The Point of LangChain — with Harrison Chase of LangChain

**Alessio** (0:10)
Hey, everyone. Welcome to the Latent Space Podcast. This is Alessio, Partner and CTO in Residence Addesible Partners, and I'm joined by my co-host, Swix, founder of SmallAI.

**Swix** (0:19)
Welcome. Today we have Harrison Chase in the studio with us. Welcome, Harrison.

**Harrison Chase** (0:22)
Thank you guys for having me. I'm excited to be here.

**Swix** (0:25)
Been a long time coming. We've been asking you for a little bit, and we're really glad that you got some time to join us in the studio.

**Harrison Chase** (0:31)
Yeah, I've been dodging you guys for a while. About seven months.

**Swix** (0:37)
I totally understand.
We like to introduce people through the official backgrounds and then ask you a little bit about your personal side. So you went to Harvard Class of 2017 You don't list what you did in Harvard. Was it CS?

**Harrison Chase** (0:48)
Stats and CS.

**Swix** (0:49)
That's awesome. Loved me some good stats.

**Harrison Chase** (0:52)
I got into it through stats, through doing sports analytics. And then there was like so much overlap between stats and CS that I found myself doing more and more of that.

**Swix** (0:58)
And it's interesting that a lot of the math that you learn in stats actually comes over into machine learning.

**Harrison Chase** (1:04)
Oh, yeah.

**Swix** (1:05)
Which you applied at Kensho as a machine learning engineer and robust intelligence, which seems to be the home of a lot of AI founders. And you started LangChain, I think around November 2023 and incorporated in January.

**Harrison Chase** (1:18)
Yeah, I was looking it up for the podcast and the first tweet was on, I think, October 24th, so just before the end of November or end of October.

**Swix** (1:26)
Yeah. So that's your LinkedIn. What should people know about you on the personal side that's not obvious on your LinkedIn?

**Harrison Chase** (1:33)
A lot of how I got into this is all through sports, actually. Like I'm a big sports fan, played a lot of soccer growing up and then really big fan of the NBA and NFL. And so freshman year at college showed up and I knew I liked math. I knew I liked sports. One of the clubs that was there was the Sports Analytics Collective. And so I joined that freshman year. I was doing a lot of stuff in like Excel, just like basic stats, but then I wanted to do more advanced stuff. So learn to code, learn kind of like data science and machine learning through that kind of like just kept on going down that path.
I think sports is a great entryway to data science and machine learning. There's a lot of like numbers out there. People like really care. Like I remember, I think sophomore, junior year, I was in the Sports Collective and the main thing we had was a blog. And so we wrote a blog.
It wasn't me.
One of the other people in the club wrote a blog predicting the NFL season. I think they made some kind of like with stats. And I think their stats showed that like the Dolphins would end up beating the Patriots and New England got like pissed about it, of course. So people like really care and they'll give you feedback about whether you're like models doing well or poorly. And so you get that. And then you also get like instantaneous kind of like we're not instantaneous, but really quick feedback. Like if you predict a game, the game happens that night. Like you don't have to wait a year to see what happens.
So I think sports is a great kind of like entry way for kind of like data science.

**Alessio** (2:43)
It was actually my first article on the Twilio blog with a Python script to like predict pricing of like daily fantasy players based on my past week performance.
Yeah, I don't know. It's a good getaway drug.

**Swix** (2:56)
And on my end, the way I got into finance was through sports betting.
So maybe we all have some ties in there. Was like Moneyball a big inspiration, the movie?

**Harrison Chase** (3:05)
Honestly, not really.
I don't really like baseball. That's like the big thing.

**Swix** (3:10)
Let's call it a lot of stats.

**Harrison Chase** (3:11)
A lot of stats.

**Swix** (3:13)
Cool. Well, we can dive right into LangChain, which is what everyone is excited about. But feel free to make all the sports analogies you want that really drives on a lot of points.
What was your GPT aha moment? When did you start working on GPT itself? Maybe on LangChain, just anything to do with the GPT API.