Unscripted: What Happened After the Mic Went Off with Douwe Kiela
10:04 Share

Welcome to Founded and Funded. Sometimes the best moments happen after we yell cut. So we couldn't resist sharing this 10 minute impromptu conversation between Madrona partner John Turow and Dao Kiela, founder and CEO of Contextual AI. This happened just after they finished recording their full length episode released last week. Enjoy.

I talk to a lot of enterprise CTOs, as I'm sure you do, and a lot of founders, as I'm sure you do. And I feel like even when this kind of technology is horizontal, we say you go to market vertically or by segment or whatever, but I don't even think that's quite right. I think the storytelling is the thing that becomes vertical or segmented. And when you speak to...

of a bank versus a CTO of a pharma company or the head partner of a law firm or whatever it would be. None of these people, their eyes will glaze over when we start to talk about chunking. If we can talk about SEC filings and the tolerances in there and whatever and a couple of things

really impactful stories that are in the language of those segments, that seems to go so far. And I've seen it myself. And even when astute customers will realize it's the same thing. And so storytelling at a time like this where there's opportunity in every direction you look feels different.

It feels like a thing that can be a superpower for you. That's cool. Yeah, I would love to be better at that actually. Vertical specific storytelling. It's not easy because it's like how vertical do you want to go? We don't want to be Hebbia or even Harvey. We want Hebbia and Harvey to be built on contextual. But the only way to do that is to maybe show that you can build Hebbia and Harvey on our platform. I'll tell you about when I've done it right and when I did it wrong.

So when I did it right was in early days of DynamoDB, which is a managed NoSQL data store. And we said Dynamo is really useful for ad tech, for gaming, and for finance, probably.

And it's because there were key use cases in each of these domains that took advantage of the capabilities of NoSQL and were not too bothered by the limitations of NoSQL. You can only have certain numbers of lookups and things like that. Astute customers could realize you could use Dynamo for whatever you wanted, but we didn't say that ever. All of our market was, we had customer references and we had reference implementations and whatever.

And that kind of helped us, the MBA speakers, cross the chasm. You plant your feet really well. When I've done it badly also shows the power of this technique. I remember I did a presentation about Edge AI. It was like 2016 at AWS re:Invent. Edge AI, and we shipped the first Edge AI product ever at Amazon. And we showed how we were using it with

Rio Tinto, which is a giant mining company doing autonomous mining vehicles. And we chose that because it's fun and sort of sparks the imagination. And we thought it would spark the imagination across a lot of domains. This is a re-invent. So it was on a Thursday, I want to say, a Wednesday or a Thursday that I did that presentation. On a Friday morning before I was going to fly out, I got...

an urgent phone call from the CTO of the only other major mining company of that scale saying, "I have exactly that problem. Can you do the same thing for me?" And I thought, "Well, gee, I aimed wrong because I picked a market of two and I already had one." But it shows that if you really put it in, you know, people don't necessarily use imagination, but if you put it in terms that are that recognizable, they can.

they can see themselves. Yeah. Isn't there the, so I heard that, uh, I think it was, uh, maybe it was Swami or someone senior in AWS said like the, the big problem in the market right now is not the technology. It's people's lack of imagination around. That sounds like a Swami thing to say. Yeah. Yeah. It's very true. Maybe Andy or is that, yeah, I don't know. It could be. And it's, you know, I would also say that the, that's a major role for founders. You know, the, um,

on this spectrum we have like there will be the you know I'll put you in a group with Sergey and Larry right? And so there's the there's the Daoist Sergeys and Larrys there's the Mark Zuckerbergs who are only PHP coders right? And there's the domain experts who are visionaries they're missionaries about solving a real problem and they understand the problem better than other people do and they are not necessarily uh

nuanced in what is possible, but they can hack it together. They can get it to work enough that they can get to a point to then sort of build a team around them. Who is the archetype there? Who is it, Larry and Sergey? So we have Larry, Zuck. I would think about, this is not a perfect example, but I would think about...

pure sales CEOs. - Like Benioff or something. - Yeah, or the guys who started Flatiron Health or Invite Media, they were not oncology experts. They understood their customers really well. Jeff Bezos was not a publishing expert.

And nor was he actually, you know, I'm not sure he wrote code at all at Amazon. I'm not sure he ever checked one line of code in a production. But deep customer empathy and, you know, conviction around that. And being willing, you know, Jeff was, the story with Jeff is that the first book that was ordered on Amazon.com from a non-Amazonian company

was not a book that they had in stock. And the team told Jeff, "Sorry, we gotta cancel this order." And Jeff said, "Like hell we do." And he got in his car and he went to every bookstore in the city. - Barnes and Noble somewhere. - Yeah, and he found it. And then he drove to the post office and he mailed it himself. - Nice. - And he was trying to make a point, but he was also saying, "Look, we're in the books business now, "and we promised our whole catalog, "the first order, better believe we're gonna honor it."

So that's what I think about. And it's, you know, you do things that don't scale and the rest. - Yeah, exactly. Yeah, doing all the crazy stuff. Like everybody is sort of, like all the VCs are saying, like, "Just do it SaaS, like no services, like focus on one thing, do it well." And all of that is kind of true. But if you want to be the next Amazon, then you also have to not follow that word. - You do things that don't scale and you figure out, you know,

You know and I know eventually you can get things to scale. One of the reasons that, and you would know this so much better than I do, one of the reasons Meta invested as early as it did in AI was content moderation. Because that was, you would like a social media business to scale with compute. But it was starting to get bottlenecked by how many content moderators you could get. And that's a lot slower and more expensive. And so...

how quickly and effectively can you leverage that up? And so... That's why they needed AI content moderation. That's why they needed AI. Content, yeah. Yeah. So, it was a lot of fun. Me and Amon were doing all the multimodal content moderation stuff. Really? Yeah, so that was powered by our code base. Wow. In what year? It was like 20...

18, 19, something like that. So that was... We did hateful memes. I don't know if you've heard of this. The hateful memes project. That was my thing. Wow. And so where that came from was because... So content moderation was pretty good on images. And it was pretty good on text. Yeah.

So like if there was some like Hitler image or whatever, or some like obvious hate speech, we would be pretty good. - That's kind of an easy one. - Exactly, but so the most interesting ones, and people had figured this out, was like multimodal. It's like I have a meme, and so like on the surface to the individual classifiers, it looks fine.

But if you put them together, it's like super racist or they're trying to sell a gun or like they're dealing drugs or things like that. So everybody at the time was trying to circumvent these hate speech classifiers by being multimodal. And then Amon and I came in and we solved it. How did you solve it? By building better multimodal models. So we had a better multimodal classifier that actually looked at both modalities at the same time in the same model.

And so you need signal from that from experts who can explain it to you. Yeah, so we had the annotated data. I mean, we just built the framework and we built the data set and we built the models and then most of the work was done by the product team. Got it. Thank you.