“One of my investors, who's very well regarded, told me that it's actually a good idea to make AIs hallucinate, because it's entertaining to watch AIs make mistakes and share that and screenshot it, and it goes viral on Twitter. And that'll give you a lot of new users. And obviously, we did the opposite of that.”
sputnikintelligence.com
Newcomer asked whether Perplexity's investors had really told Srinivas that accuracy wasn't the point. He says an unnamed investor advised letting the AI make mistakes for virality, and Perplexity built around accuracy instead.
…investors who are now working at big tech companies actually, who actually told me that it's a bad idea to give accurate answers because people actually like it if AI Race mistakes and AI hallucinates because people actually find that entertaining. And that was true. A lot of how the original. Your
investors told you AI is not
one of my investors who's very well regarded told me that, obviously, not going to name who. But one of his arguments was that it's actually a good idea to make AIs hallucinate because it's entertaining to watch AIs make mistakes and share that and screenshot it and it goes viral on Twitter. And that'll give you a lot of new users. And obviously, we did the opposite of that and we've been pretty good too. And I think fundamentally the problem we really solved is getting AIs to be accurate. And obviously that's immediately realizable in the form of an answer engine. But you can keep building on that accuracy foundations for deep research, browsing agents. Every form of eugenic action is search. You're searching for the right tool, you're searching for the right file, you're searching for the paragraph within the file, you're figuring out which files to use in a big repository of code. You're figuring out how to orchestrate different models, different connectors, how to put it all together to a final report or an answer or a web app or an artifact. Each of these individual actions is picking…
Yeah. When you look at the revenue for Perplexity and the mix, tell me what I'm wrong here, but you guys sort of began with subscription-based, right? People were paying for the pro version of Perplexity, which uses multiple models. And then you guys, I think, do some sort of token usage-based…
Yes. A big part of our revenue growth this year has been coming from consumption-based or token-based. And we charge for usage basis on our computer agent. And every subscription plan gets a certain amount of bootstrap credits. On the $200 a month plan, everybody gets a certain number of credits…
What would you say? I mean, what is the main use right now? The most amount of use of Perplexity's tools is coming from what?