logical reasoning tasks, it has been criticized for having a “woke” bias after it produced inaccurate historical images depicting Black Vikings, “diverse” versions of American Founding Fathers, and Female NHL players.

These allegations have cost the company’s parent company, Alphabet, $70 billion in market value and cast serious doubts over the reliability of its AI product ecosystem.

Yet Perplexity AI remains unscathed by such allegations, largely due to its emphasis on fact-checking with third-party sources, which makes it easy for users to verify any claims made in its output.

Breaking Down the Next Generation of Search

At first glance, Perplexity AI is a chatbot offering a similar experience to ChatGPT, Gemini, or Claude, but it offers a more unique approach.

Users can enter a prompt to research the internet on a given topic, and the chatbot will search the web and respond with a written output answering their question in real-time, alongside relevant images and links to sources used to compile the information.

There is also the option to click on follow-up questions, search for videos, or even generate images.

Another key point of differentiation is the Discover tab – essentially a news feed that users can click on for details about current events in the news cycle. This makes it easier to explore for information on what’s happening in the world than with ChatGPT or Gemini.

Under the hood, Perplexity AI uses a range of tools to bring AI-driven search experience to life.

Most notably, this includes GPT 3.5, or GPT-4, and Claude 2 for Pro users. GPT 3.5 gives it the ability to read up to 4,000 tokens in a question, while GPT-4 enables up to 32,000 and Claude 2 100,000.

At the same time, the organization uses a technique called retrieval augment generation (RAG), where a language model connects to an external knowledge base, in this case, extracting data from search engines to provide more users with more accurate outputs. The use of RAG in this context helps to reduce the chance of hallucination.

In any case, Perplexity AI’s secret weapon against the competition is trust. The emphasis on fact-checking has gone a long way toward insulating the vendor from concerns over misinformation or hallucinations (even though they are present) – because it’s so easy for users to check its sources.

Perplexity can also concentrate on the content, returning answers based on the information it finds rather than the constraints Google and co have to work with, such as SEO scores or ad placements.

As reported by Spectrum.ieee, Perplexity.ai founder Denis Yarats said: “You can think of it like the LLM does the final ranking task.

“[LLMs] don’t care about an [SEO] score. They just care about semantics and information. It’s more unbiased, because it’s based on the actual information gain rather than the signals Google engineers optimize for whatever reasons.”

The Bottom Line

Perplexity AI keeps winning fans over by providing an AI-driven search experience that enhances the traditional search engine experience offered by Google and Bing and repackages it under the banner of generative AI and chatbot-style copilots.

Even though its flagship solution isn’t immune to errors and hallucinations, this hasn’t slowed the company down, and it is succeeding in providing one of the more trustworthy search experiences online.