EP #155 - NOW PLAYING Dec. 18, 2024: Old McDonald Had A 👨‍🌾 Farm…On 💧 Water
Dec. 11, 2024

AI vs. Google: Perplexity's Search Engine Revolution

AI vs. Google: Perplexity's Search Engine Revolution

Text us your thoughts on the podcast

Can Perplexity's AI-powered search engine challenge, or for that matter, replace Google?

supporting links

1.     Perplexity AI Tutorial - How to use AI for Research [YouTube]

2.     Why I use Perplexity AI MORE than Google and ChatGPT for Content Research [YouTube]

3.     Perplexity AI Review: Features, Benefits, and Alternatives [ClickUp]

4.     Perplexity AI: Exploring AI-powered search beyond Google [Search Engine Land]

5.     The 8 AI Search Engines That You Should Replace Google With [KD Nuggets]

6.     A Conversation with Perplexity Founder & CEO, Aravind Srinivas [Thought Economics]

7.     Aravind Srinivas takes a dig at Google for 'rebranding AI offerings' [Firstpost]

8.     Perplexity CEO Aravind Srinivas says Google’s search results are cluttered [FastCompany]

9.     Lenny's Newsletter [Perplexity blog by Lenny Rachitsky]

10.   Discover Daily by Perplexity [Podcast]


Contact That's Life, I Swear

Thank you for following the That's Life I Swear podcast!!

Transcript

⏱️ 14 min read                   

Can Perplexity's AI-powered search engine challenge, or for that matter, replace Google? Have to tell you, Perplexity is very impressive. This start-up is redefining search with cutting-edge artificial intelligence, challenging the tech giant and showcasing what's possible with innovation from the ground up.

Welcome to That's Life, I Swear.  This podcast is about life's happenings in this world that conjure up such words as intriguing, frightening, life-changing, inspiring, and more. I'm Rick Barron your host. 

That said, here's the rest of this story 

Since its founding in September of 1998, Google has been the go-to source for answers whenever we’ve had a question about the world or needed to find something online.

Well, it now appears a new search engine in town may overturn that notion where other startups have failed.

This new search engine, Perplexity, was founded in August of 2022. As of late, it has quickly become one of the most talked-about products in the tech industry. Its founders, who previously worked in A.I. research at OpenAI and Meta, have created a new buzz that has tech insiders raving about it on social media. The company has also attracted significant investment from notable figures like Jeff Bezos, an early Google investor. In a recent funding round led by Institutional Venture Partners, Perplexity raised $74 million, valuing the company at an impressive $520 million.


From left: Johnny Ho, Aravind Srinivas, and Denis Yarats, co-founders of Perplexity. Courtesy of Lenny's Newsletter

Google's monopoly in the search engine market has been a formidable challenge for countless startups over the years, with many ambitious projects falling by the wayside. One such contender, Neeva, had to admit defeat last year, unable to carve out a substantial user base. However, the once unshakable dominance of the tech giant appears to be eroding. A growing chorus of disgruntled users has criticized the declining quality of Google's search results, which have become increasingly saturated with dubious and low-grade websites.

Test run

Intrigued by the buzz surrounding Perplexity, I took time to sit down with a friend of mine who is tech savvy, to take the weekend to take this new A.I. search engine out for a test run. We looked across desktops and mobile devices to evaluate comparisons of engineering consistencies. We started navigating both the free version and the premium offering, Perplexity Pro, which requires a $20 monthly fee and the appeal of advanced A.I. models and exclusive features, including the ability to integrate personal files into the search experience perfectly.

After probing this innovative platform, my tech friend and I agreed that Perplexity’s capabilities are nothing short of remarkable. My research has reinforced my belief that A.I.-powered search engines like Perplexity possess the power to erode Google's iron grip on the market, or at the very least, compel the tech titan to adapt and innovate to maintain its edge.

During the probing of Perplexity’s landscape, a question did pop in my head that I’ll get to later. The prospect of products like Perplexity rendering traditional search methods obsolete is both thrilling and unsettling, a double-edged sword that cuts through the fabric of the digital world we've come to know. 

Must admit, that Perplexity's desktop interface resembles Google's – a simple text box taking center stage on a simple landing page. Yes, design is everything.

The resemblance does fade when typing your query. When asking a question of Perplexity, it doesn't merely serve up a list of links. Instead, it conducts a comprehensive web search, employing advanced A.I. algorithms to craft a concise summary of its findings. These meticulously curated answers come complete with annotations linking to the sources the A.I. relied upon, which are conveniently displayed in a panel above the response.


Perplexity engineers contemplate ideas. Courtesy of Lenny's Newsletter 

Perplexity offers an innovative feature called "Copilot," designed to help users refine their searches by posing clarifying questions. For example, we entered ‘plan a trip to Kyoto, Japan.’ Copilot first asked ‘When are you planning to visit Kyoto?’ Copilot provided some top attractions in Kyoto to see and asked to select from the list for added details. With these details, Copilot provided a tailored list of suggestions that helped make decisive decisions.

Another standout capability of Perplexity is its precise source-specific search functionality, enabling users to delve into academic papers, YouTube videos, or other documents. This feature is a game-changer.  Say you’re needing to adjust a setting on your water heater. Traditional Google searches returned numerous unhelpful DIY links and ads for plumbing services. However, Perplexity, when restricted to YouTube videos, quickly identified the exact tutorial for your water heater model. Moreover, it extracted the key information and presented it as a clear, step-by-step guide.

At its core, Perplexity leverages OpenAI's GPT-3.5 model alongside its proprietary AI, a version of Meta's open-source Llama 2 model. Pro users can access a variety of models, including the acclaimed GPT-4 and Anthropic's, Claude.

What was very appealing about Perplexity is its human-like and honest approach when it lacks sufficient information. Instead of fabricating answers, it provides a partial response with a note like "No further details are provided in the search results." This transparency and humility are refreshing compared to other AI chat products that often deliver confident yet inaccurate responses.

The benefits of Perplexity A.I. search

Perplexity AI is a conversational AI search engine that can provide benefits over other search engines, including:

  • Learning through conversations: Perplexity AI allows users to learn through conversations instead of hunting for information in search results
  • Concise answers with source links: Users receive concise answers that are linked to their sources, which can be useful for fact-checking and academic research
  • Data analysis: Perplexity AI can collect, process, and analyze large amounts of information, which can help scientists and analysts quickly sort data and prepare detailed answers
  • Natural language queries and responses: Perplexity AI uses natural language for both queries and responses
  • Source citations: Perplexity AI provides source citations, which can help ensure accuracy
  • Flexibility: Perplexity AI offers multiple LLM options for flexibility
  • Collections: Users can create collections to store and organize their queries

Perplexity's Uphill Battle Against Google's Dominance

Google's superiority over Perplexity isn't just about having access to vast location data—it's about the sheer magnitude of its resources and reach. Perplexity, a very young start-up with a modest team of 41 working out of a co-working space in San Francisco, has impressively captured 10 million monthly active users. However, this figure is dwarfed by the billions that flock to Google.

Adding to its challenges:

·       Perplexity has yet to establish a solid revenue stream. The site currently operates without ads, and less than 100,000 users have subscribed to the premium version

·       While Perplexity hasn't ruled out the possibility of incorporating ads in the future, the current user numbers may not be enough to generate significant income 

·       Additionally, Perplexity lacks the diverse array of products that make Google's ecosystem so ubiquitous, from Gmail and Google Chrome to Google Docs.

Despite these hurdles, Perplexity's potential to shake up the search engine landscape is immense. In a recent interview, CEO, Aravind Srinivas, acknowledged Google's dominance but expressed confidence in Perplexity's nimble and innovative approach to challenging the tech titan.

"What fuels my optimism is knowing that for Google to outdo us, they'd have to essentially overhaul their entire business model," Srinivas noted, pointing out the fundamental differences between Google's ad-based revenue system and Perplexity's user-focused, A.I.-driven strategy.

All this said about the hurdles for Perplexity, let’s not forget that Google was once a small group of engineers.


Perplexity Discover Daily Podcast. Courtesy of Lenny's Newsletter 

 

Perplexity's cutting-edge approach faces considerable hurdles on its path forward

Reliance on Language Models: The core of Perplexity's engine is a language model with a staggering 175 billion parameters, demonstrating remarkable capabilities yet harboring inherent flaws. The challenge lies in delivering intelligent assistance while preserving the nuances of human interaction, especially considering these models can err on factual questions up to 21% of the time.

Biases and Inaccuracies: Perplexity's training data has been found to contain demographic biases, potentially skewing search results. In its early testing phase, around 15% of inquiries on critical topics like medical advice and financial guidance yielded inaccurate information, underscoring the monumental task of ensuring fairness and accuracy.

Scalability Challenges: While Google Search handles over 8.5 billion queries daily worldwide, Perplexity currently supports only around 500 million searches per day. Matching Google's capacity would not only demand substantial financial investment but also significant technological breakthroughs to expand its infrastructure accordingly.

Two Sides to the Coin: A Benefit for Users, a Menace for Publishers

As called out earlier, let me get back to my nagging question I felt while researching Perplexity. Both Perplexity and Google, now provide summaries of a user’s search request. At first sight of this I wonder about writers and publishers. If one can get a quick abstract of their search request, why waste your time scrolling down the page to review other links, as Google does today? Perplexity does the same, but it’s formatted in a rather subtle manner. 

In today's digital media landscape, a significant portion of the economy still relies on a steady stream of users clicking on links from Google and being served advertisements on publishers' websites.

With Perplexity, the need to visit a website often becomes obsolete – the A.I. takes on the task of browsing for you, providing all the necessary information directly on the answer page.

Seeing this is very concerning for publishers. If users can access the information they seek without ever leaving the search engine, the click-through rates to publishers' websites could plummet, devastating their advertising revenue. The rise of A.I. search threatens to upend the delicate ecosystem that has sustained digital media for years, forcing publishers to reassess their strategies and adapt to a rapidly changing landscape.

For writers whose livelihood depends on people reading their work, the thought of A.I. search engines rendering articles obsolete is deeply unsettling. It's a problem that highlights the complex interplay between technological advancement and the sustainability of traditional media models in the digital age.

The Looming Specter of A.I. Search: A Journalist's Dilemma

After browsing the landscape Perplexity and learning about similar products being developed by other start-ups, I kept coming back to the concerns raised by industry insiders. In short, they have a point. 

Suppose A.I. search engines can reliably summarize the latest developments in the Ukraine War or guide users to the best EV on the market. What incentive remains for anyone to visit a publisher's website? Why would journalists, bloggers, and product reviewers continue to share their work online if an A.I. search engine is poised to consume it and reiterate it in a different manner?

The convenience and efficiency offered by Perplexity is undeniably alluring. Yet, writers and publishers alike struggle with the nagging fear that they may be unwittingly contributing to their own professional demise by embracing this technology. If A.I. search engines become the norm, will there still be a place for the kind of in-depth reporting, analysis, and commentary that has been the lifeblood of traditional media?

The rise of A.I. search presents a complex dilemma, pitting the benefits of technological progress against the potential erosion of an entire industry. As we navigate this uncharted territory, society needs constantly weigh the advantages of Perplexity against the unsettling realization that each query one makes could be another step toward a future where some peoples skills and expertise become obsolete. 

What can we learn from this story? What's the takeaway?

Perplexity, has enormous capabilities but the existential questions remain. As the media landscape evolves, it's clear that writers, publishers, and consumers alike will need to adapt and innovate to ensure the survival of quality journalism in the age of A.I. search.

Well, there you go, my friends; that's life, I swear

For further information regarding the material covered in this episode, I invite you to visit my website, which you can find on Apple Podcasts/iTunes for show notes and the episode transcript.

As always, I thank you for the privilege of you listening and your interest. 

Be sure to subscribe here or wherever you get your podcast so you don't miss an episode. See you soon.