AI-powered Search for the Enterprise — with Deedy Das of Glean

**Alessio** (0:09)
Hey everyone, welcome to the Latent Space Podcast. This is Alessio, Partner and CTO in Residence at Decibel Partners. I'm Joian Bamakojos, Swicks, Writer and Editor of Latent Space.

**Swyx** (0:20)
Yeah, awesome. And today we have a special guest.
It's Deedy Das from Glean. Do you go by Deedy or Debarghya?

**Deedy Das** (0:28)
I go by Deedy.

**Swyx** (0:30)
It's a little bit easier for the rest of us to spell out. And so what we typically do is I'll introduce you based on your LinkedIn profile, and then you can fill in what's not on your LinkedIn. So you graduated your bachelor's and master's in CS from Cornell. Then you worked at Facebook and then Google on search, specifically search, and also leading a sports team focusing on cricket. That's something that we can dive into. And then you moved over to Glean, which is now a search unicorn in building intelligent search for the workplace.
What's not on your LinkedIn that people should know about you?

**Deedy Das** (1:00)
Firstly, guys, it's a pleasure to be here. Thank you so much for having me.
What's not on my LinkedIn is probably everything that's non-professional. I think the biggest ones are I'm a huge movie buff and I love reading. So I think I could get through, usually I like to get through 10 books-ish a year, but I hate people who count books. So I should say that number.
And increasingly, I don't like reading non-fiction books. I actually do prefer reading fiction books purely for pleasure and entertainment. I think that's the biggest omission from my LinkedIn.

**Swyx** (1:34)
What's something that caught your eye for fiction stuff that you would recommend people?

**Deedy Das** (1:38)
Oh, I recently started reading The Three-Body Problem, and I finished it, and it's a three-part series. And well, my controversial take is I did not really enjoy the second part, and so I just stopped. But the first book was phenomenal, great concept. I didn't know you could write alien fiction with physics so well. And Chinese literature in particular has a very different cadence to it than Western literature. It's very less about the...
Let's describe people and what they're all about and their likes and dislikes. And it's like, here's a person, he's a professor of physics, that's all you need to know about him. Let's continue with the story.
And I enjoyed it. It's a very different style from what I'm used to reading.

**Swyx** (2:21)
Yeah, I heard it's very highly recommended. I think it's being adapted to a TV show, so looking forward to that.

**Alessio** (2:26)
So you spend now almost four years at Glean. The company is not Unicorn, but you were on the founding team, and LLM and text interfaces are all the rage now, but you were building this before it was cool, so to speak.
Maybe tell us more about the story, how it became and some of the technological advances you've seen, because I think you started, the company started really close to some of the early GBT models. So you've seen a lot of it from day one.

**Deedy Das** (2:52)
Yeah. Well, the first thing I'll say is Glean was never started to be a technical product looking for a solution. We were always wanted to solve a very critical problem first that we saw, not only in the companies that we'd worked in before, but in all of the companies that a lot of our, a lot of the founding team had been in past their time at Google. So Google has a really neat tool that already kind of does this internally. It's called MoMA and MoMA sort of indexes everything that you'd use inside Google because they have first-party API access, who has permissions to what document and what documents exist, and they rank them with their internal search tool.
It's one of those things where when you're at Google, you sort of take it for granted, but when you leave and go anywhere else, you're like, oh my God, how do I function without being able to find things that I've worked on? Like, oh, I remember this guy had a presentation that he made three meetings ago, and I don't remember anything about it. I don't know where he shared it. I don't know if he shared it, but I do know that it was something about X, and I kind of want to find that now. So that's the core information retrieval problem that we had set out to tackle.
And we realized when we started looking at this problem that enterprise search is actually, it's not new. People have been trying to tackle enterprise search for decades. Pre-2000s, people have been trying to build these on-prem enterprise search systems.