00:35:23 · Molly O'Shea, Jonathan Siddarth
[00:35:23] S0: Okay, but how are you helping close the enterprise gap? [00:35:27] S1: With enterprises, the biggest gap is making sure that you are building custom models, which could be an off the shelf model that you take, but you fine tune it on your proprietary data. [00:35:40] S0: This is what you're doing with enterprises?
00:35:42 · Jonathan Siddarth
[00:35:42] S1: That's right. With enterprises, the old way of thinking was you would take this off the shelf model or you give everybody access to ChatGPT or Claude and call it a day. That's the SaaS era thinking, plug and play. But the best enterprises know that it's not about integrating general intelligence, it's about building proprietary intelligence. You'd want to take a model, maybe from built by one of the Frontier Labs, but you'd want to fine tune that model on your proprietary enterprise data. Your data is your moat. And fine tune it on your proprietary human intelligence. You might automate your proprietary internal workflows. It will probably be a custom model that's on prem. It will probably be a smaller model. Oftentimes we see a 500,000,000 a 10,000,000,000 parameter model. That's the regime, not a trillion parameter model. And this will be faster, more accurate. You'll have full control over governance, confidentiality. You don't want your model to help your competitors. So that's what you want to build. But the constraint for these models is: step one, you first have to prepare your data. Enterprises have lots and lots of unstructured data, but you want to convert it into structured data, into a format to fine tune these LLMs. You'd want a platform to make sure you can distill your human expertise into the models and suck up your proprietary data into a way to fine tune these models. And you'd want to build the right workflows designed for partial autonomy. You'd want to have good evals for your specific workflows. You'd want to solve safety. And enterprises need a neutral evaluator to help them pick the right models, pick the right systems to build this, and that's where we come in. The same technology that we built to help the Frontier Labs get to AGI, we are building a version of this for enterprises to build proprietary intelligence to build their own internal ChatGPT or systems built on top of chat GPT. And that's going to transform. And I'm very bullish for Fortune 500 companies that have good data collection practices. As long as you at the very least maintained good unstructured data from your prior business interactions with your customers, you have an opportunity to win because your proprietary data is your moat as as you fine tune LLMs based on that knowledge.
00:38:26 · Jonathan Siddarth, Molly O'Shea
Most companies have really amazing human expertise inside their organizations. They just have to transfer that expertise from those human minds into those machine minds and that will give them an edge. It will actually give them an edge over a startup starting from scratch. If you're GEICO, long as you do a good job of fine tuning your custom models on your underwriting data, on your claims processing data, and building systems around it to automate underwriting and claims processing, you can win. But if you don't do that, there could be an upstart AI company that starts today, that does this from scratch, that has a good data collection feedback loop. That company could jump you. But that company will still need time to acquire that data. If you're GEICO or Progressive, you have access to data. So it's your battle to lose. But to do that, you have to act fast to make sure that you own your proprietary intelligence and you don't outsource it away. [00:39:34] S0: With over $300,000,000 in revenue, I'm sure this number has changed because that was a couple months ago, but who are your customers? You can talk categorically. You don't have to tell me their names. I know some of these things are sensitive, but, like, what ways in which are you helping these different industries deploy their own models?
00:39:34 · Molly O'Shea, Jonathan Siddarth
[00:39:34] S0: With over $300,000,000 in revenue, I'm sure this number has changed because that was a couple months ago, but who are your customers? You can talk categorically. You don't have to tell me their names. I know some of these things are sensitive, but, like, what ways in which are you helping these different industries deploy their own models? [00:39:56] S1: So we are working with almost everybody that's building a foundation model. We are working with eight out of the nine frontier foundation labs. And now we've expanded to start to work with enterprises in financial services, in insurance, in pharma, and we are rapidly expanding into more verticals. [00:40:24] S0: So this is like investment banks, this is maybe private equity funds, these are like real financial institutions. That's pretty interesting.
00:40:32 · Jonathan Siddarth
[00:40:32] S1: That's right. And if the job involves analyzing data and making relatively quantifiable decisions like verifiable decisions it's a great use case. And for investment banks, private wealth management firms, these other companies in capital markets, if the job is to figure out good investment opportunities and you're looking to find alpha, AI is sort of an existential risk to ignore. Because now with this technology you can ask a question like, hey how many times did the CEO of this public company talk about this particular topic in the last twelve months and how has that changed? In the pre LLM world, this question would been a very sad piece of work for an intern or some analyst to do for many, many hours. But for LLMs this is like nothing. LLMs are really good at synthesizing gobs and gobs of information to come up with useful signal, especially if they are fine tuned on the right data sets. If you are an investment firm and you have a history of evaluating investment opportunities and making decisions, and maybe you decided to buy a stake in a company at a certain price, or maybe you decided to pass company, and you probably wrote an investment memo outlining your decisions, that's incredibly valuable data to fine tune an LLM. And you own that data. That's your edge. Your data is your moat. And I think the best firms are already doing this to help an investment analyst make better decisions faster and eventually with significantly fewer people.
00:42:35 · Molly O'Shea, Molly O'Shea
[00:42:35] S2: Sourcery is proudly sponsored by Carta. Carta is transforming the private marketplace, connecting founders, investors, and limited partners through software purpose built for private capital. Trusted by more than 65,000 companies in over 160 countries, Carta's platform of software and services lays the groundwork so you can build, invest, and scale with confidence. Carta's fund administration platform supports over 9,000 funds and SPVs, representing nearly $185,000,000,000 in assets under management, with tools designed to enhance the strategic impact of fund CFOs. For more information, visit carta.com/sorcery. That's carta.com/sourcery. [00:43:21] S0: How do you give conviction in your investors in outside entities that this is sustainable, durable revenue? The whole image of Mercor, Surge, Scale AI is that this is service based revenue and it's not long term. And we all know the VC world, investors, they love long, sticky revenue. But how do you position this to investors and how do you think about revenue yourself?