What the Rise of Perplexity Represents

Credit: DALL-E, modified by The Media Copilot

It was a big week for practical AI. Besides the Perplexity news — a tool millions of people use every day — there was also the real-world debut of the Rabbit r1 AI gadget, Adobe’s advanced AI features that arrived in the new Photoshop beta (our early impression: they’re much more handy than the previous version of Generative Fill), and Meta’s AI-powered Ray-Bans getting an upgrade with multimodal abilities so they can now interpret whatever the wearer is looking at in a not at all creepy way.

But the talk of AI town was of course Perplexity. More on that in a minute, but first a word from a new sponsor…

AI Scams Are Rising. Here’s How You Can Protect Yourself.

Scammers just got even more dangerous thanks to AI. It’s become incredibly easy to copy someone’s voice or create a deepfake. Many cases have already been reported of people impersonating family members asking for money.

Here’s how you can help prevent that: Incogni is a personal data removal service that scrubs your personal information from the web. Incogni:

Protects you from identity theft and scammers taking out loans in your name.

Prevents strangers from buying your personal information on search sites.

Get 55% off with the code COPILOT. And if you’re not happy, get a full refund within 30 days.

Try Incogni

Perplexity is the darling of this particular moment in AI. After a funding round of $63 million that put the valuation of the company over the unicorn bar of $1 billion, Perplexity appears to be the company defying the current backlash to AI hype, showing that you don’t have to be working on foundational models to create something worthy of attention — and VC dollars.

If you don’t use it, Perplexity is an AI-powered search engine. It answers user queries in a way that’s more akin to an executive summary as opposed to a list of blue links. For every query, Perplexity cites its sources prominently and usually includes some that are outside the mainstream. It even has different modes for searching academic journals, Reddit posts and more.

What I found telling is that the on-the-ground response to this week’s funding news was almost entirely positive. I didn’t see much, if any, skepticism about Perplexity, the product, on X or LinkedIn: no one was calling it a boondoggle or wondering what all the fuss was about. The consensus seems to be that Perplexity has built an experience that leverages AI to better serve a certain subset of internet users — those who want to go deeper on the topics they’re searching (think: researchers, students, journalists). And it seems the size of that group is more significant than conventional wisdom might suggest.

Riding that narrative and its current moment, Perplexity is now looking to raise even more money, $250 million, according to TechCrunch. As if on cue, there was also a report about how Microsoft is blocking the service internally (disclosure: The Media Copilot staff sometimes do consulting work for Microsoft), which suggests there’s more to Perplexity than just hype.

If you squint at this picture, you can almost see Perplexity becoming a viable alternative to Google in the way Bing never was. Or even the next Google, if you’re high.

I don’t mean to be cynical about Perplexity. I’m a fan of the product — I use it, and I teach how to use it for journalism and marketing research in The Media Copilot training classes (there’s still time to save big on our May cohorts!). But much of what it does seems easy to replicate.

(story continues below)

Perplexity’s main points of differentiation are:

The UI: Citing sources and suggesting follow-up questions are offered by pretty much any LLM these days, although the way Perplexity does them hits the bull’s-eye. Don’t underestimate the value of putting the links at the top, with thumbnails — it not only provides convenience, but because it gives source material so much prominence, it gives at least the appearance of an more ethical approach than other chat experiences (essentially it’s saying, “look, we didn’t write anything — here’s who did the hard work”). That strengthens a feeling of trust from the user.

Going deeper the mainstream sources: The big customer discovery that Perplexity made is that there is a large group of people who are underserved, or poorly served, by Google search, with its model that forces users to sift through links and to figure out which ones are worthy and which are just SEO-optimized junk. Yes, Google works well at pointing to mainstream sources, but Google was never that good at surfacing authoritative voices beyond them. By its nature, AI is better at this.

Intuiting when to go long and when to go short: After you use Perplexity for a while, you realize that the length of the summary it gives you typically varies directly with the complexity of the topic. While accusations often hit ChatGPT and other LLMs for being “lazy” on certain answers, you don’t hear that so much with Perplexity, which appears to index better on the overall utility of the answer.

Add it all up, and you can see why many people are making Perplexity their first stop for research. But you can also see that it would be relatively easy for anyone building an AI model to copy those ideas, perhaps offering them as a “research mode” in their chatbots. 

You could argue that the search generative experience (SGE) in Google is starting to do exactly this. Ask Google a question taken from the suggested queries on the Perplexity homepage, and the results page typically leads with an answer similar to what Perplexity serves up, complete with prominent card-like links to sources. 

Perplexity’s dossier-like answers are mirrored in Google’s search generative experience (click to zoom).

Regardless, Perplexity still has an edge, and that’s because its real differentiator is that it’s not Google. How this manifests is evident when you do any search on a topic that even approaches an opportunity for a transaction. Google was founded on the goal of organizing the world’s information, but it’s fundamentally an advertising engine, and it’s been optimized for more than two decades to produce search outcomes that drive commerce.

If you search, “What are the best air fryers?” on Perplexity, you’ll get one of its typical dossier-like answers, citing specific recommendations from a group of reviews sites like The Wirecutter or Reviewed. In Google, the page of search results changes in many ways: Instead of leading with a list of blue links, you’ll get several clickable pictures of air fryers with prices listed, plus a row of keyword suggestions — including brands — that you can add go your search with a click. You might even get shunted into an Amazon-like Shopping page, complete with a set of boxes to tick on the left rail.

I’m not questioning the utility of any of those features, but the overall experience makes it obvious to even the most naive internet user that the page is built to drive a transaction. And while that business has been incredibly lucrative for Google, there is a cost to pay in trust. Again, I don’t mean to question the integrity of any ad product on Google’s experience — I’m sure they’re all built to make things as transparent as possible to the user — but even honest salesmen are still salesmen, and there is something exhausting about dealing with salesmen all day.

That, I believe, is the bright light that Perplexity represents, at least in its current form: a Google alternative that doesn’t just bring a lot of interesting features to the table but also radiates a trust that mature platforms simply cannot replicate.

That glow won’t last forever, though: There is the unfortunate truth, recently cited by Verge Editor-in-Chief Nilay Patel on the Ezra Klein podcast: That, on the open internet, all platforms will eventually optimize for transactions, and perhaps that is Perplexity’s destiny as well. Given the “black box” nature of AI, that relationship between user and advertiser has the potential to be even more insidious, but that’s a whole other story.

For now, however, Perplexity offers a reprieve from a commerce-driven internet, wrapped in a user experience designed to delight the truly curious. That may not add up to the next Google, but it does show that organizing the world’s information isn’t the end of the game. You can chart a path forward by finding new ways to serve it up.

The Media Copilot is a reader-supported publication. To receive new posts and support our work, consider becoming a free or paid subscriber.

Ready to start using AI like a pro?


Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.