The Data Stack Show - The PRQL: What’s our first ever two-part episode about?

Episode Date: January 2, 2023

In this bonus episode, Eric and Kostas preview their upcoming two-part conversation with Arjun Narayan and Frank McSherry of Materialize.

Welcome to the Data Stack Show prequel. We have one of our more special episodes coming up because it's actually two episodes, Costas. We talked with the co-founders of Materialize. We had Arjun on the show, get this, a year and a half ago. Can you even believe that? More, I think, actually. And had a really interesting conversation about streaming SQL and what they're building. And of course, we like to have people back on the show.

We had him come back on and bring his co-founder, Frank McSherry, who in his own right is a fascinating individual who has done some amazing work, both in academic research and then actually building stuff on top of Timely Dataflow, which I know you're really interested in. So we covered an immense amount of material. What did you like most about the show? What was the part that you enjoyed the most? Because it really was an incredible thing. Oh, it's just like so many different things, to be honest. Like, first of all, I find it always fascinating to see the dynamics of co-founders. That says a lot about like the company and like, especially like for me and for you.

I was like, we've been in the position of a founder, right? And we know like how special these relationships between co-founders are. Absolutely. I don't know. Like I find it very interesting for me like to just sit and observe like the interactions of these two guys. So that's one thing. Like I think just for that, it's worth like going through the recordings, but also we have, first of all, we have like a company that I love to talk about because they do try to build like something that is radically different.

Which, okay, like to me, it makes it like always very interesting to hear from Arjun and like see what's going on, like and how they it makes it like always very interesting to hear from Arjun and like see what's going on, like how they have progressed and like all that stuff. So the time I think is great after like a year to see exactly how things have changed in the company and the product. And then of course we have Frank who is, I don't know, like he is special. Like he has done like a lot of things in the industry. Yep.

He has, he's well known like for his work in privacy, but also he has introduced a new processing model, which is the time flow model that we were going to be talking about. NIAID. Which, NIAID. Which is NIAID. Yes, that was the name of the project back in Microsoft when she developed that. And that's super exciting. Like if you think about it, like since MapReduce, which is the name of the zeros, like there aren't that many different models that they have up there.

One of them is NIAID. It hasn't like yet got the traction that many different models that they have. One of them is Nyad. It hasn't like yet got the traction that like MapReduce had, but it seems that it's changing. Yeah, I agree. So yeah, I think it's going to be super, super interesting to talk with them about building companies and being co-founders, but also about the technology itself and a little more about what Niantic is and why it's so different and so interesting.

Yeah, I agree. The story of how they met and then how their relationship progressed towards Materialize is fascinating. Arjun's story about what he learned at CockroachDB and Cockroach Labs, like working on CockroachDB was fascinating, especially what he learned about himself as a programmer, I thought was really interesting. And a couple other interesting things

that really stuck out that I think our listeners are going to love about this one, the decisions that they made relative to NIAID in terms of what they impose on top of timely data flow as part of materialize hearing about the relationship between timely data flow and what they're building on top and why they made those decisions was absolutely fascinating and it was one of those sections of the conversation where you sort of step back and realize these people are unbelievably brilliant and have thought

very long and very hard about this. And then the last one, which I think is an interesting teaser, is we all think about SQL as sort of one of the most ubiquitous querying languages out there, right? But Arjun's discussion around SQL dialects and ultimately why they chose Postgres flavor, and even within that, you know, they still had to make certain decisions or impose certain, you know, constraints. And that whole conversation was really fascinating because, of course, there are different flavors of sequel, but when you're doing what Materialize does, that stuff becomes very important

and they had a lot of hard choices to make. So if any of that sounds interesting, you're going to love it, especially because Brooks wasn't there and we recorded for 90 minutes. Of course, it'll be two episodes, but we went deeper, I think, than we've ever gone on multiple subjects. So you absolutely don't want to miss this one.

So subscribe if you haven't. We'll notify you when each episode comes out. Tell a friend and we will catch you on the next one.