AI: Perplexity's sharp elbows. RTZ #902

AI: Perplexity's sharp elbows. RTZ #902

I’ve discussed often how Perplexity continues to punch above its weight amongst the leading LLM AI players this AI Tech Wave . Whether it’s in fielding state of the art AI Browsers, or offering multi-LLM driven AI Search, or doing big marketing deals globally, and so much more, Perplexity continues to compete head on with the AI Big Boys. And doing it all aggressively for a while now.

And that means often having to execute on opportuntiies and ask for permission later. Or more likely, negotiate and settle.

Bloomberg explains how Perplexity continues to skirt the lines in Perplexity Is AI’s New Pit Bull”, and is worth a full read. Some highlights:

  • “Perplexity AI Inc. is being sued by Amazon.com Inc. to stop its Comet web browser from browsing Amazon’s store on a customer’s behalf.”

  • “Reddit Inc. alleged Perplexity was illegally circumventing controls to prevent the scraping of billions of posts on the message board.”

  • “Perplexity’s defense strategy is to argue that blocking it from accessing freely available websites would be antithetical to the notions of an open web.”

And the company proactively drives its own narrative through this execution:

“The way Perplexity AI Inc. frames it, the company is something of a freedom fighter for the open web. Its “answer engine” — don’t call it a chatbot — is democratizing access to information just as Google once did. And its Comet web browser, which can carry out some tasks on a user’s behalf, is a revolutionary time-saving tool.”

And sharp elbows have consequences, as even bigger peers have found out the hard way:

“But among some of its tech peers, Perplexity is becoming company non grata. Last week, Amazon.com Inc. sued the company to stop allowing Comet’s autonomous AI agent from browsing Amazon’s store on a customer’s behalf. A week earlier, Reddit Inc. alleged Perplexity, in concert with three data companies, was illegally circumventing controls to prevent the scraping of billions of posts on the huge message board so they could be cited in Perplexity’s answers. The chief executive officer of cybersecurity company Cloudflare Inc. said Perplexity was behaving “more like North Korean hackers” than a legitimate AI company.”

“The two most recent legal assaults were added to an already long list of those frustrated over how Perplexity has scraped their material: The BBC, Condé Nast, Forbes, the New York Times, News Corp., Nikkei, Merriam-Webster — all have raised complaints, accusing the AI company of going out of its way to overcome technical measures intended to protect content.”

And the company as many aggressive startups before it, is also courting litigation:

Perplexity’s defense strategy has been to become the pit bull of AI, a necessity since its survival could very well depend on legal victory. Its public relations team certainly seems well up for the scrap. Its response to Reddit was to post its response on Reddit, rousing its fans. “We won’t be extorted,” the spokesperson wrote, arguing the company doesn’t use Reddit posts to build a large language model and was therefore more akin to Google.”

“Perplexity said the suit was ultimately a “strong arm” tactic by Reddit to protect its growing business model of selling its own content to AI builders like OpenAI and Google. Perplexity says blocking it from accessing Reddit — or indeed any freely available website — would be antithetical to the notions of an open web.”

As I discussed just a few days ago, the latest major legal sparring is with Amazon, no less, with founder/CEO Aravind Srinivas going up against Amazon CEO Andy Jassy:

“Amazon’s lawsuit was “a bully tactic to scare disruptive companies like Perplexity out of making life better for people,” the company said. What’s it to Amazon if someone uses a tool to make shopping quicker or more automated? Well, a number of things: Cynics might say a bot won’t look at or be swayed by ads, nor can it be upsold. They might also point out that Amazon is developing its own AI shopping bot. But the e-commerce giant rightly points out that a third-party bot might make mistakes or ignore helpful information a human might consider when choosing the most appropriate product and delivery options.”

In the meantime, Perplexity driven by its adroit founders, is being cast in a less favorable light:

“At the heart of most complaints against Perplexity is the view that it is a dishonest actor, hiding its activities and ignoring long-established fair practices. In August, Cloudflare singled out Perplexity for using “stealth” techniques to circumvent the blocking of Perplexity’s scraping bots. Perplexity said Cloudfare was either seeking publicity by mischaracterizing Perplexity’s actions or was incompetent for not understanding them.”

Guided by its own actions across the lines, like the case of identifiying its Comet AI Browser as a Google Chrome browser:

“What Perplexity truly needs, and fast, is some allies in its fight. Other AI browsers are also cloaking their activities around the web, according to analytics firm StatCounter. Just like Perplexity’s Comet, OpenAI’s ChatGPT Atlas, and Dia, owned by Atlassian Corp., identify themselves to websites as Google Chrome rather than revealing that the visitor is using an AI browser. This is possible because the software is built using Chromium, the open-source framework behind Google’s product. But, StatCounter CEO Aodhan Cullen explained, “so also are Edge, Opera, Brave and many other browsers, but they all identify themselves as their own browser.”

“Correctly identifying the browser is basic “good hygiene,” a Cloudflare vice president, Will Allen, told me, saying the obfuscation creates significant problems for those offering web services. “How do I really know that’s an agent that’s acting on your behalf? Are you going to buy 50 flat-screen TVs because there is some loop running in the agent that got messed up?”

“So it might be expected, then, that legal complaints against Perplexity might also soon extend to competing browsers and companies, offering some strength in numbers. But then again, maybe not: OpenAI signed a $38 billion cloud-computing deal with Amazon last week, while the small development team behind the Dia browser was acquired by Atlassian, a close Amazon partner. These companies have a motivation to find a compromise.”

Part of the issue of course, is a smaller company going up against peers with FAR greater financial resources:

“No such luck for the lone ranger. Perplexity’s aggressive ethos is driven by simple necessity. It has raised just $1.7 billion to date, according to Pitchbook, compared with OpenAI’s $57.22 billion and Anthropic’s $31.75 billion. While Perplexity isn’t the kind of capital-intensive operation its competitors are — it builds its “answer engine” on models built by others — its size leaves it vulnerable to its mounting legal troubles.”
”It has made some friends: It’s paying Snap Inc. $400 million to offer Perplexity within SnapChat, and it has allocated $42.4 million to share with willing publishing partners in a revenue-sharing arrangement. But realistically, it seems doubtful the company could afford to make the number of deals it would need to maintain the service’s breadth, quality and functionality.”

For now, it’s taking the risks for the bigger rewards, and doing it ‘on the cheap’:

“Therefore, it has little choice other than to go in hard on winning the argument that it contends it shouldn’t have to. That is still a very open legal question. “You better hope we win,” Jesse Dwyer, Perplexity’s head of communications, wrote in an email, “or the internet is going to get a lot worse for a whole lot of people.” Unfortunately for him — and us, if he’s right — deeper pockets disagree.”

Perplexity in executing at a pace and path similar to countless other tech startups that found a way to mainstream escape velocity in their early days with super hustle and grit. Uber comes to mind as a recent example, amongst many.

What’s unusual this time in this AI Tech Wave, is Perplexity is competing in a field where it requires tens if not hundreds of billions just for a seat at the table.

And Perplexity barely has some of those resources, far less than their peers And thus the likely need for bigger friends sooner than later. Stay tuned.

(NOTE: The discussions here are for information purposes only, and not meant as investment advice at any time. Thanks for joining us here)





Want the latest?

Sign up for Michael Parekh's Newsletter below:


Subscribe Here