What is it that's unique about either
your model or just the tool, the generative tool?
Yeah. So thank you for having me here.
So perplexity is an AI powered conversational answer engine.
It's not a search engine. So you ask a query, you ask a question,
you just directly gives you the answer. The main difference from other chat bots
like transcript, is that you always get the sources.
So it only says what exists on the web already and it clearly tells the user
where every part of the answer varies in coming from on the web.
One of the backers or investors is in video, but are they also supporting you
in compute? Did you get access to an A100 cluster,
for example? So NVIDIA is an investor in US
strategically where we partner with them to work on frameworks for training on
inference, but we have all our compute cluster on other hyperscalers like EWR.
Competition. I get that the uniqueness is that the
sauce is clearly labeled, but you're directly taking on board.
Mm hmm. Is the sourcing enough to sort of
differentiate in that market? Yeah.
So our perspective is that A.I. chad bots are going to be
of of wide variety there. It's going to be a chat board for you to
go in and brainstorm new ideas and interact with it to, like, ask for ideas
for birthday presents or generate essays or write poems.
There's going to be character. I like chat boards for chatting with
personalities, but there has to be one single chad board which you rely on for
accurate ly correct information on the web, and that market is there for the
taking, and we want to go for that. And 10 million people on a monthly basis
do. I'm interested in your desire to go out
and build this oven because you were an open eye as a researcher you've been
working with deep mind. Why did you feel that something new was
necessary to be created? I always wanted to try an
entrepreneurial journey so it was not meant to go and create a company for
search. We initially started offices to
brainstorm and create products using the large language models as an upcoming
technology, much before it was even called Generative A.I., it was just that
we stumbled on this amazing idea of combining large language models and web
search together in the form of a new product called Perplexity.
So you raised this money. Is that to be put towards R&D?
Compute power is a little bit of marketing because I'm kind of interested
is how you got into the hands of 10 million people.
Yes. So we've really been fortunate to have
completely organic growth. We haven't paid for any of these users.
We obviously announced all our product features on Twitter and like people
follow us there and they get to know about us.
And there's a lot of word of mouth growth for us.
But our spend is largely going to be on scaling, adoption and usage, and that's
going to be spend more on compute resources.
As you point out, Caroline and I have experienced some of that word of mouth
about perplexity. Here's the thing.
The story with Openai was about the push to commercialize.
Do you feel as a as a founder and leader of that company, that pressure?
Well, honestly, I think people just want progress, and progress can be reflected
in terms of clear user growth and adoption and not just for revenue.
So as you get more and more users, there's always going to be plenty of
opportunities for you to monetize either through subscriptions or other forms of
monetization. And therefore, right now all our
investors are pretty aligned with this that we actually want to scale more in
terms of getting growing a larger user base.