Live on Bloomberg TV

00:00I don't even know where to start at the moment. Every day there's a new headline beheads. So when you talk about the soda, the hype and the reality. Where are you on the hype reality spectrum right now in terms of what we're seeing in real world deployment? I add first, thanks for having me on. Excited here to talk to you about a I as you mentioned, we've been investing in a Iot Greylock for over a decade. And we believe the shift that we're seeing now around generative A.I. is equal in importance to this. Earlier shifts from cloud to on premise software and the shift from desktop to mobile. I think every company is going to be any company and everything we do at work and how we live is going to be transformed by A.I.. So I set out thinking a lot of this is real. One mental model that my partner Reid Hoffman and I wrote about to think about how I might impact our lives. Is this concept of a co-pilot for everybody? So get hub co-pilot is an initial co-pilot product that's used by developers that helps developers write better code. Imagine a world where co-pilot scapegoat for lawyers, helping them write better briefs for doctors, helping them do medical diagnoses for salespeople, helping them sell software. I think if we're having this conversation three years from now, all of us will be using these products that fuel like co-pilot in the way we work and live. Set some context for the audience. Reid Hoffman, your partner. He's kind of one of the founding members of Open A.I.. He's on the board. He invested through his own means, his own charity, charitable funds. But I know you guys track what open a ISE during really closely. We'll get onto your portfolio companies that for you right now. Who or what is leading the way in artificial intelligence? I think you have to start with opening ISE. Give them immense credit. What they've done with the transformer model architecture with products like Dolly, GBP 3, Chachi team, new models that we think will come out later this year. They're really at the front lines of bringing these technology to lots of application developers and users. And as you know, they expose these and AP ISE and we're fortunate to be involved with a number of companies that are leveraging those API ISE on the text and the image side to build products that solve different business use cases and consumer use cases. You said that to your mind, every company will be an A.I. company. Could you explain a little bit more what you meant? I think if you look at the capabilities of these models on the generation side, whether they're generating text, image, video code on the classification and discriminative side, looking at a portfolio of loans and predicting what portfolio, which loan is going to default, you're going to see every business have to have any ISE strategy. And we're seeing that already. You talk about Microsoft and the release of their new products from being Google announced, barred and this fast to catch up. There are a number of early stage companies that are leveraging game in different ways. If you don't have any ISE strategy, you're going to miss a very important computing with. This is February of 20, 23, November 20, 22. Chat G.P.S. kind of gets released to the world, yet there are investors like you to have been at this for a while. My question is, how do you feel about that? You know that now there's a broader interest that I guess you've been trying to drum up interest in this well before November 20. Absolutely. As I said, we've been investing in now for more than a decade. And we can talk about some of our companies are scaling today and having real impact. But I think Chachi Petey was an important moment. I've heard it described as the iPod iPhone moment for generative A.I. and I think the power of that interface and the way that end users could actually use it in the real world. It felt like magic and I think it showed everyone how they can be used. But as you pointed out, that we've been investing in there for a long time. So, for example, we're investors in a company called Crossed Out Process, a generative A.I. company that services contact center. So if you called Horizon today, rising contact center agents are using generative A.I. and the way they interact with you, leveraging the technology Christ has built. And that's just one of many examples you brought up that you could name that that will bring up on the screen your portfolio companies. How do you value them accurately? You know, look at those names you're invested in. How do they make money is on the top line? One question, but actually what we're being asked more often is valuations, valuations, valuations. Absolutely. So it's hard to talk about valuations in a general sense. But what I would say is that Greylock, where early stage investors, we invest in a small set of companies and we invest with a 10 year plus time horizon. Right depending on the business. There are different ways to think about valuation. Some of these businesses are enterprise company. So crust as a SAS company. It leverages A.I. to build its source products and with abnormal security, a cybersecurity company that leverages A.I.. These are durable business models that have stood the test of time and we think will lead to high growth, profitable, enduring businesses. I want to ask you about snorkel, because I know that you have published blog posts yourself gonna pull costs to kind of use them as an example. As a case study, how do you discern these startups that are promising? First is the start. You know, what I'm hearing from venture capitalists and your peers is, well, I'm getting 100 DAX through my desk. About 1 percent of them are actually focused on A.I.. The rest just claim to be safe. Using that as a case study. Explain to me how you've identified this as something with real promise. Absolutely. So when we make an investment, we look at two things. One is what is the problem that the entrepreneur and team is trying to solve and what type of market opportunity does it represent? And the second is, do they have a durable, compounding technical advantage in solving that problem? Circle is a great example. So snorkels a data science and machine learning platform for the enterprise. It enables large banks and insurance companies and government agencies to take advantage of all of these advancements. We're seeing in a I when we first met the team, we were blown away by this team's background, having come out of Stanford, having published a number of interesting articles and results around ways I can be used in enterprise contacts. And fast forward to today. They're serving many of the largest Fortune 500 accounts and enabling them to take something that feels like Chachi Petey, but actually use it on top of their own data and in use cases that they can have in production today, driving business value. I would ask you a bit about the read through to enterprise and sort of broad of software companies. There's this idea of right, that Microsoft may give Azure an advantage because those customers using Agile can get access to the open API API ISE. So that makes it more attractive. Do you think that makes sense as a business proposition for the bigger tech companies? I think all of the cloud players are going to need to have any ISE story. I think the relationship that Microsoft has with opening II is a really interesting and smart one. We're going to see similar similar products get released by Google by NWS. That would be my prediction. And I think as you serve the next wave of applications that leverage A.I. as a cloud provider, you have to have a ISE capabilities built in and you better believe that that's going to be a dimension on which customers are going to discern. We have to talk about ethics and we have to talk about concern. The biggest concern is accuracy. When it comes to generative A.I., you'll take place. I'm glad you brought this up. And this is why I'm so happy that we have things like Chachi CBT being these products actually in the real world so that users can interact with them. And these problems get surfaced. That wouldn't happen if these things were being built in the laboratory. I'm confident that these things in the real world with the right guard rail and the human feedback systems, we are going to see dramatic improvement and we're going to see models and products that are trustworthy. Speak the truth and act with high integrity. The phrase A.I. arms race comes up every day now. Is that a fair way of describing what we're seeing in the market right now? I think it is. If you believe, as I do, that A.I. is a wave on the same level of impact as the shift to mobile or the shift to cloud, everything is up for grabs. We're seeing innovation and disruption everywhere. You talked earlier about search. So just one of many markets where we're seeing a lot of disruption. We saw what Microsoft did with being Google, announcing Bard and catching up newer companies like media in our portfolio that are building A.I. native approaches to search. We're going to see it disrupt every market.

No schedule data available.

The Race to Build a ChatGPT-Powered Search Engine

Greylock Partners General Partner Saam Motamedi joins Ed Ludlow to discuss how likely Microsoft's AI-powered Bing search engine is to take over Google. Plus, his thoughts on AI investing trends and the state of the VC ecosystem. (Source: Bloomberg)

More From Bloomberg Technology
All episodes and clips