WEBVTT - Sequoia Partner Konstantine Buhler Talks Agentic AI 

0:00:02.520 --> 0:00:09.080
<v Speaker 1>Bloomberg Audio Studios, Podcasts, Radio News. Hillo, I'm so excited

0:00:09.119 --> 0:00:11.479
<v Speaker 1>to speak to Constantina, and of course it's all about AI.

0:00:11.560 --> 0:00:15.800
<v Speaker 1>He also had a great conversation with Jensen Wango Nvidia. Constantine,

0:00:15.880 --> 0:00:18.680
<v Speaker 1>thank you for joining us. I know and everyone's talking

0:00:18.720 --> 0:00:21.720
<v Speaker 1>about AI. Everyone's talking about also this deal between open

0:00:21.760 --> 0:00:22.720
<v Speaker 1>ai and AMD.

0:00:23.360 --> 0:00:27.400
<v Speaker 2>Your first thoughts, well, it's an exciting new partnership. We

0:00:27.400 --> 0:00:30.920
<v Speaker 2>were with Jensen Huang this morning and the relationship between

0:00:31.000 --> 0:00:34.400
<v Speaker 2>Nvidia and open Ai goes back a long way. In fact,

0:00:34.400 --> 0:00:37.400
<v Speaker 2>we were talking about how their very first AI factory,

0:00:37.479 --> 0:00:40.559
<v Speaker 2>the dgx one, was delivered to open ai back in

0:00:40.600 --> 0:00:44.239
<v Speaker 2>twenty sixteen to Elon Musk actually at the Sequoia offices.

0:00:44.800 --> 0:00:46.920
<v Speaker 2>And so it's just a reminder that all of this

0:00:46.960 --> 0:00:50.360
<v Speaker 2>AI revolution is powered by really impressive hardware.

0:00:50.720 --> 0:00:54.080
<v Speaker 1>You're very focused on agenta AI. We're the biggest opportunities

0:00:54.120 --> 0:00:55.800
<v Speaker 1>and why Agenda AI of.

0:00:55.760 --> 0:00:58.800
<v Speaker 2>Course, so agentic AI is AI that doesn't just make

0:00:58.840 --> 0:01:01.520
<v Speaker 2>a decision, it actually takes an action on it, and

0:01:01.640 --> 0:01:04.720
<v Speaker 2>we are now in the moment where it's really starting

0:01:04.720 --> 0:01:08.199
<v Speaker 2>to make a difference in workflows. I give a couple examples.

0:01:08.720 --> 0:01:12.160
<v Speaker 2>One is actually for a security We're investors in a

0:01:12.160 --> 0:01:16.119
<v Speaker 2>company called Expo that started off trying to find vulnerabilities

0:01:16.160 --> 0:01:19.360
<v Speaker 2>in software and now ranks number one in the world

0:01:19.560 --> 0:01:23.560
<v Speaker 2>on global vulnerability leader boards for finding vulnerabilities and fixing

0:01:23.560 --> 0:01:28.560
<v Speaker 2>them for the good guys. Another example, well, we only

0:01:28.680 --> 0:01:32.360
<v Speaker 2>actually service companies that have shown that it is we're

0:01:32.400 --> 0:01:36.440
<v Speaker 2>finding vulnerabilities on their own platforms. So it's for companies

0:01:36.480 --> 0:01:39.399
<v Speaker 2>that have been approved and gone through the vetting and the.

0:01:39.360 --> 0:01:41.480
<v Speaker 1>Second and you were saying, yeah, the second way to

0:01:41.520 --> 0:01:42.039
<v Speaker 1>do so.

0:01:42.080 --> 0:01:45.399
<v Speaker 2>Another example is our portfolio company, Harvey, which is actually

0:01:45.400 --> 0:01:50.080
<v Speaker 2>helping lawyers to actually complete their work faster. They're now

0:01:50.160 --> 0:01:53.680
<v Speaker 2>at over seven hundred customers and fifty eight countries. But

0:01:53.840 --> 0:01:56.640
<v Speaker 2>just to make it very real, my sister is a

0:01:56.680 --> 0:02:00.200
<v Speaker 2>general counsel at a tech company and she texts me

0:02:00.280 --> 0:02:02.560
<v Speaker 2>a couple of weekends ago and she said, Harvey just

0:02:02.600 --> 0:02:05.880
<v Speaker 2>saved my Friday night And I said, what do you mean.

0:02:06.040 --> 0:02:08.880
<v Speaker 2>She said, well, I got a contract from my CEO.

0:02:09.400 --> 0:02:11.320
<v Speaker 2>Usually it would take three three and a half hours

0:02:11.360 --> 0:02:13.919
<v Speaker 2>to turn it over, used Harvey got it done in

0:02:14.080 --> 0:02:18.160
<v Speaker 2>thirty minutes. And so that's us actually feeling AI and

0:02:18.200 --> 0:02:21.359
<v Speaker 2>agentic KI finishing workflows in the real world, but it

0:02:21.440 --> 0:02:22.040
<v Speaker 2>cost t I.

0:02:22.000 --> 0:02:24.160
<v Speaker 1>Mean, things are getting competitive, right, So how do you

0:02:24.240 --> 0:02:25.400
<v Speaker 1>back the right company?

0:02:26.160 --> 0:02:28.160
<v Speaker 2>So we always look at the team at the end

0:02:28.160 --> 0:02:31.440
<v Speaker 2>of the day, especially for early stage investing. We're looking

0:02:31.480 --> 0:02:35.200
<v Speaker 2>for the founders that are going to endure and persist

0:02:35.240 --> 0:02:37.960
<v Speaker 2>through ups and downs. We know that AI is going

0:02:38.000 --> 0:02:42.480
<v Speaker 2>to have its moments of excitement but also moments of difficulty,

0:02:42.560 --> 0:02:45.040
<v Speaker 2>and we look at the long duration and in long

0:02:45.120 --> 0:02:46.639
<v Speaker 2>duration we think it's undervalued.

0:02:46.680 --> 0:02:49.399
<v Speaker 1>Still, what are the other impacts that actually this could

0:02:49.440 --> 0:02:52.080
<v Speaker 1>have on the economy? But also, you know there's so

0:02:52.160 --> 0:02:54.280
<v Speaker 1>much spence. How are you sure that whatever you spent

0:02:54.400 --> 0:02:57.880
<v Speaker 1>on AI you get the returns in a reasonable time frame.

0:02:58.200 --> 0:03:01.080
<v Speaker 2>Sure. So we're here at the Citial Securities conference with

0:03:01.120 --> 0:03:03.959
<v Speaker 2>a lot of leading investors and they're all talking about

0:03:04.320 --> 0:03:06.840
<v Speaker 2>what's the return on that AI? And that's what twenty

0:03:06.880 --> 0:03:09.240
<v Speaker 2>twenty six is going to be all about. How do

0:03:09.280 --> 0:03:12.840
<v Speaker 2>we actually see that return? And there's a few engines

0:03:12.880 --> 0:03:14.600
<v Speaker 2>that are going to power that. The first is new

0:03:14.680 --> 0:03:18.640
<v Speaker 2>monetization strategies. To date, most of the monetization of AI

0:03:18.760 --> 0:03:21.680
<v Speaker 2>has been subscription revenue, and you pay for a subscription

0:03:22.280 --> 0:03:25.000
<v Speaker 2>and you get a product. Well, there's a couple more

0:03:25.280 --> 0:03:28.440
<v Speaker 2>ways that companies are now monetizing. One is through usage,

0:03:28.760 --> 0:03:30.600
<v Speaker 2>so the more you use the product, the more you pay,

0:03:30.760 --> 0:03:33.600
<v Speaker 2>kind of like a real service. The other is actually

0:03:33.600 --> 0:03:36.520
<v Speaker 2>through e commerce. So last week open ai announced that

0:03:36.560 --> 0:03:40.760
<v Speaker 2>you can actually complete entire purchases on chat GPT, so

0:03:40.880 --> 0:03:43.200
<v Speaker 2>you can actually do research for the product you want

0:03:43.240 --> 0:03:46.720
<v Speaker 2>to buy and buy it through Shopify or ats completely

0:03:46.760 --> 0:03:49.360
<v Speaker 2>on the AI platform. That's something we're excited about. We

0:03:49.400 --> 0:03:51.760
<v Speaker 2>wrote a thesis that this is going to become an

0:03:51.760 --> 0:03:54.840
<v Speaker 2>important way to do commerce, and we believe that will

0:03:55.040 --> 0:03:57.680
<v Speaker 2>continue to be. The other one is new modalities, so

0:03:57.760 --> 0:04:01.640
<v Speaker 2>everybody is used AI for text based modalities. You talk

0:04:01.720 --> 0:04:04.280
<v Speaker 2>to the AI and it responds to you and text.

0:04:04.640 --> 0:04:08.160
<v Speaker 2>We've seen voice hit the mainstream. Voice has gotten so

0:04:08.320 --> 0:04:11.720
<v Speaker 2>good that with companies like Sierra, which is a portfolio

0:04:11.760 --> 0:04:15.440
<v Speaker 2>company that does customer support, you can have entire conversations

0:04:15.560 --> 0:04:20.120
<v Speaker 2>with AI flawlessly, or companies like Listen Labs does market

0:04:20.160 --> 0:04:24.279
<v Speaker 2>research incredibly well with voice. And then in twenty twenty

0:04:24.320 --> 0:04:28.080
<v Speaker 2>six we'll see video start to really take center stage.

0:04:28.120 --> 0:04:31.600
<v Speaker 2>We obviously have seen Sora and Nano Banana and those

0:04:31.640 --> 0:04:34.200
<v Speaker 2>perform incredibly well. It's just the beginning for them.

0:04:34.960 --> 0:04:36.640
<v Speaker 1>Talk to me a little bit about the partnership that

0:04:36.680 --> 0:04:39.080
<v Speaker 1>you have with Citadel Securities. I mean, this was a

0:04:39.120 --> 0:04:42.080
<v Speaker 1>large amount of money from Sequoia and Citadell. Is this

0:04:42.120 --> 0:04:44.960
<v Speaker 1>a blueprint of more of these kind of agreements?

0:04:45.279 --> 0:04:47.800
<v Speaker 2>Of course, So we were the first outside investors in

0:04:47.839 --> 0:04:51.159
<v Speaker 2>Citdel Securities in twenty twenty one. We're proud to have

0:04:51.160 --> 0:04:54.159
<v Speaker 2>invested a billion dollars in the business and a lot

0:04:54.160 --> 0:04:56.359
<v Speaker 2>of the reason we invested in the company was the

0:04:56.480 --> 0:05:00.760
<v Speaker 2>technology platform that they operate on and have built. And

0:05:00.839 --> 0:05:04.920
<v Speaker 2>this technology platform has changed an entire industry. So if

0:05:04.960 --> 0:05:08.400
<v Speaker 2>you think back thirty forty years, the way that securities

0:05:08.400 --> 0:05:13.240
<v Speaker 2>were traded was really in person services. Two people would

0:05:13.520 --> 0:05:17.200
<v Speaker 2>find a price and then transact in person. And they,

0:05:17.400 --> 0:05:20.479
<v Speaker 2>especially Ken, had a vision of the future in which

0:05:20.720 --> 0:05:22.560
<v Speaker 2>you're going to actually be able to discover that price

0:05:23.160 --> 0:05:27.880
<v Speaker 2>using algorithms, predictive algorithms, and then actually be able to

0:05:27.920 --> 0:05:32.240
<v Speaker 2>complete that digitally. And we've actually seen that become the way.

0:05:32.400 --> 0:05:35.520
<v Speaker 2>The vast majority of equities are now traded digitally, and

0:05:35.560 --> 0:05:39.080
<v Speaker 2>it's just the beginning. So only eight percent of fixed

0:05:39.080 --> 0:05:42.920
<v Speaker 2>income is traded digitally. There's another ninety two percent to go,

0:05:43.400 --> 0:05:47.080
<v Speaker 2>we see big expanding markets, and frankly, it also reminds

0:05:47.160 --> 0:05:49.840
<v Speaker 2>us of the AI revolution and the agentic revolution, where

0:05:49.839 --> 0:05:52.359
<v Speaker 2>it starts off manual and then you get AI to

0:05:52.360 --> 0:05:54.640
<v Speaker 2>make a prediction and then like an agent, you actually

0:05:54.640 --> 0:05:55.599
<v Speaker 2>complete it digitally.

0:05:56.360 --> 0:05:58.240
<v Speaker 1>I mean, this is basically an industry being disrupted.

0:05:58.720 --> 0:05:58.840
<v Speaker 2>Right.

0:05:59.000 --> 0:06:01.640
<v Speaker 1>Is there still one industry that we haven't thought about

0:06:01.680 --> 0:06:04.880
<v Speaker 1>that you're looking at, ah that will be significantly disrupted?

0:06:05.480 --> 0:06:10.240
<v Speaker 2>There always is, There always is an industry for disruption,

0:06:10.360 --> 0:06:12.480
<v Speaker 2>but also for innovation. You know, we want to make

0:06:12.480 --> 0:06:16.120
<v Speaker 2>people's lives better. I think an industry that could have

0:06:16.160 --> 0:06:19.040
<v Speaker 2>a lot of innovation that would help everyone is health care.

0:06:20.080 --> 0:06:21.599
<v Speaker 2>Why do I say that, Well, there's a lot of

0:06:21.640 --> 0:06:24.800
<v Speaker 2>spend in healthcare. It's incredibly important, and we're not all

0:06:24.800 --> 0:06:28.680
<v Speaker 2>getting the services that we could. So my wife is

0:06:28.720 --> 0:06:31.800
<v Speaker 2>a surgeon in health care. She has from her health

0:06:31.800 --> 0:06:34.720
<v Speaker 2>system hundreds of patients that could come through and get

0:06:34.760 --> 0:06:37.839
<v Speaker 2>those procedures. She still spends a lot of time writing

0:06:37.880 --> 0:06:41.359
<v Speaker 2>notes and actually doing tasks that an AI can help with.

0:06:41.760 --> 0:06:44.360
<v Speaker 2>In fact, now an AI is helping so her health system.

0:06:44.440 --> 0:06:47.400
<v Speaker 2>She uses open evidence, which is now used by forty

0:06:47.400 --> 0:06:50.920
<v Speaker 2>percent of physicians. She uses a Bridge, which does transcription,

0:06:51.680 --> 0:06:54.680
<v Speaker 2>and that actually helps her complete that busy work faster

0:06:55.160 --> 0:06:58.040
<v Speaker 2>so she can spend more time with patients. I'm very

0:06:58.080 --> 0:07:00.800
<v Speaker 2>hopeful that one of the areas that AI helps with

0:07:00.960 --> 0:07:01.479
<v Speaker 2>is healthcare.

0:07:01.680 --> 0:07:03.560
<v Speaker 1>Constantine, thank you so much for your time today. That's

0:07:03.600 --> 0:07:05.600
<v Speaker 1>Constantine Bluler there from Sequoia.