WEBVTT - Greylock Partner & LinkedIn Co-Founder Reid Hoffman Talks Investing In The AI Sector

0:00:02.320 --> 0:00:06.559
<v Speaker 1>Bloomberg Audio Studios, podcasts, radio.

0:00:06.320 --> 0:00:09.399
<v Speaker 2>News, which is why I'm delighted to speak to Read Hoffman,

0:00:09.520 --> 0:00:12.600
<v Speaker 2>not just LinkedIn co founder and Graylock partner, but a

0:00:12.720 --> 0:00:16.560
<v Speaker 2>serial entrepreneur and somebody who has founded several AI companies

0:00:16.560 --> 0:00:19.200
<v Speaker 2>and will continue to do so. And what I've learned

0:00:19.320 --> 0:00:22.280
<v Speaker 2>is that one hundred million US dollars is the clearing

0:00:22.360 --> 0:00:25.800
<v Speaker 2>price in the market for frontier model AI talent. But

0:00:26.520 --> 0:00:29.880
<v Speaker 2>we're laughing, but actually that's the debate here. Why is

0:00:29.880 --> 0:00:32.800
<v Speaker 2>one hundred million dollars reasonable or unreasonable?

0:00:33.320 --> 0:00:36.360
<v Speaker 1>Well, so, essentially, one of the things that's good about

0:00:36.400 --> 0:00:39.159
<v Speaker 1>having an open marketplace of anything, including labor, is that

0:00:39.200 --> 0:00:42.120
<v Speaker 1>you essentially get multiple bidders. And that's what's that a price.

0:00:42.520 --> 0:00:44.479
<v Speaker 1>And so the fact that someone goes, I'm willing to

0:00:44.479 --> 0:00:47.120
<v Speaker 1>spend one hundred million dollars in order to get this talent,

0:00:47.400 --> 0:00:51.239
<v Speaker 1>you think that's crazy. Why would it possibly higher than

0:00:51.479 --> 0:00:54.880
<v Speaker 1>the vast majority of CEOs of contruc Well, if that one,

0:00:55.320 --> 0:00:59.640
<v Speaker 1>what you're judging is that one talent will potentially create

0:01:00.000 --> 0:01:02.520
<v Speaker 1>millions of dollars of value for you, whether it's in

0:01:02.680 --> 0:01:06.000
<v Speaker 1>a social network, whether it's in you know kind of uh,

0:01:06.200 --> 0:01:08.200
<v Speaker 1>you know, our search engine anything else as a way

0:01:08.200 --> 0:01:11.320
<v Speaker 1>of doing it. And so if you think that, then

0:01:11.360 --> 0:01:12.720
<v Speaker 1>that's just a bold bet.

0:01:13.040 --> 0:01:15.480
<v Speaker 2>What Sam Altman, the CEO of open ai, said was

0:01:15.480 --> 0:01:19.000
<v Speaker 2>that that particular figure which he had accused Meta of

0:01:19.000 --> 0:01:21.920
<v Speaker 2>offering to some open ai staff, what he claimed was

0:01:21.959 --> 0:01:24.800
<v Speaker 2>it was crazy that he sounds like, it's not crazy.

0:01:24.840 --> 0:01:27.560
<v Speaker 1>Really, well, I don't think it's crazy when we know

0:01:27.640 --> 0:01:30.679
<v Speaker 1>that AI is going to transform all businesses in all industry,

0:01:30.720 --> 0:01:33.839
<v Speaker 1>in the entire tech industry. And these are trillion dollar

0:01:33.959 --> 0:01:36.400
<v Speaker 1>companies et cetera at the high end and the whole

0:01:36.440 --> 0:01:39.360
<v Speaker 1>stack all the way through undown in a billion today.

0:01:39.400 --> 0:01:40.840
<v Speaker 2>For a brief moment, one of them was a four

0:01:40.880 --> 0:01:43.240
<v Speaker 2>trillion dollar company exactly please continue.

0:01:42.959 --> 0:01:47.080
<v Speaker 1>Yes in video. And so the uh you know, so

0:01:47.720 --> 0:01:51.400
<v Speaker 1>saying hey, this piece of talent is a risk adjusted

0:01:51.440 --> 0:01:55.000
<v Speaker 1>bet that that makes a difference in that that's just

0:01:55.520 --> 0:01:58.600
<v Speaker 1>a risk bet. You might say it's a dumb risk bet,

0:01:59.040 --> 0:02:02.040
<v Speaker 1>but it's not done to do if those are the

0:02:02.040 --> 0:02:02.840
<v Speaker 1>stakes you're playing for.

0:02:03.760 --> 0:02:07.000
<v Speaker 2>I'm a student of Silicon Valley history. You were there

0:02:07.240 --> 0:02:10.200
<v Speaker 2>with respect, and there was a time where CEOs would

0:02:10.200 --> 0:02:13.080
<v Speaker 2>email each other, you know, and say, don't take my people, Okay,

0:02:13.120 --> 0:02:15.200
<v Speaker 2>I won't if you don't take my people. We had

0:02:15.240 --> 0:02:18.239
<v Speaker 2>this sort of poaching agreements saga in the mid to

0:02:18.440 --> 0:02:21.840
<v Speaker 2>early two thousands. Is this different in this? Is it

0:02:21.919 --> 0:02:23.839
<v Speaker 2>just the reality of a marketplace?

0:02:25.320 --> 0:02:27.800
<v Speaker 1>Well, I think it's much. Look. I think when you

0:02:27.880 --> 0:02:30.440
<v Speaker 1>email and say don't take each other people, that's bad

0:02:30.560 --> 0:02:33.839
<v Speaker 1>for anti trust, that's bad for labor laws, that's bad

0:02:33.919 --> 0:02:36.560
<v Speaker 1>for the right thing to do relative to talented people.

0:02:36.560 --> 0:02:40.160
<v Speaker 1>And obviously, as the founder of LinkedIn, I'm kind of

0:02:40.160 --> 0:02:43.480
<v Speaker 1>in the hey, we should match the talent towards best

0:02:43.520 --> 0:02:46.720
<v Speaker 1>possible outcome. Now, so that was a bad thing to

0:02:46.720 --> 0:02:49.000
<v Speaker 1>be doing. I'm glad we're not doing it now now.

0:02:49.160 --> 0:02:51.399
<v Speaker 1>What we do still, of course want, is we want

0:02:51.440 --> 0:02:54.480
<v Speaker 1>people to engage in serious work at serious companies, as

0:02:54.520 --> 0:02:57.120
<v Speaker 1>opposed to like, hey, I'm being paid this much here

0:02:57.200 --> 0:02:59.720
<v Speaker 1>this month and that much there this month. You've got

0:02:59.760 --> 0:03:02.040
<v Speaker 1>to be like, no, no, let's get the work done. Let's

0:03:02.040 --> 0:03:04.480
<v Speaker 1>build the new future. That's more of my worry on

0:03:04.520 --> 0:03:06.679
<v Speaker 1>the bidding, not whatever prices it gives to.

0:03:06.639 --> 0:03:10.440
<v Speaker 2>You are no longer a board participant of open Ai,

0:03:10.520 --> 0:03:12.800
<v Speaker 2>so I put that out there. But you understand the

0:03:12.840 --> 0:03:16.440
<v Speaker 2>company and you understand its structure. Why is this sort

0:03:16.480 --> 0:03:19.120
<v Speaker 2>of restructure held up, and is that having an impact

0:03:19.120 --> 0:03:21.280
<v Speaker 2>on their ability to hire or retain talent?

0:03:21.320 --> 0:03:23.760
<v Speaker 1>Do you think I haven't heard that it's had an

0:03:23.760 --> 0:03:26.160
<v Speaker 1>impact a negative impact on our ability had because everyone

0:03:26.160 --> 0:03:28.679
<v Speaker 1>knows that Open Air is one of the amazing new

0:03:28.760 --> 0:03:32.200
<v Speaker 1>tech companies of the current badge. I think that the

0:03:32.240 --> 0:03:35.440
<v Speaker 1>restructure is it's a very challenging thing to kind of

0:03:36.040 --> 0:03:38.320
<v Speaker 1>move from where a five oh one C three to

0:03:38.360 --> 0:03:42.440
<v Speaker 1>where a public benefit corp, especially when you have lawsuits

0:03:42.680 --> 0:03:46.000
<v Speaker 1>being filed and refiled and all of the sudden, you know,

0:03:46.280 --> 0:03:48.720
<v Speaker 1>academics saying it's a terrible thing to happen. And so

0:03:49.200 --> 0:03:52.200
<v Speaker 1>I don't know exactly what the impediments are, but it's

0:03:52.240 --> 0:03:53.920
<v Speaker 1>a challenging process.

0:03:53.560 --> 0:03:56.160
<v Speaker 2>If you were to guess. And again, I'm delighted to

0:03:56.200 --> 0:03:57.400
<v Speaker 2>be set with you on a day where there was

0:03:57.440 --> 0:04:02.000
<v Speaker 2>such newsflow Open AI and io Johnny I thing it closed.

0:04:02.360 --> 0:04:04.360
<v Speaker 2>What do you suspect it is that they're building?

0:04:04.840 --> 0:04:08.360
<v Speaker 1>Well, so I don't know, I'm told, but like if

0:04:08.360 --> 0:04:09.880
<v Speaker 1>it were me, I build a phone.

0:04:09.800 --> 0:04:12.320
<v Speaker 2>A phone because the reason I ask you the root

0:04:12.320 --> 0:04:14.520
<v Speaker 2>of it is, every day, how do I interact with

0:04:14.560 --> 0:04:16.640
<v Speaker 2>a generative AI tool, in particular through which is an

0:04:16.640 --> 0:04:18.040
<v Speaker 2>app through my smartphone.

0:04:18.279 --> 0:04:21.520
<v Speaker 1>Yeah, yes, well exactly, No, No, a smartphone. And look,

0:04:21.560 --> 0:04:24.480
<v Speaker 1>I'm you know, I'm a little older. I still use

0:04:24.520 --> 0:04:26.080
<v Speaker 1>a you know, laptop every swapen.

0:04:27.279 --> 0:04:29.919
<v Speaker 2>You have, as I said at the top, been a

0:04:29.960 --> 0:04:35.479
<v Speaker 2>serial entrepreneur. Recently, you've also backed brain implant space. Yes,

0:04:35.760 --> 0:04:39.160
<v Speaker 2>that is a somewhat crowded field. Why did you move

0:04:39.200 --> 0:04:39.960
<v Speaker 2>into that area?

0:04:40.040 --> 0:04:42.320
<v Speaker 1>Read Well, it's a little bit different than brain implant

0:04:42.360 --> 0:04:45.520
<v Speaker 1>It's it's called SAMAI and it's ultrasound. Okay. Now, by

0:04:45.520 --> 0:04:48.640
<v Speaker 1>the way, ultrasound might actually be one of the really

0:04:48.680 --> 0:04:50.640
<v Speaker 1>great things for both reading and writing from the brain,

0:04:50.680 --> 0:04:52.799
<v Speaker 1>because you can read but you don't have to stick

0:04:52.880 --> 0:04:55.800
<v Speaker 1>things in right. And part of the reason I backed

0:04:55.800 --> 0:04:57.839
<v Speaker 1>it is because they're using AI. I'm basically an AI

0:04:57.920 --> 0:05:01.960
<v Speaker 1>investor to get this ultra low frequency much less energy

0:05:02.000 --> 0:05:04.960
<v Speaker 1>than your phone in terms of focusing it on one

0:05:05.040 --> 0:05:08.880
<v Speaker 1>part of the brain. The tool set for creating brain

0:05:08.960 --> 0:05:12.040
<v Speaker 1>therapeutics on all kinds of things, anything from anxiety, this

0:05:12.200 --> 0:05:16.880
<v Speaker 1>serious dementia stuff is it's an amazing possible tool set,

0:05:16.920 --> 0:05:18.599
<v Speaker 1>and that possibility is the thing I invested.

0:05:18.720 --> 0:05:20.800
<v Speaker 2>I found that so interesting that you moved into that field.

0:05:20.800 --> 0:05:23.640
<v Speaker 2>I'm jumping around a bit that across all of this

0:05:23.720 --> 0:05:26.960
<v Speaker 2>is the anxiety of jobs. Many of your peers have

0:05:27.040 --> 0:05:30.120
<v Speaker 2>come out and made bold proclamations of how many jobs

0:05:30.160 --> 0:05:33.000
<v Speaker 2>would be eliminated by AI. I've not had the chance

0:05:33.040 --> 0:05:35.240
<v Speaker 2>yet to ask you for your prediction, and so may I.

0:05:35.560 --> 0:05:39.120
<v Speaker 1>So yes, So look the far future. No one really

0:05:39.200 --> 0:05:41.720
<v Speaker 1>knows anyone who says definitively there's gonna be lots of

0:05:41.720 --> 0:05:44.000
<v Speaker 1>fewer jobs, or definitively if there's gonna be lots more jobs.

0:05:44.440 --> 0:05:48.560
<v Speaker 1>To say that definitively is to undercut the credibility of

0:05:48.560 --> 0:05:51.479
<v Speaker 1>what you're saying. Now, I describe this as the cognitive

0:05:51.520 --> 0:05:54.640
<v Speaker 1>industrial Revolution. I think it has the same impact as

0:05:54.640 --> 0:05:58.240
<v Speaker 1>the industrial revolution, but now in like knowledge work and

0:05:58.279 --> 0:06:01.200
<v Speaker 1>information work and language work and everything, there's both the

0:06:01.200 --> 0:06:03.320
<v Speaker 1>good and the bad of that. The good is none

0:06:03.320 --> 0:06:05.600
<v Speaker 1>of us are of our current modern society works about

0:06:05.600 --> 0:06:09.560
<v Speaker 1>the industrial revolution. That's what creates the economics and prosperity

0:06:09.600 --> 0:06:14.320
<v Speaker 1>for a middle class, for democracies, for Western societies, et cetera.

0:06:14.800 --> 0:06:17.000
<v Speaker 1>On the other hand, the transition is very difficult, right,

0:06:17.080 --> 0:06:20.279
<v Speaker 1>So the transition into the industrial age was challenging. So

0:06:20.360 --> 0:06:22.440
<v Speaker 1>I think we're both going to get the transition, which

0:06:22.480 --> 0:06:25.320
<v Speaker 1>will be a lot of job transformation, initial job loss,

0:06:26.000 --> 0:06:30.680
<v Speaker 1>eventually I think job regaining and transformation and ultimately a

0:06:30.680 --> 0:06:31.479
<v Speaker 1>lot of prosperity.

0:06:31.480 --> 0:06:34.479
<v Speaker 2>But we need to embrace let's titus all together through

0:06:34.560 --> 0:06:38.120
<v Speaker 2>Reed Hoffman the investor. How have you adjusted your financial

0:06:38.160 --> 0:06:41.960
<v Speaker 2>planning and investment strategy based on the talent what we discussed,

0:06:41.960 --> 0:06:43.200
<v Speaker 2>but it also your jobs outlook.

0:06:43.600 --> 0:06:47.279
<v Speaker 1>Well, so I'm still investing a ton in AI companies

0:06:47.279 --> 0:06:50.640
<v Speaker 1>and tech companies, and for example, people forg going to say, well,

0:06:50.800 --> 0:06:53.320
<v Speaker 1>you don't need fewer software engineers. And actually, in fact,

0:06:53.640 --> 0:06:55.680
<v Speaker 1>I think not only are we going to we're not

0:06:55.720 --> 0:06:58.320
<v Speaker 1>going to need fewer. I think we'll still have they

0:06:58.400 --> 0:07:01.599
<v Speaker 1>call it ten million software engineers, developers, you know kind

0:07:01.600 --> 0:07:05.200
<v Speaker 1>of uh, you know across the you know, Western world.

0:07:06.120 --> 0:07:08.760
<v Speaker 1>But I think what's going to happen is actually even

0:07:08.880 --> 0:07:11.640
<v Speaker 1>you are going to start doing software development because everyone's

0:07:11.680 --> 0:07:14.680
<v Speaker 1>going to have a co pilot for doing some of

0:07:14.720 --> 0:07:17.720
<v Speaker 1>their tasks with a software engineer helping them, right, And

0:07:17.800 --> 0:07:21.040
<v Speaker 1>so we're going to have many, many more software engineers.

0:07:22.000 --> 0:07:25.640
<v Speaker 2>Reid Hoffman, Greylock partner, LinkedIn, co founder, serial founder of

0:07:25.680 --> 0:07:26.560
<v Speaker 2>AI companies,