The MAD Podcast: How AI Gets Built — with Matt Turck
00:00:01 · S0, George Sivulka, Matt Turck
[00:00:01] S0: Hello, and welcome back to the MAD podcast, a series of conversations with leaders from across the AI, machine learning, and data landscape by FirstSmart Capital. Today's episode features a conversation with George Civolka, CEO [00:00:16] S1: and founder [00:00:17] S0: of Hevia, the AI productivity tool, for a conversation on using generative AI and large language models to save employees in industries like fintech and government from mundane data tasks. This conversation was recorded live at last week's Data Driven NYC. And as always, if you're liking the MAD podcast, don't forget to subscribe, leave us a rating or a comment. Now, here's Matt and George Cebulka. [00:00:47] S2: Very good. Welcome, George, to Data Driven NYC. So you are the CEO of Hebiah. Hebiah was founded in 2020 with a mission to reinvent the way we work. So we're going to talk about what that means. Since then, you have raised $30,000,000 across three rounds of financing from people people here may have heard of, like Peter Thiel, Jerry Yang, a founder of Yahoo, Rams three Ram, one of the first investors in Google, and also our friends at Index Ventures, including Mike Volpe, who spoke at this event, actually in this room, exactly where he's sitting. So congratulations on starting the company and all the success so far. So I'd love to start with your journey. You were like a research scientist, you've done all sorts of interesting things. So walk us through the journey that led to starting the company.
00:01:42 · George Sivulka
[00:01:42] S1: Yeah, happily. And I'll have send Mike a picture so that he can see us up here. The... I think the genesis of Hebia begins pretty far back when I was a PhD student, and I was one of the few math students who actually wanted to teach. What they did was they threw me in a class, which is pretty famous at Stanford, called Math 51, where there is, you know, hundreds, many hundreds of students. It kind of reminds me of this hall. And I got to see those students as they kind of left Stanford, continued on their careers, and years later or months later or whatnot, whenever I synced back in with them, probably the most common job outside of working at Fang was, oh, I'm going to go work, I'm going to be a knowledge worker, I'm going go work at Morgan Stanley or Goldman Sachs. And I had some of my most brilliant students do this, you know, like actually savant level mathematicians that I was scared of would end up, you know, in these other roles. And when I would think back, they would be, you know, you could see in their face, in their micro expressions that they had not smiled in months, you know, and some of them, you know, had gone through like physical transformations, yes, other, you know, it was really, really bad. And the thing that Stanford scratches into your head, kind of imprinted over and over and over again in every entrepreneurial class is, you know, build a company where there's a lot of pain. And I had never seen like such smart people going through so much hell, like doing such stupid things over and over and over again, and you know, I said, well, this is a good place to build a company and forgot about it. And I continued all my research. Was a computational neuroscientist, so I was just kind of, know, heads down, I'm going be an academic. To me, professors made more money than God, I was really excited to be a professor. And at one point, some of my friends in the lab that I was working on started to show me the earliest transformers, this is like 2016, 2017, and there was a distinct moment where I remember them doing something that I'd never seen done before, and I kind of freaked out. And then I saw it again when Google Translate got really good. I kind of freaked out again.
00:04:18 · George Sivulka, Matt Turck
And then my friends started to apply it to search, to neural information retrieval, and I kind of freaked out again, and I freaked out three times, and I said, okay, there's something here. Long story short, I ended up creating a Jupyter notebook, which ran a very early version of a neural information retrieval model, and I gave it to some of my former students that were working in the Morgan Stanley Menlo Park office to look over one type of document called the DefM14A. I don't know if many of you know what that document is. It's like 400 pages long. It's kind of really, really gnarly to process. And they started to copy the Jupyter Notebook and paste it and copy and paste it, and apparently a bunch of people that I'd never met started to use this software. And at one point, it kind of hit me, you know, there's this huge amount of pain. This is a thing that I've heard about three times. I think it's a good time to bet on large language models, and I said, hey, I'll take a leave of absence. And I have never gone back from that leave of absence. The genesis of Hebia and kind of the product that we've built is all around taking the smartest people in the world that do very silly, mundane, honestly stupid tasks, and using the power of large language models to help them not have to do that. [00:05:50] S2: Great, great. What a great story. Turning this into fast forwarding to today, what does Hebia do?
00:05:50 · Matt Turck, George Sivulka
[00:05:50] S2: Great, great. What a great story. Turning this into fast forwarding to today, what does Hebia do? [00:06:02] S1: So, yeah, there's kind of a product evolution as time has gone on and we've refined that customer pain. And what we noticed was, you know, people didn't want to have to learn how to run a Jupyter Notebook if they were bankers. A lot of the time, people didn't really just want a simple search index, like a Chroma or Pinecone database or whatnot, and some gendered response over them. The people that really were power users of Hebia wanted more than anything else to wrangle large language models and apply them in different ways in almost like a WYSIWYG fashion, and so what we built was a productivity tool. It started off being very geared towards search and neural information retrieval and making that readily available to everyone, and then kind of as people added more and more features, we built out a suite of different large language model tools where anyone with no technical training can wrangle them and wrangle any documents to get any sort of output, whether it's an ETL pipeline, whether it's, you know, a search that you'd like to expose to users, whether it's a chat bot that we can spin up for enterprises in two seconds, or even just other ways to extract and find data that you might like. So the one line is we're the first large language model native productivity tool. [00:07:25] S2: Okay. So how does that manifest then for your customers? How do they deploy the product? To do what?
00:07:25 · Matt Turck, George Sivulka
[00:07:25] S2: Okay. So how does that manifest then for your customers? How do they deploy the product? To do what? [00:07:35] S1: What are some of the use cases you observed? A really similar analog to what Hebia does actually harkens back to Excel in 1985, where Excel took a technology, the SQL database, and allowed just normal humans to wrangle it in an application and create value. And they started in a niche in fixed income, and now governments run on Excel, in two thousand years archeologists will be looking at Excel files, you know, and to me it's actually the most important software to ever have been made. On a similar vein, I think that large language models are this new, very powerful technology like SQL once was, and we're trying to build Hebia to be just like Excel honestly, a way to programmatically wrangle in a WYSIWYG way those large language models. And so what that means is it's an application that you can have in your browser or on your computer and you can have any documents that are easily ingestible, and you could have any large language models that are kind of easily loadable over those documents, and you could run different workflows. [00:08:45] S2: So it's a bring your own large language model kind of thing? So if I'm, know, whatever, a bank and I wanna deploy or an asset manager, I wanna deploy Hebia, I can I can pick and choose the models or does it come with battery included? [00:09:02] S1: It comes with batteries included, but you can also pick or choose your models. [00:09:05] S2: Okay.
00:09:05 · Matt Turck, George Sivulka
[00:09:05] S2: Okay. [00:09:06] S1: And I think the thing that's very interesting is no one really cares what the back end of Excel looks like. It just works, you know? And it's really easy to get data into Excel. And it's really easy to run analyses and get data out of Excel in the form that you'd like. For Hebia, a lot of the technology that we built, a lot of what's proprietary is actually that input and output flow and then that flexibility. [00:09:30] S2: And then do you need to customize per use case within an enterprise and then per vertical or is it so powerful that you don't? [00:09:42] S1: We not... So we work with many of the largest asset managers, we work with many banks, we work with the US government and we have not needed to fine tune or kind of personalize any part of the software. In the same way where that would be kind of, you know, crazy to fine tuning Microsoft Word, we've tried to build these things in a way that's very easy to use for anything. [00:10:07] S2: Okay. As I was prepping for this, there were some use cases that were mentioned on the websites around, let's see, financial due diligence, market research, knowledge management. Do you want to talk to those?