BG2 with Bill Gurley & Brad Gerstner | MANG VC Gone Wild, Can You Trust AI Valuations? & More

00:00:00.000 | You know, the number one question I get from founders

00:00:02.720 | who come in here,

00:00:03.560 | the number one question I get from my LPs

00:00:05.680 | is where are we in this correction?

00:00:08.160 | What stage of grief are we in, Bill?

00:00:12.100 | - And when does it end?

00:00:13.260 | (upbeat music)

00:00:15.840 | - Hey man, good to see you.

00:00:26.720 | - Good to be seen.

00:00:27.600 | - Of course, you know, to the audience,

00:00:29.320 | do your own homework, make your own investing decisions.

00:00:32.040 | We are not your investment gurus.

00:00:35.080 | So Bill, talk to me a little bit

00:00:36.880 | about what you've been thinking about this week.

00:00:38.600 | - Yeah, so one of the things I've been thinking about lately

00:00:41.040 | and we'll go into it in more depth

00:00:42.800 | is how some of the big decisions

00:00:46.520 | and there've been kind of sequence of events

00:00:49.180 | in the large LLM market,

00:00:51.200 | I think are gonna create some pretty big market distortions

00:00:54.680 | that might be felt, you know, along the way

00:00:57.600 | by a number of different players.

00:00:58.960 | So I wanna go in deep on that.

00:01:01.240 | What about yourself?

00:01:02.180 | - Well, I was prepping for our annual LP update today.

00:01:05.800 | I mean, I've been doing these now for over 15 years

00:01:08.440 | and it's always that time of year where I stop,

00:01:12.120 | forces me to telescope out, I think about valuations,

00:01:15.720 | about what's going on with Mag7 over the last year,

00:01:18.840 | about long run tech compounding, about this AI cycle

00:01:22.800 | and what's gonna happen this year.

00:01:24.720 | Then I also had this, you know, I guess my first tweet

00:01:28.600 | that went over a million views

00:01:30.280 | over the course of the last week,

00:01:33.120 | which, you know, it's still amazing to me

00:01:36.480 | that because of regulatory capture,

00:01:38.980 | people just aren't getting their calcium CT exam.

00:01:41.420 | They're still like enslaved to looking at their cholesterol.

00:01:44.640 | We saw the tragic news on that Warriors coach this week,

00:01:48.440 | you know, so I'm just thinking everybody over the age of 40

00:01:52.120 | needs to get this CT scan, but--

00:01:54.440 | - Hey, let me ask you a quick question on this

00:01:56.160 | before we dive in.

00:01:57.000 | So you shared this with me.

00:01:59.120 | I had had a spiked LDL on just a simple test

00:02:03.640 | and I went and did this and it was pretty simple.

00:02:07.220 | I mean, I was in and out, you know, in a flash.

00:02:09.680 | Like it's not like it was a big invasive thing.

00:02:13.400 | And so if it's so simple and so powerful,

00:02:16.240 | why do you think the establishment's fighting it?

00:02:18.760 | - You know, I think we establish standards of care

00:02:21.860 | in this country and in this case,

00:02:23.440 | the standard of care is to track people's cholesterol

00:02:27.120 | and no doctors have an incentive to do anything

00:02:30.240 | other than standard of care.

00:02:31.600 | And so I think they worry about liability.

00:02:33.560 | I think other doctors just aren't on top of it,

00:02:36.520 | but the reality is a CT scan,

00:02:38.440 | you don't need a doctor referral.

00:02:40.280 | It costs less than a hundred bucks, as you know,

00:02:42.320 | takes less than 30 minutes in and out, non-invasive at all.

00:02:46.000 | And it actually will tell you whether or not

00:02:49.080 | you have plaque in your artery.

00:02:50.840 | So, you know, it's, you know, I saw,

00:02:53.760 | I saw after this tragic death this week,

00:02:56.320 | the head of preventative cardiology at Stanford,

00:02:59.160 | Dr. Marin said, you know,

00:03:01.060 | perhaps had he had a calcium CT, he'd be alive today.

00:03:03.500 | So it's just a no brainer.

00:03:05.000 | I'm thrilled that you did it.

00:03:07.240 | I know we've been on this with all of our friends

00:03:09.760 | and so that's good news.

00:03:12.240 | - Awesome.

00:03:13.120 | - Well, let's dig in.

00:03:14.240 | We may as well start with the hottest topic of the day

00:03:18.220 | in Silicon Valley, which is AI.

00:03:20.180 | You know, you stirred the pot a bit this week.

00:03:23.480 | - And it was reacting to a tweet out of your firm.

00:03:26.160 | So put the tweet up, let's tell the story.

00:03:30.160 | So put the tweet up.

00:03:31.120 | What was the point of the tweet and what was in the graph?

00:03:34.760 | Like what was being discussed?

00:03:37.620 | - You know, as you know, so Apoorv on my team

00:03:40.680 | who helps cover AI,

00:03:42.080 | he started looking into the amount

00:03:44.960 | of venture capital investment.

00:03:46.240 | Everybody's talking about whether or not

00:03:48.280 | we're in a VC winter.

00:03:49.580 | But if you look at the aggregated amount of VC capital

00:03:52.640 | that's being invested, it's a big number.

00:03:54.640 | But if you deconstruct it a little bit,

00:03:56.360 | what you see is that we have this explosion

00:03:59.520 | in venture capital investing

00:04:01.000 | coming out of four companies, right?

00:04:02.800 | He called them MANG.

00:04:04.240 | Microsoft, Amazon, you know, NVIDIA and Google.

00:04:08.480 | And so as you can see on this chart,

00:04:10.840 | we go from almost no venture capital investing,

00:04:14.120 | you know, out of these folks six or seven years ago

00:04:16.440 | to all of a sudden this $25 billion last year

00:04:20.460 | in VC investing, it led to the question like,

00:04:23.140 | why is this happening?

00:04:24.940 | And you know, it's not being distributed equally.

00:04:27.740 | It's only going to a few companies.

00:04:29.340 | So really what's going on?

00:04:31.060 | And so I turn to you, you know, you retweeted it.

00:04:34.260 | You had some things to say, what's going on?

00:04:36.260 | - Yeah, and let me start with two kind of high level thoughts

00:04:41.260 | before I drill in.

00:04:42.660 | The first thing is, you know, it's awesome

00:04:45.500 | that this new paradigm has come along.

00:04:48.700 | One of the things that makes venture capital investing,

00:04:51.860 | being a part of Silicon Valley so much fun

00:04:54.860 | is you always get to move on to the new thing.

00:04:57.600 | It's a learn it all mentality all the time.

00:05:00.620 | And it's just super invigorating when the new thing pops up

00:05:03.900 | and everyone gets to go play with it and talk about it.

00:05:06.460 | And you have to learn it.

00:05:07.380 | And if you don't, you get left behind

00:05:09.020 | and it's a big part of the ecosystem.

00:05:11.240 | So it's exciting that there's a new gold rush

00:05:14.580 | and 'cause I'm about to say something

00:05:17.220 | that's going to sound cynical.

00:05:18.260 | And so I want to start with that.

00:05:20.260 | The second thing is, you know, I don't,

00:05:22.620 | some of the things I'm going to say,

00:05:23.820 | I don't have perfect visibility, obviously,

00:05:26.580 | inside these large companies, exactly what they're doing.

00:05:29.540 | But I have an intuition of what's happening.

00:05:31.380 | If anyone out there, you know,

00:05:33.220 | after I make these statements says,

00:05:35.140 | no, Bill, you got it all wrong.

00:05:36.720 | You know, let me know and we'll correct it

00:05:38.540 | and we'll talk about it.

00:05:40.000 | So what I wrote, I'll just read it.

00:05:41.880 | I said, this is what happens when you invest with credits

00:05:44.920 | that allow you to goose your own revenues.

00:05:47.320 | So I think if we think about this historically,

00:05:50.840 | Microsoft found itself in a position

00:05:53.640 | where it realized, one, that AI could have a massive impact

00:05:58.160 | on the products they already have.

00:05:59.840 | And that's been proven, you know,

00:06:01.540 | in the development world with CoPilot,

00:06:03.520 | now they're implementing it for Office

00:06:05.640 | and all the productivity apps.

00:06:07.560 | So they knew it was powerful.

00:06:10.200 | Second, they felt they were behind.

00:06:12.240 | And so they embraced open AI.

00:06:14.720 | And obviously the relationship between Satya

00:06:17.200 | and Sam is quite well known at this point.

00:06:20.680 | As part of that relationship, they made a quote,

00:06:24.760 | and I definitely use quotes, investment in open AI.

00:06:29.000 | And we've been, the world's been told a big part

00:06:32.160 | of that investment wasn't cash dollars,

00:06:34.840 | but credits for cloud services.

00:06:37.360 | - Correct.

00:06:38.200 | - And what we then speculated happened after that

00:06:41.560 | is the other large cloud service providers

00:06:45.440 | became fearful of loss of relevance

00:06:48.960 | or loss of market share because of this type of transaction.

00:06:53.120 | And so we started seeing copycat transactions

00:06:56.320 | happen along the way.

00:06:57.960 | And so the reason I think that,

00:07:01.120 | well, I'll walk into some of the details

00:07:03.000 | of what could happen.

00:07:03.940 | But what I fear is that this is happening

00:07:06.760 | at such a large level,

00:07:08.240 | and maybe we haven't seen the end of it, right?

00:07:10.080 | I think there's reason you can see more,

00:07:12.760 | that it's gonna create a market distortion.

00:07:14.960 | And if we think back to the last cycle,

00:07:17.800 | I lived through the market distortion

00:07:20.080 | created by zero interest rates

00:07:22.320 | and the Vision Fund and the Vision Fund copycats,

00:07:26.280 | and all of a sudden billions of dollars

00:07:27.760 | are piling into companies.

00:07:28.800 | And the reason this matters,

00:07:30.360 | if you're a player in the ecosystem,

00:07:32.320 | is if there's a massive market distortion,

00:07:35.400 | the rules that you've been taught to live by

00:07:38.280 | can all of a sudden either not apply

00:07:41.000 | or there's new rules that apply.

00:07:42.800 | It can get messy.

00:07:43.640 | The playing field can get messy.

00:07:44.720 | - Okay, so let's break it down for just a second.

00:07:47.180 | So I think I understand what you're saying.

00:07:51.160 | So you're basically saying Microsoft decided

00:07:54.960 | to make a big investment in open AI.

00:07:57.520 | Amazon decided to make a big investment into Anthropic,

00:08:01.200 | just as two examples.

00:08:03.000 | And of the billions of dollars that they're investing

00:08:06.360 | at these very high valuations,

00:08:09.360 | those startup companies need to spend those billions

00:08:14.360 | back on the services from the people

00:08:17.360 | who gave them the money.

00:08:19.160 | - And they have the need, right?

00:08:20.520 | They have the need either for training

00:08:22.120 | or even a lot of them are reselling their software

00:08:26.320 | packaged with the compute.

00:08:29.520 | And so it's like a value added service, right?

00:08:32.760 | We're putting AI on top of a CPU

00:08:35.640 | that you would rent otherwise.

00:08:37.320 | Now, let's think about this from both players' side.

00:08:40.400 | From the big company players' side,

00:08:42.720 | I would just say definitively,

00:08:44.720 | and this could get me in trouble,

00:08:46.600 | but this is low quality revenue.

00:08:49.000 | Like this is flat out low quality revenue.

00:08:51.680 | And I'm certain that they've got their auditor

00:08:55.080 | to sign off on it.

00:08:56.520 | But I don't think there's any way you can argue

00:08:59.080 | that it's high quality revenue.

00:09:00.440 | Here's an example I'll give you.

00:09:02.680 | Well, first of all, it's cashless, right?

00:09:04.240 | It's cashless revenue.

00:09:05.440 | So when the credit's reused, you get zero cash coming in

00:09:09.600 | and we know you have an amortized cost

00:09:11.840 | against that big capex that you made.

00:09:14.120 | - And the reason you have zero cash coming in, Bill,

00:09:16.600 | is because you gave the cash,

00:09:18.840 | you gave it to them as a form of investment

00:09:21.120 | and they're turning around and handing it to you.

00:09:22.960 | So from a company's perspective,

00:09:25.120 | it's like taking out of your left pocket

00:09:26.920 | and putting it in your right pocket.

00:09:28.480 | - Well, yeah.

00:09:29.320 | I mean, a big skeptic would say

00:09:31.040 | you're using your balance sheet

00:09:32.280 | to drive your income statement,

00:09:34.480 | which should be a no-no.

00:09:36.560 | The other way I think to highlight the low quality revenue

00:09:41.600 | is imagine there's a startup.

00:09:43.360 | I came up with a cool name for the startup.

00:09:44.960 | It's called the Ultra Hosting Company, UHC.

00:09:48.040 | So UHC got a bunch of money from venture capitalists.

00:09:52.000 | They built a big server farm

00:09:54.480 | and their only customers they have

00:09:57.600 | are companies that they went out

00:09:59.480 | and gave credits away to as a form of investment.

00:10:02.160 | That's 100% of their customers.

00:10:04.880 | So they have tons of revenue as they reuse these things

00:10:09.160 | and zero cash flow whatsoever.

00:10:11.240 | And so I think that highlights it, right?

00:10:15.400 | 'Cause some people will say,

00:10:16.240 | oh, it's just a small percentage

00:10:17.720 | of this big company's revenue.

00:10:20.000 | I'm like, it doesn't matter.

00:10:21.400 | It's still, it is what it is.

00:10:22.760 | - So Bill, if we steel man why this might be okay, right?

00:10:28.360 | Because I don't hear you saying this isn't illegal.

00:10:30.880 | You're not saying it's a violation of GAAP.

00:10:32.440 | You're not saying they're defying their auditors.

00:10:34.560 | What you're saying, I think at a minimum

00:10:37.760 | is that if Microsoft is putting money in,

00:10:41.040 | or I hear you saying two things.

00:10:42.200 | If Microsoft's putting money into OpenAI at $90 billion,

00:10:46.480 | I think the first thing I'm hearing you say

00:10:48.040 | is you gotta be a little bit skeptical of that valuation

00:10:51.040 | because it's not exactly an arm's length transaction.

00:10:54.000 | - Well, yeah, and here's another,

00:10:55.520 | here's a way to really drill in on that,

00:10:57.440 | going back to Ultra Hosting Company.

00:10:59.560 | Let's assume for the sake of this discussion

00:11:03.560 | that the service being provided is a commodity.

00:11:05.880 | And one could argue putting a bunch of NVIDIA servers

00:11:10.240 | and GPUs in a cluster and renting them to you

00:11:13.400 | is a commodity service, right?

00:11:15.760 | So if I'm competing with you to invest in this startup

00:11:20.760 | for you to turn around and use my commodity service,

00:11:24.720 | isn't there a strong argument

00:11:26.360 | that the way I would get there

00:11:27.600 | is by taking a price up to a level

00:11:29.880 | where you choose me over them?

00:11:32.200 | And so all of a sudden, you've got definitive proof

00:11:35.840 | that the way you would win the war,

00:11:37.440 | and you get another benefit.

00:11:39.360 | You get revenue, you get market share.

00:11:41.640 | So at the very least, there's an argument

00:11:43.600 | of might you be maximizing this versus that?

00:11:47.160 | And so yes, I think the valuations could be superfluous

00:11:50.440 | as a result of this.

00:11:51.440 | And that's one of the many market distortions

00:11:54.000 | that would happen as a result of this activity.

00:11:56.840 | The second thing is this could all come to an end.

00:12:00.120 | And so I think for the big hosting providers,

00:12:03.680 | having revenue that is non-typical,

00:12:08.680 | let's just call it atypical,

00:12:10.600 | could backfire if we reach a point

00:12:13.600 | which this type of activity is no longer done

00:12:17.480 | for whatever reason.

00:12:18.360 | - Okay, now let's describe a scenario

00:12:20.360 | that I think you would have less issue with, right?

00:12:23.040 | If $10 billion was just invested into OpenAI

00:12:28.040 | by the five leading SandHill firms,

00:12:32.640 | and then if OpenAI in an arm's length commercial transaction

00:12:36.160 | decided to buy $10 billion worth of services

00:12:40.160 | from Microsoft to Azure to train GPT-5, GPT-6, et cetera,

00:12:45.160 | you're saying like that's no problem.

00:12:47.640 | The problem that you see here

00:12:49.320 | is that that's coming from Microsoft.

00:12:51.120 | So I guess when we look at this,

00:12:55.760 | the indicator that there's a problem from you might be,

00:13:00.000 | do these companies have the ability

00:13:02.880 | to raise this type of capital?

00:13:04.440 | Because if they're just substituting Microsoft

00:13:06.680 | for somebody else who could provide the services,

00:13:09.160 | that's one thing.

00:13:10.280 | But if they don't have the ability to raise this capital

00:13:13.280 | from alternative sources,

00:13:14.760 | it would seem to be more evidence of your case.

00:13:19.080 | - Well, and there's another element

00:13:21.840 | that you're touching on 'cause we read,

00:13:23.800 | you might've seen an article in the information this week

00:13:26.680 | about how Google's having to change

00:13:29.800 | their compensation policies to keep up

00:13:32.160 | with the comp that's coming out of the major AI companies.

00:13:36.680 | And we've heard talk of what I would call

00:13:40.800 | pretty early secondaries at OpenAI.

00:13:43.160 | And so if, let's just call it accelerated liquidity

00:13:49.160 | is part of what it takes to get a killer AI engineer,

00:13:54.160 | then you will have to have funding other than just credits

00:13:59.360 | to be able to fund those secondaries.

00:14:01.000 | That's one thing.

00:14:03.080 | But yeah, I agree with you.

00:14:04.160 | Let me talk briefly about why I think this could be

00:14:07.200 | a problem for those players

00:14:09.760 | and even for the other smaller companies in the ecosystem.

00:14:14.280 | So I believe that once you have these credits

00:14:18.880 | inside your company,

00:14:20.640 | somebody is certainly gonna make the argument, right?

00:14:24.760 | Or they're either gonna be ignorant to the cost

00:14:27.320 | because it's now not a real cash cost, right?

00:14:30.840 | I have these huge credits that I'm using

00:14:33.440 | and or they'll fool themselves into thinking

00:14:36.880 | if I sell my service below the replaceable cost

00:14:42.400 | of the credit, that's really a negative gross margin sale.

00:14:46.920 | But I bet you a ton of people walk into that world.

00:14:50.840 | And if there's a price war for AI services,

00:14:54.640 | all these kinds of things,

00:14:55.800 | I could easily imagine people pricing below the credit cost.

00:15:00.080 | So just another distortion that could happen.

00:15:02.760 | I think the difficult thing as an investor for Altimeter,

00:15:06.120 | we looked at all of these businesses, all of these models,

00:15:11.400 | and there were two things that were really difficult for us.

00:15:13.760 | One was the complexity of the transaction.

00:15:17.480 | Just looking at open AI and trying to understand

00:15:22.480 | the nature of the relationship with Microsoft.

00:15:25.440 | Again, I wanna stipulate from the start, as I've said,

00:15:28.840 | I think AI is gonna be bigger than the internet itself.

00:15:31.560 | I mean, using compute to build intelligence

00:15:36.560 | is a powerful thing for all of us.

00:15:39.000 | So I'm actually happy as both a human

00:15:41.560 | and happy as an investor that in fact,

00:15:43.960 | these models are getting funded

00:15:46.240 | and that they're being built.

00:15:47.640 | But I'm just saying as an arm's length investor,

00:15:50.000 | looking at these valuations, it was hard for me to even,

00:15:53.320 | and I've been doing this a long time,

00:15:54.920 | to even understand the nature of the security I was buying

00:15:58.560 | and the relationship with these companies.

00:16:00.200 | So I think that's why a lot of firms like Altimeter

00:16:03.560 | have had trouble getting to a yes decision.

00:16:06.080 | Set aside the fundamental decision,

00:16:08.840 | which is can these companies actually generate

00:16:12.720 | a lot of durable and ongoing revenue from this

00:16:15.600 | if we have open source providers

00:16:17.440 | who are going to collapse the price of the market

00:16:19.840 | down to zero?

00:16:20.680 | So for us, we haven't invested in any of these.

00:16:23.280 | These were the two bigger challenges,

00:16:24.680 | but I will say I haven't invested,

00:16:27.200 | but I've sat here and thought to myself,

00:16:29.920 | man, I may be missing the biggest thing in the world

00:16:32.160 | because open AI and Anthropic and these other companies,

00:16:36.440 | their valuations have continued to skyrocket.

00:16:39.480 | The usage is very clear.

00:16:41.760 | The teams that they built are absolutely remarkable.

00:16:44.880 | What they're putting into the market is terrific,

00:16:47.280 | but I do think that it is a really important thing

00:16:49.880 | that you're pointing out, which is at a very minimum,

00:16:54.880 | I think we can say that the participation of MANG,

00:16:59.280 | Microsoft, Amazon, Nvidia, and Google,

00:17:02.560 | is distorting the price in the market

00:17:05.480 | in a way that wouldn't have occur

00:17:07.560 | if it was all arm's length transaction

00:17:09.520 | with financial investors.

00:17:10.600 | Yeah, and I'm also not saying anything negative

00:17:14.560 | about the companies or the technology or what they built.

00:17:17.040 | My primary point is that they've done

00:17:20.520 | these unusual transactions.

00:17:22.600 | That has become, that has been mimicked, right?

00:17:27.400 | It's become a bit viral and the competitors

00:17:30.120 | have had to do it as well.

00:17:31.760 | And now it's at such a scale where I think

00:17:34.080 | it could distort the market that we're in.

00:17:36.480 | In the ways that I talked about.

00:17:37.800 | I'll tell you one other negative externality of this,

00:17:40.200 | one other fallout of doing this, right?

00:17:42.840 | If we have the largest companies in the world

00:17:44.840 | who are effectively anointing the winners

00:17:46.920 | with their capital, it makes it really hard

00:17:49.560 | for a true startup that has to raise money

00:17:51.840 | from venture capitalists who don't have $25 billion

00:17:54.960 | to deploy to be able to raise the capital

00:17:57.560 | required to compete.

00:17:59.360 | It's a sport of kings, sport of kings.

00:18:01.840 | And look, one person said to me,

00:18:04.680 | well, if Amazon can't buy a vacuum cleaner company,

00:18:07.840 | what are they supposed to do with their capital?

00:18:09.560 | So it may be that the lack of M&A leads to people

00:18:14.560 | to be more experimental with how they wanna deploy CapEx

00:18:20.800 | and get usage off of that CapEx.

00:18:22.880 | That could play a role here as well.

00:18:24.680 | - Well, I think that the thing I'm gonna be looking for,

00:18:27.680 | Bill, we know there is a lot of secondary transaction

00:18:31.440 | being done at that $90 billion.

00:18:33.160 | In all of these, there's a lot of secondary being done.

00:18:35.560 | You've talked at length about secondary

00:18:38.240 | at these early stages as being a warning sign

00:18:41.520 | and not good for the culture of these companies.

00:18:44.920 | One of the things I'm gonna be looking for is,

00:18:47.440 | do they- - By the way,

00:18:48.680 | maybe it's a great thing for the players.

00:18:51.720 | They say, don't hate the player, hate the game.

00:18:54.120 | And in some ways, it's just equivalent

00:18:57.800 | to the latest sports figure getting the breakthrough deal.

00:19:02.800 | I'm talking about for the AI engineers themselves.

00:19:06.000 | - So is the answer to this, Bill,

00:19:08.000 | like if both of these companies,

00:19:09.400 | if OpenAI and Anthropic went public,

00:19:12.240 | washed their cap tables out,

00:19:13.760 | now had to raise the money from the public markets

00:19:15.880 | to the extent they were burning it,

00:19:17.360 | would you feel better about the health of the situation?

00:19:21.680 | - Yeah, by the way, health implies, once again, it's negative.

00:19:27.040 | I just think it's really different.

00:19:29.080 | And when you have different factors,

00:19:33.280 | these large externalities in the market that we all play in,

00:19:37.600 | all of a sudden the rules are different

00:19:39.160 | and the game plays out in a different way.

00:19:41.160 | That's my only warning is watch out.

00:19:43.920 | I think there will be ramifications of having done this.

00:19:47.400 | That's my main thing.

00:19:48.320 | - The one thing I'm pretty excited about

00:19:51.320 | is we have a lot of competition, right?

00:19:54.160 | It's very clear this is viewed internally

00:19:56.960 | as somewhat existential at Microsoft,

00:20:00.160 | Amazon, Nvidia, Google, et cetera.

00:20:03.760 | And you're gonna have a lot of competition.

00:20:06.080 | These companies have incredible balance sheets.

00:20:08.880 | As a result of this incredible investment,

00:20:11.880 | we're probably going to accelerate the path to AGI.

00:20:16.400 | So I'm glad they are deploying their capital,

00:20:18.880 | but I agree with you.

00:20:20.800 | It makes it very difficult

00:20:22.040 | for the venture capitalists out there.

00:20:24.120 | And for other founders,

00:20:25.480 | if you're a founder and you wanna compete in this,

00:20:27.720 | you know, in the model game,

00:20:29.680 | you know, and you don't have their backing,

00:20:31.240 | there is no chance.

00:20:32.960 | - And while we're here, I can't help but mention it.

00:20:35.400 | It's, you know, you mentioned open source.

00:20:38.640 | I personally am just a massive believer in open source.

00:20:43.200 | And there's a topic we're gonna get to later

00:20:45.040 | that I think we'll all come back to it.

00:20:47.520 | But it's so powerful for society

00:20:50.280 | that these ideas can be shared so openly.

00:20:53.280 | And for me, it's been a sad reality

00:20:57.080 | that some of the larger LLM players

00:20:59.760 | have literally attacked open source directly

00:21:02.880 | and are telling, you know, regulators to try and disable it.

00:21:06.840 | And when I see that, you know,

00:21:09.040 | because I've never seen it this early in a market,

00:21:11.600 | well, I've never even seen attacking open source before.

00:21:15.440 | And I recognize it's highly competitive,

00:21:18.000 | but when I see it, it makes me skeptical.

00:21:19.840 | And I look at the LLM models.

00:21:22.040 | I mean, they're scaled up on parameter count

00:21:24.440 | and the width of the attention window.

00:21:27.040 | There easily could be limitations to that.

00:21:29.760 | Like, if you just think about how optimization models work,

00:21:34.760 | they could run out.

00:21:35.760 | Like, it's not infinite scaling on those types of things.

00:21:39.240 | And then a very smart,

00:21:41.240 | I was having a conversation with Melanie Mitchell

00:21:43.160 | from the Santa Fe Institute is a very smart AI specialist.

00:21:47.200 | And she thinks data may be what causes the asymptote.

00:21:50.720 | In other words, what new data

00:21:52.640 | are you going to put in the training model?

00:21:54.320 | It's already sucked everything up.

00:21:56.160 | And so those things could cause a bit of an asymptote,

00:22:00.680 | a bit of a ceiling and the open source models,

00:22:03.440 | at least on all the performance tests that are being published

00:22:06.160 | are just running fast behind.

00:22:08.440 | And so I could see why the players might try

00:22:13.400 | and cut off open sources needs.

00:22:15.920 | It makes me disgusted.

00:22:18.120 | I really hate it.

00:22:19.320 | - I also think it's impossible.

00:22:20.680 | You saw that thread out of Zuckerberg this week,

00:22:24.720 | where he said by the end of the year,

00:22:25.840 | they're going to have the equivalent

00:22:26.880 | or this year 600,000 H100 GPUs running in super compute.

00:22:31.880 | And that he was, and there's been a lot of debate on this

00:22:35.880 | and he made it clear, they're committed to open sourcing AGI.

00:22:39.880 | And it just reminds me of the conversations

00:22:43.600 | that you've seen Elon have on X.

00:22:46.840 | And he's raising capital for his own X.AI.

00:22:51.280 | But we have a company called OpenAI

00:22:54.440 | that clearly is a closed model and we have-

00:22:57.320 | - And used to be open source.

00:22:58.760 | - Right.

00:22:59.600 | - And now attacking open source.

00:23:00.880 | - And we have the founder who everybody thought was,

00:23:05.200 | perhaps not in favor of open source

00:23:07.360 | who's actually running the open source model now.

00:23:09.320 | So I think the competitive landscape looks great.

00:23:12.680 | We have not seen any bumping up against those scaling laws.

00:23:16.120 | I think, you know, Alad said in a tweet yesterday,

00:23:19.560 | we're going to have four or five companies

00:23:20.960 | that hit chat GPT-4 level this year with their models.

00:23:23.960 | I think that's exactly right.

00:23:26.280 | I think it's going to be an exciting time,

00:23:27.600 | but maybe, what do you think about-

00:23:29.280 | - We can circle back to-

00:23:30.440 | - Let's jump over to our second topic.

00:23:32.160 | - Okay, great.

00:23:33.000 | So you just had your big LP meeting.

00:23:36.040 | Why don't you, if you're willing to,

00:23:38.240 | pull the curtain back a little bit

00:23:40.240 | and tell the world what you're talking about.

00:23:42.960 | - Well, you know, as you know,

00:23:45.360 | we cover both public and venture markets.

00:23:48.280 | And, you know, I had this thesis 15 years ago

00:23:52.280 | as a founder that venture capital companies

00:23:56.640 | were going to stay private longer.

00:23:58.000 | They were going to scale faster.

00:23:59.400 | They were going to have more impact on the public markets.

00:24:02.400 | I think that's played out exactly

00:24:03.920 | how we thought that it would.

00:24:05.160 | I mean, you have 40, 50, in the case of ByteDance,

00:24:08.120 | a $300 billion still venture company,

00:24:11.400 | private company that's not public.

00:24:13.200 | And so the insights you can glean from being in that market

00:24:16.400 | really in order to the value on the public side

00:24:19.440 | and then from public back to venture.

00:24:21.600 | This morning, you know, like the things we went through,

00:24:24.360 | it causes me to telescope out and think about this.

00:24:26.560 | Last year, we saw this multiple expansion.

00:24:29.720 | I would call it really reversion to the mean

00:24:32.080 | as the market played catch up to the big pullback, right?

00:24:35.680 | Remember 2022, and maybe we can pull up this chart

00:24:38.800 | that we showed on software valuations.

00:24:41.480 | But, you know, if you think about the pullback that

00:24:43.720 | happened, Mike Wilson said, we're

00:24:46.920 | going to have a hard landing at the beginning of last year.

00:24:49.680 | Larry Summers was causing a panic about interest rates.

00:24:53.400 | Things like Meta and Uber and Nvidia,

00:24:55.760 | they all had massive pullbacks.

00:24:57.080 | Think about it.

00:24:57.720 | Meta was trading at six times earnings.

00:25:00.880 | And so if you look at this chart, what this chart shows,

00:25:03.640 | and this is public software kind of valuations.

00:25:07.080 | And you can see that we were over 100%

00:25:10.760 | above the 10-year historical valuation,

00:25:14.280 | that blue line in the middle of '21 and '22.

00:25:18.000 | We troughed at the beginning of '23 at about 35%.

00:25:22.000 | And what was that, Brad?

00:25:23.080 | Was that this kind of, we live in a COVID world forever now,

00:25:27.480 | software is the only place that the world exists?

00:25:30.760 | That kind of thinking.

00:25:31.680 | Right, remember, we were talking about this chart

00:25:35.600 | in our own chats.

00:25:37.240 | And we were saying, man, this doesn't make any sense.

00:25:40.560 | But we knew it was zero interest rates,

00:25:43.240 | the ZURP environment that was leading to this.

00:25:45.200 | I just listened to a pod you and I did at Soan Conference

00:25:48.240 | in May of 2021, Bill.

00:25:50.360 | I should have been able to save myself a lot more money

00:25:52.600 | because we were worried about interest rates and multiples

00:25:55.760 | and inflation in May of '21 before the Fed started acting.

00:25:59.320 | And that was part of it.

00:26:00.960 | But remember, everybody said, well,

00:26:02.640 | maybe it's different this time.

00:26:04.560 | Because we pulled forward digitization,

00:26:07.640 | nobody's going to leave their house.

00:26:09.600 | Everybody's going to have to buy everything online

00:26:12.000 | and do everything online.

00:26:13.080 | So I think people tried to justify these multiples.

00:26:17.000 | But at the start of last year, we

00:26:19.800 | were down to 30%, 35% below the 10-year historical average.

00:26:26.440 | We had a big run up.

00:26:27.640 | You saw a lot of these names, Meta, Uber, Nvidia,

00:26:30.360 | up over 100% over the course of the last year.

00:26:35.920 | But that brings us to where we are today.

00:26:38.120 | So you really got to go.

00:26:39.200 | Maybe we can bring up this next chart, which

00:26:42.920 | is we shared this with our investors.

00:26:45.760 | We said, OK, great.

00:26:47.240 | The beginning of 2023 was an incredible opportunity,

00:26:51.680 | Meta trading at six times earnings.

00:26:53.880 | But what about today?

00:26:55.160 | Well, this shows you the multiple expansion

00:26:59.000 | that we saw in '23.

00:27:00.680 | And now both tech and non-tech are

00:27:03.880 | trading at a premium to the 10-year average.

00:27:07.440 | So the blue line represents tech.

00:27:09.400 | This is the Qs.

00:27:10.600 | We're trading at about a 36% premium to the 10-year average.

00:27:15.440 | I just showed you software.

00:27:16.720 | Software is still trading at a discount

00:27:18.880 | because people are more skeptical of software.

00:27:21.160 | But when you look at the Qs, Nvidia, Microsoft,

00:27:24.200 | some of these bigger names, it really shows you that premium.

00:27:26.920 | And then even non-tech-- and this

00:27:28.880 | is the one that's a bit of a head scratcher to me--

00:27:31.440 | non-tech is trading at a premium, a 12% premium

00:27:36.040 | to the 10-year average.

00:27:37.600 | And then finally, a show by comparison.

00:27:42.240 | If we look at this next chart, the big three, what I call

00:27:45.520 | Microsoft, Nvidia, and Meta, they, too,

00:27:50.240 | are trading at a premium, about a 14% premium.

00:27:53.400 | But if you look at their peg ratio,

00:27:54.960 | this is growth-adjusted on the bottom of that slide.

00:27:59.600 | They're basically trading in line

00:28:01.520 | with where they've traded the last 10 years.

00:28:05.320 | So what does all this tells us?

00:28:07.840 | It tells--

00:28:08.800 | Hey, can I-- let me interrupt.

00:28:10.800 | And then we're going to go back to what does all this tell you.

00:28:15.440 | One thing that we've never spent a lot of time talking about

00:28:18.880 | is that the largest companies in our world

00:28:22.920 | have some of the highest growth rates.

00:28:25.400 | And I think that's unprecedented.

00:28:28.080 | Why do you think that's happening?

00:28:30.640 | Well, I'll tell you.

00:28:31.960 | Remember, back at Harvard Business School,

00:28:35.080 | they taught us the diminishing returns of scale.

00:28:38.880 | Remember the-- I think it was Lou Gershner wrote the book

00:28:42.000 | Elephants Can Dance?

00:28:43.200 | Yeah.

00:28:44.200 | There was this idea that elephants can't dance,

00:28:46.800 | that companies get large, they can't innovate,

00:28:49.880 | they can't earn a great return on capital.

00:28:52.280 | And so eventually, they get competed away.

00:28:56.120 | I would suggest that we actually have a new phenomenon going on

00:28:59.480 | in the world, which is an increasing advantage of scale,

00:29:03.480 | not a decreasing advantage of scale.

00:29:05.800 | And why is that?

00:29:07.000 | Well, because if you want to train on 600,000 H100s

00:29:12.480 | like we just talked Zuckerberg doing with the Lama 3 model,

00:29:15.840 | you have to have a business that is generating

00:29:18.880 | the massive cash piles that he's generating in order

00:29:22.040 | to do that.

00:29:22.960 | And so I do think, like you said,

00:29:25.280 | it's a very unique moment in time.

00:29:28.400 | Now listen, that doesn't mean that you're there forever.

00:29:31.800 | I've been out there saying Google's got a lot of challenges

00:29:34.960 | as they try to transition their search monopoly to their answer

00:29:38.680 | monopoly.

00:29:39.640 | I think they're wrapped around an axle in terms

00:29:41.920 | of doing the things they need to do to catch up and to compete

00:29:47.440 | and to protect that search monopoly.

00:29:50.160 | But I do think that this is a moment in time

00:29:52.200 | where there are those increasing advantages of scale.

00:29:55.440 | Listen to this figure.

00:29:56.960 | We expect-- our forecast is that Microsoft and Amazon,

00:30:04.480 | Snowflake, they'll all accelerate their growth rates

00:30:07.040 | this year.

00:30:08.080 | Accelerating your growth rate-- as an old stock analyst,

00:30:11.160 | accelerating your growth rate at that scale

00:30:13.960 | is unheard of, unheard of.

00:30:16.680 | And so that's kind of the takedown, that we were--

00:30:20.280 | OK, so I took you off on that.

00:30:22.480 | Let's go back.

00:30:24.440 | You presented a lot of data, a lot of historic--

00:30:29.400 | to here we are now.

00:30:30.760 | Now what does it mean?

00:30:32.000 | What does that mean for people going forward?

00:30:34.040 | Yeah, I mean, listen.

00:30:34.960 | I think that, as the charts have showed, all attack,

00:30:39.720 | if you look at the Qs combined, they certainly aren't the deal

00:30:43.400 | they were at the start of '23.

00:30:45.400 | Like, the amazing thing is that we had the lowest exposures

00:30:49.240 | at the start of '23.

00:30:51.240 | Investors were so nervous they weren't investing,

00:30:54.080 | and yet things were being given away for free.

00:30:56.560 | And now we see some of that money coming off the sidelines,

00:30:59.400 | out of those money market accounts,

00:31:00.880 | into all of these names.

00:31:02.880 | After, meta has moved from $90 a share to $380 a share.

00:31:08.800 | I still think there are great returns to be had here,

00:31:11.680 | but the returns are much more normalized.

00:31:13.840 | I think the return to target in our portfolio is 20% to 30%,

00:31:18.080 | whereas the start of last year, Bill, it was like 80%.

00:31:22.480 | And we saw those returns play out.

00:31:24.920 | So that's the big thing.

00:31:25.880 | The world is normalized.

00:31:27.440 | Multiples have normalized.

00:31:28.840 | It's not going to be as easy as it was at the start of '23.

00:31:32.120 | I want to show one other chart, because I

00:31:33.960 | think this is the thing that we don't talk

00:31:37.600 | about enough in this business.

00:31:39.000 | And this is long-term, long-run compounding in tech.

00:31:43.600 | You know how it goes with our friends.

00:31:45.120 | We're always talking about much more shorter-term stuff.

00:31:50.160 | But this chart, I had my team pulled together.

00:31:53.440 | And I said, if we look over the last 10 years,

00:31:58.600 | what have earnings compounded at for tech versus non-tech?

00:32:03.680 | And what have stock prices compounded

00:32:05.800 | at at tech versus non-tech?

00:32:07.520 | Is all this just a bunch of crazy people

00:32:09.440 | in Silicon Valley that are running up prices and multiples?

00:32:12.520 | Or is there a reason that tech has grown faster?

00:32:16.440 | And if you look at this, over the last 10 years,

00:32:20.000 | technology companies have compounded earnings at 13%.

00:32:24.640 | And their stock prices have compounded at 17%,

00:32:27.720 | so a little bit faster than earnings have compounded.

00:32:30.920 | But if you look at non-tech, so if you take the S&P

00:32:33.680 | and you strip out all the tech companies,

00:32:35.800 | they've only compounded earnings at about 6%.

00:32:40.600 | And their stock prices have grown at about 8%.

00:32:43.800 | So let me ask you this question, Bill.

00:32:46.280 | Technology has gone from 5% of global GDP

00:32:49.440 | to 15% over the course of the last 15 years.

00:32:52.800 | When we have this conversation five or 10 years now,

00:32:55.600 | is tech going to be more or less than 15% of global GDP?

00:33:01.120 | I'm going to say more, but with an asterisk.

00:33:04.320 | So your chart has left off 2009 and 1999.

00:33:10.920 | And if you were--

00:33:13.440 | the argument you're making can be used holistically

00:33:16.680 | at any point in time.

00:33:18.360 | But if your point of entry is '99, '09, or the top of--

00:33:24.080 | the end of 2021 here, you're not in a good place.

00:33:28.640 | So maybe you have to--

00:33:30.240 | what do you call when you roll in?

00:33:32.240 | Like dollar cost average or something?

00:33:35.320 | Price of entry matters.

00:33:36.880 | But what I would suggest to you is it's almost certain

00:33:40.240 | that tech's going to be a larger portion of the global GDP

00:33:42.960 | in 5, 10 years because technology is becoming

00:33:46.520 | more important every day in every company's life,

00:33:48.600 | every person's life.

00:33:49.720 | The second thing I would tell you

00:33:51.200 | is that I think that technology in aggregate

00:33:53.600 | will continue to out-earn non-technology companies.

00:33:57.800 | Why?

00:33:58.280 | Because of the age of efficiency.

00:34:01.120 | Why? Because of an AI super cycle.

00:34:03.680 | And so what I said basically, if you're playing from home,

00:34:08.800 | I think the biggest free lunch in all of investing

00:34:12.360 | is the asymmetric bet over the long haul on technology

00:34:16.560 | companies compounding.

00:34:17.720 | Now, the hard thing, as you know, Bill,

00:34:19.680 | is are you able to pick the ones that do the best versus the

00:34:23.400 | ones that don't do great?

00:34:24.560 | Because we know lots of technology companies

00:34:26.720 | get wiped out, get wiped out.

00:34:29.040 | So I mean, I think unless you do it for a living,

00:34:31.240 | you probably just buy the index and you're betting

00:34:33.320 | on technology out-compounding.

00:34:36.080 | For us, we have a lot of fun trying

00:34:39.080 | to pick the ones that are going to be the winners versus--

00:34:41.480 | What are the best--

00:34:42.680 | I don't know if you know, but just for listeners,

00:34:45.240 | what are the best tech indexes?

00:34:48.160 | Or are there other ones?

00:34:49.600 | Listen, there are a lot of ETFs like growth software ETFs

00:34:52.880 | or growth internet ETFs.

00:34:55.600 | You can invest in the Qs.

00:34:57.120 | It turns out that even the SPY, which is the S&P 500 index,

00:35:01.800 | is very quickly becoming a tech index.

00:35:04.840 | And so I think a big mistake-- here's an interesting one.

00:35:07.880 | A big mistake that a lot of technology investors have made

00:35:12.680 | is nobody feels smart just investing

00:35:15.600 | in the big companies.

00:35:16.760 | They want to present to their friends the company nobody's

00:35:20.000 | ever heard of that turns into a 10-bagger or a 50-bagger

00:35:22.960 | or a 100-bagger.

00:35:24.280 | It's a much more interesting conversation

00:35:26.040 | to have a cocktail party presenting that idea

00:35:28.560 | than it is presenting the idea for Microsoft, which

00:35:31.160 | is going to compound at 15%, which

00:35:32.880 | has led investors to be grossly under-leveraged

00:35:37.760 | to the largest companies that have done the best

00:35:40.160 | and over-leveraged to the rest of the technology complex.

00:35:43.200 | And we just mentioned you had an unusual--

00:35:46.600 | I mean, I can't articulate how unusual

00:35:50.240 | it is for the biggest companies to be succeeding,

00:35:52.640 | because my entire career up until now,

00:35:56.880 | the large companies become a laughingstock.

00:35:59.440 | And Microsoft had a window, Freesatche,

00:36:02.520 | where it was in that place.

00:36:04.240 | It was considered to be like HP or DEC or IBM.

00:36:08.160 | And it was just assumed that the startups would come along

00:36:12.080 | and roll them, and that the big company gets stodgy and lead

00:36:15.600 | footed and falls behind.

00:36:17.720 | So this is a different world we're in.

00:36:20.640 | Yeah, I agree with that.

00:36:21.840 | Hey, how about we shift gears here a second?

00:36:26.120 | So that's public markets.

00:36:27.200 | Great year in '23.

00:36:28.480 | '24 going to be a little more challenging.

00:36:30.320 | People panicked when we started the year.

00:36:32.040 | Mag 7 was down, I don't know, 5%.

00:36:34.720 | Now it's up a lot.

00:36:36.800 | And so we'll see where it shakes out.

00:36:39.960 | But VC, man, we've talked a lot about this,

00:36:45.640 | where we are in this VC correction.

00:36:48.520 | The number one question I get from founders

00:36:50.920 | who come in here, the number one question I get from my LPs

00:36:53.840 | is, where are we in this correction?

00:36:56.400 | What stage of grief are we in, Bill?

00:37:00.240 | And when does it end?

00:37:02.440 | Well, look, for reasons that you and I have talked about a lot,

00:37:05.720 | I think the most differentiated element of this correction,

00:37:10.520 | versus '09, versus '01, is the amount of capital

00:37:15.040 | that the companies went into the correction with,

00:37:18.240 | the speed at which they lowered cost afterwards.

00:37:22.320 | So there wasn't denial.

00:37:23.640 | Everyone got along pretty quickly.

00:37:25.640 | And therefore, the elongated window, months of cash--

00:37:30.600 | it'll encourage a lot of our companies

00:37:32.240 | to track months of cash, just burn rate divided by how much

00:37:35.440 | you have in the bank account.

00:37:37.080 | That was elongated.

00:37:38.200 | And so things are taking long to play out,

00:37:41.840 | because the day of reckoning has been pushed out.

00:37:44.760 | Now, one thing that's really wild to me, in '01 and '02,

00:37:50.120 | when a company went bankrupt, it was news.

00:37:52.760 | Every single one of them, when they shut down,

00:37:55.040 | there was this website called F'd Company that literally

00:38:01.280 | tracked them.

00:38:02.080 | And the VCs were--

00:38:03.520 | For those at home who couldn't see your air quotes, Bill,

00:38:06.960 | that was F'd Up Company.

00:38:09.960 | Yeah.

00:38:10.560 | And the guy that ran that site actually

00:38:12.640 | runs a really cool company in the audio space

00:38:17.080 | that helps you post on Spotify.

00:38:19.400 | Anyway, we were all vilified.

00:38:22.720 | We were just taken down.

00:38:24.040 | Oh, you've got these dumb companies.

00:38:26.080 | That doesn't seem to be happening.

00:38:27.880 | You guys stumbled on this chart.

00:38:29.680 | You can pull up--

00:38:30.760 | I think it's from Carta data or something.

00:38:33.280 | But the first 3/4 of 2023 saw--

00:38:38.920 | what is it-- around 180, 1/4 of shutdowns.

00:38:43.080 | And so our run rate-- let's assume the run rate's like 200.

00:38:46.320 | We're at like 800 a year shutdown.

00:38:49.080 | No one's writing about it.

00:38:50.200 | I mean, we had convoy.

00:38:52.040 | There's a few high-profile ones where people wrote about.

00:38:54.840 | For the most part, I guess with the election in Ukraine

00:38:59.200 | and Gaza, there's so much going on that maybe people just

00:39:04.280 | don't give a shit.

00:39:05.000 | But it's wild that it's happening so quietly.

00:39:09.120 | So that's one thing.

00:39:10.560 | I guess it is happening, even though--

00:39:12.360 | I mean, it's not so quiet in Silicon Valley.

00:39:16.520 | Like, I feel for a lot of founders and, frankly,

00:39:21.240 | early-stage venture capital firms,

00:39:23.080 | even when companies are doing well,

00:39:25.160 | they're really having challenges raising

00:39:27.400 | a follow-on round of capital.

00:39:29.400 | Yeah, and if you look at this other chart,

00:39:32.400 | series A and series B type funding, it's way down.

00:39:35.920 | And that's where you take out the credit funding

00:39:39.640 | we were talking about and get that out of the mix.

00:39:42.800 | And so, yeah, it's tough.

00:39:46.280 | You and I have discussed this many times,

00:39:48.320 | but I think the number one thing a founder can do

00:39:52.120 | is to, as quickly as possible, get

00:39:55.680 | in touch with what your real actual valuation is

00:39:59.400 | and then ask yourself, what do I need to do structurally

00:40:03.360 | to give this company a fighting chance,

00:40:05.280 | knowing that that's reality?

00:40:07.240 | And I think a lot of people live in a reality distortion field.

00:40:12.440 | This, to me, is something that I'm deeply passionate about

00:40:18.920 | because I think it's a unique moment in time

00:40:23.160 | right now, where I think, because of what happened

00:40:27.920 | in Zerp, Bill, founder-friendly, this term founder-friendly,

00:40:33.000 | it kind of was like just telling the founder what

00:40:36.240 | they wanted to hear.

00:40:38.120 | Whereas, maybe pull up this tweet from Jam and Ball,

00:40:42.160 | my team last week, and I'll just read you what he said.

00:40:47.960 | Lots of talk about getting fit, but too few boards

00:40:51.480 | and founders are having honest conversations.

00:40:55.800 | I've always preached to the team here,

00:40:57.640 | founder-friendly is doing the right thing

00:40:59.360 | and being truthful with founders,

00:41:01.520 | like the letter to Meta on time to get fit,

00:41:03.600 | saying up front what's too often discussed outside the boardroom

00:41:08.600 | and behind their backs, even if it means making hard decisions

00:41:12.440 | like layoffs, selling, shutting down, down rounds, et cetera.

00:41:17.200 | And so I look at the situation today, frankly,

00:41:21.280 | I don't know whether it's because venture capitalists are

00:41:24.320 | a lot younger, a lot of people just moved into the system,

00:41:27.120 | they haven't seen a drawdown before,

00:41:29.480 | whether they're on too many boards and spread too thin,

00:41:32.280 | so they just can't handle the deluge of conversation.

00:41:35.360 | But my biggest concern is that boards haven't even

00:41:39.360 | been having the tough conversations, right?

00:41:42.720 | We certainly have the examples like you pointed out.

00:41:46.800 | And take something like Klarna, and I

00:41:49.600 | think Sequoia deserves some credit here.

00:41:52.040 | If you look at Klarna or you look

00:41:53.680 | at what happened in Instacart, I mean, Klarna's 2021 round,

00:41:58.440 | right, $46 billion led by SoftBank, Sequoia, Dragoneer,

00:42:03.640 | Silverlake in that round, $46 billion.

00:42:06.400 | They do a round in '22 at $6.7 billion, right,

00:42:11.920 | Sequoia and Silverlake.

00:42:13.320 | That, to me, is a great set of people around a board table

00:42:17.680 | having an honest conversation with the founder

00:42:20.440 | and getting the business reset.

00:42:22.560 | You cannot-- it's not good for the employees.

00:42:26.160 | It's not good for the company.

00:42:28.720 | It's not good for the investors if you're

00:42:31.200 | living in delusion when it comes to your cap table

00:42:35.880 | and to your evaluation.

00:42:37.120 | So I think we're starting to see more momentum of this.

00:42:40.320 | I certainly know in our portfolio,

00:42:42.320 | we're pushing really hard on every board that we're on,

00:42:45.680 | on having those conversations, getting liquid.

00:42:48.080 | If you don't have a model that you feel comfortable with

00:42:51.120 | to get back to that valuation, you can sell the business.

00:42:53.640 | But you've got to sell, restructure, or take

00:42:56.240 | the business public.

00:42:57.120 | I mean, those are the doors.

00:42:58.720 | Yeah, people get overly focused on this last round

00:43:02.000 | valuation thing.

00:43:03.040 | Like, it is--

00:43:04.040 | I can't tell you how many founders

00:43:07.640 | I've had a conversation with where

00:43:09.280 | it's clear the number one objective in their function

00:43:12.800 | about the next financing is to clear

00:43:15.160 | the bar of the last round.

00:43:16.640 | And it just shouldn't matter that much.

00:43:19.280 | Especially after we just went over the waterfall

00:43:21.840 | that we did, you've got to just get that out of your head.

00:43:24.800 | It's so stupid.

00:43:26.040 | You've created an artificial constraint

00:43:29.800 | that's just not that big a deal at the end of the day.

00:43:32.760 | I used to have this chart--

00:43:34.200 | I wish I had prepared it-- but of the logos

00:43:37.000 | of the public companies that traded below their offering

00:43:39.640 | price.

00:43:40.360 | And it's some of the best companies in the world.

00:43:42.760 | It includes Amazon and Salesforce.

00:43:44.280 | And Google.

00:43:45.040 | And yeah, a whole bunch of great companies.

00:43:47.920 | And yet, if you go to IPO and you're arguing over price,

00:43:52.120 | they'll say, the one thing you can't possibly let happen

00:43:55.920 | is you trade below your offering price,

00:43:58.040 | even though some of the greatest companies in the world did it.

00:44:01.000 | And so I think that's a similar kind of silly constraint

00:44:04.160 | that people throw out there.

00:44:05.320 | I would say this about your initial comment.

00:44:08.720 | Some of the board members are just too young.

00:44:11.640 | They have never lived through it.

00:44:13.000 | They don't know.

00:44:14.200 | They were trained in the last 10 years.

00:44:16.240 | They've only seen up.

00:44:18.200 | Some of them are naive, which means

00:44:21.040 | they haven't studied economics and finance to the extent

00:44:23.960 | that they should.

00:44:25.280 | I might point them to one of my favorite blog posts,

00:44:29.280 | the keys to the 10x revenue club.

00:44:31.600 | Because Silicon Valley mostly lives on price

00:44:36.160 | to revenue multiples.

00:44:38.120 | It is the conversation du jour in Silicon Valley.

00:44:42.560 | And it's one of the most naive ways

00:44:44.840 | you could value a financial asset in the world.

00:44:48.360 | And so people just need to sharpen their pencils

00:44:50.800 | to a certain extent.

00:44:51.600 | Yeah, I mean, just a couple of comments there.

00:44:53.800 | I mean, you and I both know, we talked about it even

00:44:57.600 | at the top, price to revenue.

00:44:59.800 | So you see this all the time in software.

00:45:02.360 | It's just a shorthand for investors

00:45:05.440 | to get to discounted free cash flow.

00:45:07.920 | But here's the problem.

00:45:09.320 | Because I live in the public markets,

00:45:11.080 | we do DCFs all the time.

00:45:13.320 | But how many people in Silicon Valley

00:45:15.920 | even build a model on the company, right?

00:45:19.000 | Like there's just not a lot of that that goes on.

00:45:21.000 | If you're doing seed in series A,

00:45:22.800 | that's not what this is predicated on either.

00:45:26.120 | But when you're investing in a business at over a billion

00:45:29.000 | dollar valuation, you sure as the hell

00:45:31.560 | better be building a model and understanding

00:45:34.320 | what the exit value is going to be based

00:45:36.840 | upon a normalized multiple into the public markets.

00:45:40.120 | And so I think way too little of that was going on.

00:45:43.320 | We're seeing these corrections beginning to occur.

00:45:47.200 | But if I go back to my first question

00:45:49.520 | I had for you on this topic, Bill, I went back

00:45:52.520 | and I saw a talk you did in 2012.

00:45:56.280 | And somebody said, oh, 2008 wasn't so bad.

00:45:59.480 | We've bounced right back, et cetera.

00:46:01.480 | And you said, you know, listen, 2008, 2009

00:46:06.440 | wasn't so bad because the bubble never got that big.

00:46:09.040 | And here's the interesting thing about what you said.

00:46:11.720 | You said, I don't know if we'll ever see a bubble again

00:46:16.200 | like we saw in 1999.

00:46:19.120 | So was the ZURP bubble as big as the bubble we saw in 1999?

00:46:25.760 | And will it take us the same four or five years

00:46:30.080 | that it took in 1999 for the world

00:46:32.640 | to work through that entire backlog?

00:46:37.200 | Well, it was wildly different in terms

00:46:39.840 | of one thing, which is the amount of capital

00:46:41.600 | being thrown around.

00:46:42.920 | And so the companies--

00:46:45.760 | in '99, a huge round was $20, $30 million.

00:46:49.880 | And so we went well past that with billion dollar rounds

00:46:53.120 | and whatnot.

00:46:53.920 | And so from that standpoint, I think

00:46:56.360 | you could argue maybe the price height of the bubble

00:47:00.040 | may have been similar, or maybe even lower,

00:47:02.400 | because then we were giving companies

00:47:04.440 | with no revenue huge prices.

00:47:05.880 | But the capital intensity of the bubble,

00:47:09.120 | I would say is a more important metric, was bigger.

00:47:12.720 | It was bigger.

00:47:13.680 | And how long will it take?

00:47:15.320 | I don't know.

00:47:15.960 | I mean, I'd love to see companies

00:47:17.760 | start going public again.

00:47:19.040 | You and I have talked about this.

00:47:20.840 | I don't think the window's closed.

00:47:22.720 | Like, you may not like the valuation.

00:47:24.760 | Like, you've got to decide.

00:47:26.680 | You may be afraid to be public.

00:47:29.080 | You may have had someone tell you it's too hard.

00:47:31.560 | Like, all those things.

00:47:32.560 | But it's not closed.

00:47:33.600 | Like, it's just a matter of price.

00:47:35.520 | And it just takes a while for everyone

00:47:39.800 | to get busy getting busy after you've come through--

00:47:42.200 | For sure.

00:47:42.680 | I may even take it a step further, Bill.

00:47:45.280 | There is a voracious, voracious appetite for IPOs.

00:47:49.720 | Normally, we have over 100 tech IPOs a year.

00:47:53.400 | And we've basically gone two years with no tech IPOs.

00:47:56.720 | Well, we had three, but whatever.

00:47:59.120 | No, you had a handful.

00:48:01.160 | But I'll tell you this.

00:48:02.560 | Yeah.

00:48:03.480 | Most of the folks who are on the buy side, right?

00:48:06.800 | So think about whether you're a long only fund, a capital

00:48:10.160 | group, or TRO, or whether you're Altimeter, or CO2, or Tiger,

00:48:14.200 | et cetera, right?

00:48:15.840 | If all you're buying is Mag 7, you're

00:48:18.200 | feeling a little bit uncomfortable.

00:48:20.000 | You're looking for new ideas all the time, right?

00:48:23.560 | And so I think the only thing holding us back

00:48:26.920 | is to find high-quality companies that

00:48:29.560 | are going to the public markets and accepting

00:48:31.640 | the valuations of the public markets, which

00:48:34.680 | is the market-clearing price.

00:48:36.200 | I think you're going to see the dam really

00:48:39.600 | break on this this year, particularly

00:48:41.240 | in the back half of this year.

00:48:42.520 | And the reason is because people got to raise capital, right?

00:48:46.040 | And I just think they're coming to grips with it.

00:48:48.520 | And I thought we would see some IPOs

00:48:50.600 | that were either related to recapping or raising capital.

00:48:54.320 | And sometimes what you need first

00:48:58.400 | is the derivative instrument that then makes you realize,

00:49:02.520 | oh, shit, I really got to get public

00:49:04.120 | and convert everyone to common.

00:49:05.560 | That happened to Square, actually.

00:49:07.960 | They had done a time bomb derivative thing that

00:49:11.960 | then made you go public.

00:49:13.800 | And so maybe we've got to have those first,

00:49:16.520 | and then we'll have more.

00:49:18.240 | But I would have thought you'd seen more.

00:49:21.240 | I think a lot of people also have this silly argument

00:49:26.720 | that we've got to wait for someone big to go first.

00:49:29.160 | And I actually think we'll probably

00:49:31.480 | see a smaller company with a courageous founder

00:49:35.400 | step through the window.

00:49:38.080 | We don't have to wait around for Stripe.

00:49:39.720 | That's kind of a silly notion.

00:49:41.040 | You asked a question earlier that I just

00:49:43.320 | wanted to touch on I just remembered,

00:49:44.840 | which was, why is it taking longer for founders, maybe,

00:49:47.880 | to come around to the valuation adjustment?

00:49:51.680 | I think one of the reasons is that

00:49:53.720 | with all the secondary liquidity that occurred this time around,

00:49:58.480 | you know the hardest thing is, if you ever anchor your net

00:50:01.520 | worth, if you ever look at, oh, I own 20% of this company,

00:50:05.080 | it's valued at $10 billion.

00:50:06.800 | So therefore, I'm worth $2 billion.

00:50:09.720 | And if you set that anchor in, and worse yet,

00:50:12.200 | if you start living your life that way,

00:50:15.480 | and then all of a sudden, you have this dramatic reset down

00:50:20.400 | 80%, and you're like, oh, I'm not worth $2 billion.

00:50:23.720 | Instead, somebody says, your company's really worth

00:50:26.520 | a billion or $2 billion.

00:50:28.520 | All of a sudden, go back to the stages of grief.

00:50:32.400 | What's that stage?

00:50:33.400 | Denial.

00:50:34.600 | And then anger.

00:50:35.800 | No doubt.

00:50:36.280 | Right?

00:50:37.720 | I think it plays a role.

00:50:38.720 | You're nowhere close to it.

00:50:39.920 | What I finally see happening this year,

00:50:42.080 | I think 2024 in terms of the stages of grief,

00:50:45.240 | is the year of acceptance.

00:50:47.640 | I think people are just going to have to get liquid.

00:50:49.960 | They're going to accept the prices that they have to accept.

00:50:52.440 | Because frankly, there are no more soft banks

00:50:55.000 | to bail them out.

00:50:56.760 | And the public market's not going

00:50:58.120 | to overpay, because the public market knows

00:51:00.400 | what the clearing price is.

00:51:01.800 | And so that, I think, gets founders

00:51:03.920 | to the stage of acceptance.

00:51:05.080 | But I think we probably have another quarter or two

00:51:08.040 | to get there.

00:51:09.200 | I think we've beaten that one to death.

00:51:10.800 | By the way, I'll say something.

00:51:13.520 | You made me think of something that I will share, maybe

00:51:16.280 | on behalf of all venture capitalists,

00:51:18.640 | and maybe to make everyone's lives easier.

00:51:22.920 | One of the easy defaults you go to in the middle

00:51:26.160 | is, oh, I'll just go to my insiders and ask for a bridge.

00:51:29.800 | And I will tell you, at least all the data I've seen,

00:51:33.360 | the success stories coming off a bridge are few to none.

00:51:38.680 | That's why we always refer to them as peers,

00:51:40.880 | rather than bridges.

00:51:42.200 | Peers, as in walking off the end of a pier.

00:51:46.040 | Yeah, like it doesn't get you to the other side.

00:51:49.960 | It just takes you out into the water.

00:51:52.560 | You're much better off talking through a recap with investors

00:51:58.120 | than you are doing.

00:51:59.400 | You're just piling delay upon delay.

00:52:03.320 | You're just setting yourself up for more failure.

00:52:05.680 | It reminds me of a related subject, Bill.

00:52:08.400 | I remember at the start of last year,

00:52:10.240 | we were talking about who were the companies that

00:52:12.560 | were going to follow Facebook in terms down the path of layoffs

00:52:16.480 | and getting fit.

00:52:17.640 | And somebody produced an article.

00:52:21.080 | And they said, once we started to see a trickle of companies,

00:52:27.120 | they hired some consultants.

00:52:28.360 | And they said, how many people should we lay off?

00:52:30.400 | And they all magically came up with 13% or 14%.

00:52:34.200 | And we're like, why 13% or 14%?

00:52:36.680 | And they said, well, it's bigger than 10%.

00:52:39.880 | And it's not as hard as 20%.

00:52:42.600 | And I think that this is the other piece of evidence

00:52:46.760 | we have here that people just didn't get the drill.

00:52:50.400 | We always say, do it all up front.

00:52:52.480 | Get it over with.

00:52:53.760 | And get on to rebuilding your business with unit economics

00:52:56.840 | that makes sense.

00:52:57.840 | But unfortunately, if you just look

00:52:59.760 | at the number of layoffs being announced,

00:53:01.560 | I think we now are up to 40 or 50 big companies.

00:53:05.760 | I mean, just recently Wayfair and eBay and go through the list.

00:53:09.240 | But these were all decisions that

00:53:11.080 | could have been made at the beginning of '22.

00:53:13.640 | I'm wondering, at the beginning of '23,

00:53:16.600 | it's shocking to me that we're in the first quarter of '24.

00:53:22.280 | This correction started more than two years ago.

00:53:25.680 | Why are people just now getting to the conclusion

00:53:28.840 | that they should be getting fit?

00:53:30.200 | There's a lot of wishful thinking.

00:53:35.520 | And quite frankly, let's paint it in a very different light.

00:53:41.760 | Startups aren't created by pessimists.

00:53:44.720 | They're created by blind optimists, right?

00:53:47.280 | And so your most likely mindset, if you're

00:53:51.840 | the type of founder who runs at walls,

00:53:54.720 | is that you're going to figure it out, that it'll be better.

00:53:58.560 | I'll make it work.

00:53:59.600 | I'll get back to where we were.

00:54:01.560 | That's how you're programmed to behave.

00:54:03.760 | Although I'll tell you, the story of Elon

00:54:08.240 | is always up and to the right.

00:54:11.360 | People need to celebrate the stories about Elon's survival

00:54:16.920 | and Elon not hiring people and Elon making it

00:54:21.960 | through the near-death experiences of 2008.

00:54:25.800 | What he did at Twitter, laying off 75% of the people--

00:54:28.760 | I saw somebody tweet the other day, Bill.

00:54:31.360 | They said, I guess Twitter isn't going to fall over, right?

00:54:35.880 | I mean, the platform is as vibrant

00:54:38.480 | and the product development cycle

00:54:39.800 | is as vibrant as it's ever been.

00:54:41.200 | Why don't we maybe want to jump onto topic four?

00:54:45.120 | Sure.

00:54:45.840 | So I think through our discussion,

00:54:49.440 | we'll get back to why this relates to VC and tech.

00:54:52.720 | But a lot of people found it quite interesting,

00:54:58.240 | let's say, the new leader of Argentina's speech

00:55:02.240 | at the World Economic Forum, Mille.

00:55:07.760 | And we'll put a link in, but I'm sure most everyone's

00:55:10.160 | seen it by now.

00:55:11.240 | Obviously, very different from all the other talks

00:55:15.320 | that are there.

00:55:16.280 | But why was it meaningful to you?

00:55:18.840 | Well, the first thing I did when I listened to the speech--

00:55:24.800 | and it's something I've thought about a lot--

00:55:27.040 | was Mille's an economist.

00:55:30.520 | When's the last time you heard a politician

00:55:32.800 | give a major address?

00:55:35.240 | And he starts it off by talking about the empirical evidence

00:55:39.240 | and starts quoting data.

00:55:40.880 | So my first question was, I want to see this data.

00:55:45.040 | So I had one of my analysts go back and pull the data, which

00:55:48.760 | was the underlying support data that maybe we can bring up.

00:55:52.600 | And what this data shows-- so this

00:55:54.560 | is the years it takes to double global GDP per capita.

00:56:00.400 | Now, first, maybe we should just start off by saying,

00:56:02.560 | why should we care about this?

00:56:05.040 | So GDP really is the excess, the progress, the prosperity that's

00:56:12.760 | created by a fixed amount of labor and capital

00:56:16.120 | in the world.

00:56:16.640 | We have a fixed number of human beings.

00:56:18.300 | They can only work so many hours a day.

00:56:20.440 | And we have a fixed amount of capital in the world.

00:56:22.680 | And what are all the things and services

00:56:24.640 | that we can produce with that?

00:56:26.120 | And that is the global GDP.

00:56:28.360 | Well, what he noted in this chart

00:56:30.080 | shows that basically from year zero to the year 1800,

00:56:37.360 | we had almost zero GDP globally.

00:56:41.040 | And we had very little prosperity.

00:56:44.840 | We had very little excess.

00:56:46.760 | One might think of it in a very primitive way,

00:56:49.360 | even before the year zero, that we were hunter-gatherers

00:56:52.680 | and we basically lived a subsistence life, OK?

00:56:56.720 | So in that case, he said, takes you 3,500 years

00:57:00.280 | for global GDP to double.

00:57:01.880 | And then something crazy happens around the year 1800.

00:57:05.240 | Remember, Adam Smith's Wealth of Nation

00:57:07.600 | about free market capitalism published in like 1775, 1780.

00:57:13.000 | And so we have free market capitalism,

00:57:15.520 | market democracies introduced into the world

00:57:18.000 | about that point in time.

00:57:19.240 | And just coincidentally, we start

00:57:21.800 | seeing an acceleration in the rate of global GDP growth.

00:57:26.200 | So the period 1820 to 1900, it only

00:57:29.520 | takes 87 years to double GDP.

00:57:32.720 | And then 1900 to 1950, only 60 years.

00:57:36.480 | And by the time you get to the year 2000,

00:57:39.280 | we're doubling global GDP every 20 to 30 years, OK?

00:57:45.400 | Which results in this chart, the next chart, which is--

00:57:51.120 | this is just such a shocking hockey stick chart to look at,

00:57:55.120 | OK?

00:57:56.480 | So this is the increase in global GDP

00:58:02.760 | over the course of the last 2,000 years, right?

00:58:08.200 | You almost have none.

00:58:09.080 | And then it starts going up a hockey stick.

00:58:11.560 | And then we said, let's forecast it forward

00:58:14.040 | for the next 100 years.

00:58:15.600 | And if we forecast it forward and we just

00:58:17.640 | assume that we have the same rate of GDP being added

00:58:24.360 | every year that we have this year,

00:58:25.680 | so we're not assuming any acceleration from AI,

00:58:28.880 | no systems getting better, you just

00:58:32.240 | see the advancements caused by this.

00:58:34.240 | And Millet was, of course, arguing

00:58:36.400 | that this was the result of free markets and competition.

00:58:40.520 | Yep, no doubt.

00:58:42.520 | So Bill, I think if we look at this, one of my questions

00:58:47.400 | for you gets back to this.

00:58:50.360 | Why should we care so much about these empirical facts

00:58:54.880 | and this speech?

00:58:56.160 | And what are the risks to this?

00:58:58.480 | So I would highly encourage listeners to read two books.

00:59:06.200 | The first one is The Rational Optimist by Matt Ridley.

00:59:09.560 | And the second one is How Innovation Works, also

00:59:12.080 | by Matt Ridley.

00:59:13.120 | And they should be read in that order.

00:59:15.560 | One's almost a sequel of the first one.

00:59:18.160 | But in The Rational Optimist, Matt

00:59:22.080 | talks about the underlying mechanisms

00:59:24.360 | that lead to the data you just shared

00:59:26.880 | and what Millet was talking about.

00:59:29.400 | And he sums it up in around two functions.

00:59:34.840 | One is ideas being exchanged.

00:59:37.240 | He calls it ideas having sex.

00:59:40.000 | I'll give you an example.

00:59:42.160 | And we can go back to early agriculture.

00:59:44.280 | If I learn how to plow and put seeds in the ground

00:59:47.600 | and I go to a town to trade, I can tell someone else

00:59:52.000 | how to do it.

00:59:52.560 | I can show them how.

00:59:53.800 | And now they can go and do it on their own.

00:59:56.680 | And that exchange of idea was free.

01:00:01.040 | But it had a magnificent lift to prosperity.

01:00:05.840 | It's ironic because when I wrote down

01:00:07.920 | that I wanted to use that example, the thing that

01:00:10.240 | immediately popped in my mind was in the AI world,

01:00:14.640 | the DeepMind paper about the attention window,

01:00:17.920 | attention is all you can need.

01:00:19.800 | This concept was done inside of Google

01:00:23.120 | as an open source concept, immediately copied

01:00:26.720 | by all the other players.

01:00:28.000 | So open AI doesn't exist today if that's patented and controlled.

01:00:33.240 | It just doesn't, which is another irony

01:00:35.480 | around this open source argument because they've

01:00:37.960 | been affronted massively from a major--

01:00:41.600 | but the idea of being shared-- and this

01:00:43.720 | is also why I'm such a massive open source proponent

01:00:46.920 | because I think it's so relevant to prosperity for the masses.

01:00:50.480 | If ideas can be shared, there's zero cost but infinite lift.

01:00:54.320 | And then the second part of it in Ridley's book

01:00:57.440 | is commerce.

01:00:58.720 | For all the reasons Adam Smith talked about,

01:01:01.000 | it's just a way to allocate resources.

01:01:02.960 | So if you get both of those things humming,

01:01:06.080 | you get massive success.

01:01:07.800 | It's super interesting if you look at the history of China

01:01:11.760 | because they've opened up and closed multiple times.

01:01:14.800 | So they used to be like a third of world GDP.

01:01:17.960 | And then they closed their borders and quit trading.

01:01:20.440 | And they went down way into the low 200s or something.

01:01:25.120 | And then, of course--

01:01:27.400 | by the way, let me share something on my screen if I can.

01:01:30.880 | I'm going to try.

01:01:32.120 | Yeah, right here.

01:01:33.040 | So this is my favorite test of AI bot.

01:01:36.480 | I asked the very simple question,

01:01:38.800 | what single human being has brought the most humans out

01:01:41.600 | of poverty?

01:01:42.600 | And ChatGPT got it right, of course.

01:01:45.480 | Deng Xiaoping, who brought capitalism to China

01:01:49.200 | and brought 500 million people out of poverty.

01:01:54.240 | No other human being, not the most--

01:01:57.960 | not Mother Teresa, no altruist, no socialist,

01:02:02.440 | has come anywhere close to this number that

01:02:05.440 | was done merely by unleashing the human potential that

01:02:10.600 | was latent inside of China by allowing for idea sharing

01:02:15.880 | and commerce to have this massive impact.

01:02:19.040 | And I even think there's an argument in the past,

01:02:23.600 | let's say, three or four years, that by stopping commerce,

01:02:29.480 | or at least restricting it a bit and stopping trade,

01:02:33.000 | has led to at least the shakiness

01:02:35.720 | within the Chinese economy, which

01:02:37.400 | may have led to what happened in Silicon Valley in that meeting.

01:02:43.200 | And like, hey, maybe we need to get this thing back on tracks.

01:02:47.200 | Because of that, you asked what could harm it.

01:02:50.840 | One of my favorite thinkers and professors

01:02:54.760 | is a gentleman named Ricardo Hausman at the Kennedy School.

01:02:57.760 | He's Venezuelan.

01:02:58.800 | And he was giving a presentation once about what

01:03:01.000 | went wrong in Venezuela.

01:03:02.600 | And he said they attacked the invisible hand.

01:03:05.120 | And I love that phrase.

01:03:06.720 | Attacked the invisible hand.

01:03:08.480 | Yeah.

01:03:08.960 | And so these people--

01:03:11.000 | and I'm sure they're well-meaning,

01:03:13.520 | but that view capitalism as the cause of poverty

01:03:20.560 | or negative prosperity.

01:03:21.720 | When the data all says it's the opposite,

01:03:23.720 | it's actually the thing that brings people out

01:03:25.960 | of prosperity, that appears to be the way

01:03:28.480 | you get yourself in trouble.

01:03:29.800 | Yeah.

01:03:30.480 | I think you meant out of poverty and into prosperity.

01:03:33.920 | Out of poverty and into prosperity.

01:03:35.760 | And that seems to me, I think, the reason

01:03:39.120 | that this speech, in particular at Davos, right?

01:03:43.440 | Because there's this phenomenon that's

01:03:45.120 | been criticized there for years, that there's

01:03:47.680 | a move toward perhaps collectivism,

01:03:51.640 | that we need to move away from free markets.

01:03:54.560 | And I think people set up this false dichotomy,

01:03:57.680 | like that people are absolutists on free markets

01:04:00.280 | and they don't think that there should be any regulation,

01:04:02.880 | no government's kind of anarchist view,

01:04:05.400 | or that you should have this very collectivist organization.

01:04:10.000 | It seems to me like the United States has constantly

01:04:13.960 | been in search of a balance between embracing

01:04:18.080 | its DNA of free market capitalism,

01:04:21.080 | while at the same time putting in basic protections

01:04:24.360 | for the least fortunate among us,

01:04:25.880 | for those who can't take care of themselves,

01:04:27.760 | for the people who don't benefit equally in society.

01:04:31.200 | But I think what we've seen is over the last few years,

01:04:33.840 | some concerns among the two of us

01:04:36.960 | included, that the pendulum swung too far,

01:04:40.760 | that there was a lot of anti-capitalist,

01:04:43.120 | anti-free markets, almost anti-democratic fervor,

01:04:47.440 | that somehow it was unfair.

01:04:49.520 | It wasn't equitable to all participants.

01:04:55.120 | Go ahead.

01:04:56.620 | I would just say the social safety

01:04:58.960 | net is funded by the progress that's

01:05:02.360 | created from capitalism and tech and growth and all the things

01:05:06.360 | that we talked about.

01:05:07.280 | And if you eliminate the latter, there

01:05:10.480 | is no money for the former, at least in the long run.

01:05:14.320 | Well, that's exactly what I think Millet's argument is.

01:05:18.880 | And perhaps bringing this bill full circle

01:05:21.760 | back to Silicon Valley, back to tech investing,

01:05:25.160 | people--

01:05:26.760 | I have family members and others,

01:05:28.880 | they say, you've done well.

01:05:30.200 | Why are you still doing this?

01:05:32.520 | And it may sound a little cheesy,

01:05:34.000 | but I say the very nature of what we do, I think,

01:05:37.760 | is a public good.

01:05:39.600 | And I think about it in this way.

01:05:41.320 | Human progress-- I'm not just talking venture capitalists.

01:05:46.280 | In fact, less about venture capitalists.

01:05:48.600 | I'm talking about risk takers.

01:05:50.520 | I'm talking about founders.

01:05:52.160 | I'm talking about the engines of the innovative system.

01:05:56.080 | In fact, the rule of law around bankruptcy

01:06:00.160 | was a pretty novel concept in this country.

01:06:02.840 | But we wanted to make a deal.

01:06:04.400 | We were basically saying, if you're

01:06:06.440 | going to be the risk taker, if you're

01:06:08.240 | going to put your neck on the line

01:06:09.800 | to move our country forward, if it doesn't work out,

01:06:12.880 | we want to give you some protection

01:06:15.400 | so it doesn't ruin the rest of your life.

01:06:17.200 | Because we were trying to create a system of incentives

01:06:20.200 | for people to take that risk.

01:06:22.320 | And so when I look at where we are today, one of the things

01:06:28.720 | that you saw in the empirical facts

01:06:30.920 | is the accelerating rate of capitalism, right?

01:06:35.160 | And it seems to me part of the reason for that acceleration

01:06:38.680 | is that systems compound upon systems.

01:06:42.000 | These are non-linear, right?

01:06:44.800 | And so mobile compounded a lot faster because of the internet.

01:06:49.600 | Internet because of microcomputers.

01:06:54.360 | The cloud because of mobile and the internet.

01:06:57.080 | AI, all those things were preconditions to AI, right?

01:07:02.800 | You just made me think of something that--

01:07:05.720 | if you look at what Deng Xiaoping did in China

01:07:08.520 | and how quickly--

01:07:10.760 | and you've been over there and met with the founders.

01:07:13.920 | I mean, and you've seen comments from Moretz and others.

01:07:18.160 | Just like, at the least, they're equally good.

01:07:21.640 | And arguably, in some ways, in some dimensions, better.

01:07:25.080 | Certainly harder working from the cultural standpoint.

01:07:30.320 | And that happened pretty damn fast, right?

01:07:34.040 | We used to say there was only Silicon Valley,

01:07:36.200 | like there's no other place like it.

01:07:38.800 | And then China very quickly mimicked it, very quickly

01:07:42.000 | in the span of time.

01:07:43.360 | And it makes me wonder if you had,

01:07:46.320 | like, for say, inside of Russia, the type of embrace

01:07:50.360 | of capitalism and free trade that you

01:07:53.320 | did under Deng Xiaoping.

01:07:55.440 | You might see the same damn thing.

01:07:57.120 | People are certainly smart enough.

01:07:59.600 | I think you're spot on.

01:08:01.600 | I was having a conversation on Sunday with Mike Milken.

01:08:04.960 | And he's starting the Center for the American Dream.

01:08:07.440 | And Mike is very, very concerned about move away

01:08:11.320 | from just free market capitalist democracy.

01:08:14.680 | And he said to me something that struck me.

01:08:16.480 | He said, Brad, when I travel the world,

01:08:19.080 | it dawns on me that Vietnam, everybody there

01:08:24.040 | wants to be an entrepreneur.

01:08:25.600 | Everybody there is running hard after capitalism.

01:08:29.400 | In the Middle East, we see the same thing occurring.

01:08:32.040 | And he's like, I just want to make sure

01:08:34.680 | that we continue to underscore in this country

01:08:37.360 | that it is the thing that caused us to be at the top of the heap.

01:08:41.200 | It is what led to the progress.

01:08:43.240 | And we need to protect that.

01:08:44.600 | Because like you said, you can't take it for granted.

01:08:47.280 | And that's certainly-- that's even more true

01:08:49.280 | if you look at the last 20 years.

01:08:52.480 | Like, American industrialism could have arguably

01:08:55.840 | been a post-World War II thing, with all the factories

01:08:59.840 | that were blown out.

01:09:00.840 | But if you look at just the companies that

01:09:04.160 | lead our market caps today, they're all venture backed.

01:09:07.720 | And they're all started within the past 30 or 40 years.

01:09:11.040 | And most of them started by immigrants.

01:09:13.720 | Most of them started by people who didn't start with a lot,

01:09:16.520 | like the amount of economic mobility

01:09:18.520 | that we have in this country.

01:09:21.240 | And so I would say that was a hell of a conversation.

01:09:26.080 | I've enjoyed it, like I always do.

01:09:27.800 | We've been doing these things for a long time.

01:09:30.120 | But why don't we leave it there?

01:09:31.840 | It's a good note to end on.

01:09:33.520 | That was a lot of fun.

01:09:34.920 | And until next time, BG2 is out.

01:09:38.480 | Take care.

BG2 with Bill Gurley & Brad Gerstner | MANG VC Gone Wild, Can You Trust AI Valuations? & More | E01

Chapters