#19 - Lightspeed India

Balaji Srinivasan (0:00)

All right, Hemant, welcome to The Network State Podcast. So, you're going to introduce yourself, you're at Lightspeed, we worked here at A6 NEET.

Hemant Mohapatra (0:06)

Go ahead. Yeah, thank you for having me. Been looking forward to this for a long time, Balaji. Yeah, so for the folks that don't know me, I'm Hemant. I'm Mohapatra, I'm one of the partners at Lightspeed. Lightspeed is a global inter-capital firm that invests across different regions. Basically, we are a raver on the per-ship edge.

Balaji Srinivasan (0:22)

And India is like 1.8 bill, something like that.

Hemant Mohapatra (0:25)

India managed about 1.5 to 1.8 bill. India capital, very early stage focus. You also have a growth front there. Presence in China, presence in Europe, presence in the US. And the better we don't have an office or a team, we certainly haven't messed them. Like in America, African vacuums. Yeah, and then before Lightspeed, I was at SSEM-Z with MQ. And then I had known you for maybe 7, 8 years before that. Oh, it's a fun story that I don't think you would remember.

Balaji Srinivasan (0:49)

Yeah, go ahead.

Hemant Mohapatra (0:50)

So before SSEM-Z, I was at Google.

Balaji Srinivasan (0:52)

Yes.

Hemant Mohapatra (0:53)

And I was part of this small team of people led by a guy named David Glaser, a genius Louis Dai, who saw the wave of... His thesis was that in 10 years from then, the next Nobel laureate in bio-MPA, biostatistician. Yeah, at Marex, not a vet lab.

Balaji Srinivasan (1:13)

Which just happened with domesticics. Yeah.

Hemant Mohapatra (1:16)

It just happened with me two years back. So, and what was Google is all going to be there? You're going to be the computer, the computer for that. So he pulled together a small group of people, and he called it Google Genomics. And one of the people you want to work with was a company called Consul. That's right.

Balaji Srinivasan (1:32)

My friend.

Hemant Mohapatra (1:33)

Yeah. So you were there. So I remember getting on a bunch of phone calls, trying very hard to get in touch with you, and then getting on a phone call with you, trying to convince you to do all the analysis.

Balaji Srinivasan (1:42)

I may remember this. That's right. And I think we were, I think we were using AWS at the time or something like that.

Hemant Mohapatra (1:48)

Probably.

Balaji Srinivasan (1:48)

And I was just like, I mean, Google was great, but I was, I remember actually, was that the time of App Engine?

Hemant Mohapatra (1:55)

App Engine was popular there.

Balaji Srinivasan (1:56)

Yeah. And the problem was App Engine before Google Cloud was something that if you're just doing exactly what Google wanted you to do, it worked. And, but then if it was almost like a blog, where like if you want to do just a blog, it was fine. If you want to do something outside of it, it wasn't right.

Hemant Mohapatra (2:12)

It was just like Java. I think about Java is it will let you not do anything stupid, but won't let you do anything smart either.

Balaji Srinivasan (2:18)

Yeah.

Hemant Mohapatra (2:19)

Yeah.

Balaji Srinivasan (2:20)

That's right. That's right. And I think, I mean, Cloud has obviously evolved a lot since then. And, you know, with the whole genomics thing, by the way, what's funny about, sorry, we're going to say something more. No, no, no. So with the whole genomics thing, that is a space which should be much bigger. And it's kind of been quietly growing. Like, I mean, 23andMe recently went bankrupt, unfortunately. But with 23andMe, anstry.com, like sequencing has really radically increased, right?

But it hasn't increased as much as it should. Like, everybody should have a personal genome on their computer. And so it should be an important file. It should be an app at the top of mind. And part of the reason it hasn't is that FDA attacked the whole space and basically said that you can't offer interpretations of genetic tests and results and so on and so forth. And we're now very fortunate that it's gotten beating back, because with AI, you can now interpret genetic tests, diagnostic software can do that. It can do that better than a doctor can. The entire concept of, oh, you can't use AI to interpret your own test results was a kind of thing that FDA would have blocked AI from doing. Instead, it's an unregulated mathematical advice. All those things have to be censored. It has to be bounded through a doctor. So we may be able to take a little crack at that now that the regulatory state can't block that anymore. Or at least much less powerful than it was. And so AI-enabled medicine is going to be a big thing, but it required so much infrastructure to break past the regulatory state, the American regulatory state in order to do that. And I don't know if you guys, was that on your radar at that time? No. Not the regulatory part probably.