Feb 15, 2024 · 43m · big-technology
Perplexity CEO Aravind Srinivas: How AI Challenges Google
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
In this episode of the Big Technology Podcast, Perplexity AI CEO Aravind Srinivas discusses how generative answer engines are disrupting traditional web search and Google's core business model. He examines the shifting economics of digital information retrieval, AI foundation model dynamics, publisher attribution, and the future of conversational search.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. Alex holds 25.1% of the talking time here. How this is scored →
Alex as informed peer 3.9 Guest teaching 4.4 Guest disagreement 1.7 Alex pushing back 2.8
speaking balance: gold is Alex, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 23:16 Dismissing incumbent data moat
Srinivas dismisses the host's premise that Google's data volume is an insurmountable moat, explaining that modern generative models only require a fraction of that data.
Hardest push from Alex ▶ 29:44 Host pushes publisher survival concernsAlex challenges Srinivas's assertion that citation and awareness are sufficient, arguing directly that publishers cannot survive without actual web traffic and page views.
Biggest teaching moment ▶ 23:33 Explaining the LLM paradigm shiftSrinivas explains to the host how next-word prediction on generic web text replaced traditional historical click-data moats.
Alex holds their own ▶ 31:01 Host cites Google ad earnings statisticsAlex cites specific financial figures regarding Google's incremental ad growth and Madison and Wall research to pressure Srinivas on search monetization.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | Alex as informed peer | Guest teaching | Guest disagreement | Alex pushing back | Why |
|---|---|---|---|---|---|---|
| Startup Adoption Barriers and the Boundless TAM of Information | 4 | 5 | 2 | 3 | Alex questions why fewer generative AI startups are breaking through given the hype, citing Character AI's apparent lack of usage. Srinivas politely corrects the host's framing, explaining the demographics and use cases of Character AI versus utility products. | |
| Google's Innovator Dilemma and Business Model Constraints | 4 | 6 | 2 | 2 | Alex asks about Google's internal culture and talent issues. Srinivas delivers a detailed breakdown of the Innovator's Dilemma, explaining how changing the ten-blue-links model directly threatens Google's CPC/CPM business model and stock valuation. | |
| The Silicon Valley AI Talent War and Tech Layoffs | 3 | 3 | 1 | 2 | Alex asks about recruiting from Google amidst layoffs. Srinivas shares an anecdote about a 4x counteroffer and speculates on how Google handles high-compensation layoffs. | |
| Redefining Search into Conversational Problem Solving | 4 | 5 | 1 | 2 | Alex prompts Srinivas to define what makes conversational search different from traditional search. Srinivas explains how modern search acts like an advisor rather than ten competing salespeople. | |
| Navigational Search, Answer Engines, and Wikipedia Integration | 5 | 4 | 2 | 4 | Alex pushes back on the total addressable market and whether Perplexity can replace Google completely or just coexist with it. Srinivas articulates the spectrum between navigational search engines and answer engines. | |
| Levelling the Playing Field: Model Intelligence Over Raw User Data | 3 | 7 | 3 | 3 | Alex highlights Google's massive user data advantage. Srinivas directly reframes the premise, arguing that LLMs drastically reduce the need for raw user data and equalize competition against incumbents. | |
| Content Attribution, Publisher Relations, and Real-Time Information | 6 | 4 | 2 | 5 | Alex draws on his experience as a publisher to question publisher relations, content attribution, and site crawling rights. Srinivas outlines Perplexity's academic citation model and acknowledges publisher monetization remains an open question. | |
| Rethinking Digital Advertising, E-Commerce, and User Alignment | 5 | 5 | 2 | 4 | Alex cites Google ad revenue figures and challenges the intrusion of ads into conversational AI. Srinivas responds with a vision of transactional conversion and commerce assistance. | |
| Securing Backing from Jeff Bezos and the Memo Process | 4 | 5 | 2 | 3 | Alex inquires about fundraising from Jeff Bezos and the upcoming model landscape from OpenAI, Anthropic, and Meta. Srinivas breaks down expected AI capability milestones such as multimodality, reliability, and reasoning. | |
| Podcast Conclusion and Where to Find Perplexity | 1 | 0 | 0 | 0 | Standard brief podcast outro wrapping up the episode and sharing Perplexity's web address. |
Statements from this episode (18)
Srinivas: 50% to 60% of Character.AI Users Are Under 20
“It's more popular among the younger generation, like 50 to 60% of their user base is under the age of 20. Because they're using it to talk to imaginary anime characters and things like that.”
Srinivas: Perplexity's Total Addressable Market Mathematically Has No Upper Bound
“This product has no upper bound, honestly. Because it's upper bounded by the number of people on the planet multiplied by each individual's curiosity and Curiosity has no upper bound either, right? So therefore by mathematically the Tam of perplexity has no up…”
Srinivas: Google Has More Talented People Than Even OpenAI
“I mean, you have to be insane to think you have more talented people than Google. I don't even think OpenAI has that, even though, like, they have a massive concentration of talent density today.”
Srinivas: Google Can Easily Build Perplexity But Cannot Deploy It
“It's not that they cannot build a product perplexity. They can very easily do that literally today. Like they probably already have something internally like that, but they cannot roll it out to every Google user.”
Srinivas: Google quadrupled an engineer's pay to stop Perplexity recruitment
“There was one amazing candidate that I tried to recruit from Google. This candidate used to work in the, I mean, he still works there, in the Google search team. Like, it's not one of the AI people. And the moment he told them that he's gonna join us, they qua…”
Srinivas: Google has no incentive to save search users' time
“And there is no incentive for Google to save your time in searches, by the way, because
the whole point for them is that you spend as much time opening, like, 10 tabs on Chrome the browser they control, and give, give the analytics to all of these independen…”
Srinivas: AI search is creating a new conversational query behavior
“There is like a new segment of searches that these products like perplexity or ChatGPT are creating, which is like, you know, making people actually ask well-informed questions because we were not asking questions until now. We were just entering keywords. Lik…”
Srinivas: User question formulation is the key bottleneck for AI search
“There's number one skill is number one skill that actually is a bottleneck for these products to really take off is stability to ask good questions.”
Srinivas: Wikipedia will not build its own GPT
“No, they're not going to build their own GPT. I've spoken to him and he uses our product and he likes our product a lot.”
Srinivas: AI outside Google is now higher quality than Google's internal AI
“So the first time, for the first time, we have had intelligence outside Google, like artificial intelligence outside Google, that's of higher quality than what is inside Google.”
Srinivas: Incumbents' data volume advantage has disappeared with generative AI
“So the advantages that the incumbents had in terms of having a large volume of data has gone away. I'm not saying you don't need user data at all. All I'm saying is you need a fraction of what was needed earlier for the first time.”
Srinivas: Perplexity AI summaries constitute fair use of publisher content
“I think that's very fair use of other people's content. We're not stealing it. We're actually, like, just being a middleman between them and the end reader, and we're giving them more visibility, right?”
Srinivas: Perplexity link clicks will be more valuable than Google clicks
“Our sense is also that the value of a link click on perplexity will be more than on Google. You know, your pitch has to be a way higher intent to still leave the site and go.”
Srinivas: Perplexity will not extract all search value like Google did
“We perplexity is not trying to be creating a trillion dollar economy
On top of ads or on top of all the searches that we get and trying to take all of it and not give away much to others. Like nobody won, only Google won in the past era where, you know, the te…”
Srinivas: Amazon faces Google-like ad conflicts with conversational shopping
“For Amazon to do this hard, obviously, because they do rely on advertising revenue on amazon.com to keep amazon.com profitable, so they have, like, similar problems to Google there”
Srinivas: Anthropic will release a model better than GPT-4 in 2024
“I actually think they will end up creating a model better than GPT-IV this year. Like, it, it's sort of almost guaranteed to happen. So I believe it's going to happen with CLAW-III.”
Srinivas: OpenAI will stay ahead of competitors with GPT-4.5 or GPT-5
“I'm sure there's a GPT 4.5 or five that will stay ahead. So it really is going to be a cat and mouse game there where Anthropics playing catch up and OpenAI is ahead through multimodal capabilities, reasoning capabilities, and things, things like that.”