Ep5. Stock Comp, AI Cold War, Valuations for LLM | BG2 with Bill Gurley, Brad Gerstner, & Bob Mylod

00:00:00.000 | We even have a court criticizing Elon for taking an options package where he made no

00:00:06.220 | money unless he saved the company from bankruptcy.

00:00:09.680 | Meanwhile, the CEO of his arch rival, who created no shareholder value over a period

00:00:15.240 | of five years, is making tens of millions of dollars a year in our shoes.

00:00:19.320 | Hey, man, great to see you.

00:00:33.800 | Good to see you.

00:00:34.800 | Good to have you back.

00:00:35.800 | Glad to be here.

00:00:36.800 | Man, I raced up here from NVIDIA GTC.

00:00:38.920 | How was that?

00:00:39.920 | I mean, it's a, you know, I texted Lincoln yesterday when I was down there for the keynote

00:00:46.200 | and he said, you know, I said, how'd your day go at school?

00:00:48.480 | And he said, good.

00:00:49.480 | He's like, hey, Pops, what are you up to?

00:00:51.280 | And I said, you know, I'm just watching the future.

00:00:54.320 | And he said, he asked a great question.

00:00:56.600 | He said, well, why is this any different than all the other computers in the world, right?

00:01:01.040 | And I said, you know, like we've talked about before, we're going from this era where computers

00:01:06.960 | were highly efficient calculators.

00:01:09.840 | And I said, you know, we're not far off from computers being able to reason and think and

00:01:15.240 | being smarter than every human who's ever lived on the planet.

00:01:18.920 | So it's, I mean, it's an exciting time, disruptive time.

00:01:22.520 | We're going to talk about that more later in the show.

00:01:25.680 | But we thought for episode five, we'd mix it up a little bit.

00:01:29.240 | I mean, one of the things you and I talked about when we started the pod was not only

00:01:33.880 | do we want to talk about the topics of the day, but we also wanted to do some deep dives

00:01:37.940 | on topics that you and I have talked about over the years.

00:01:41.240 | You know, I think about the topics that you've covered on Above the Crowd, whether it's network

00:01:44.960 | effects or whether it's a rake too far.

00:01:47.780 | And so, you know, one of those things we're going to talk about today.

00:01:52.040 | After that, we're going to get into some of these Google Apple rumors.

00:01:55.460 | We're going to talk a little bit more about the TikTok fallout, et cetera.

00:01:59.360 | But one of the, we've been pretty passionate and talked a lot, argued a lot over the years

00:02:03.400 | about stock-based compensation.

00:02:06.000 | And it's so important to Silicon Valley because it's all about incentives and alignment.

00:02:11.240 | It's such a special part of what makes this place great, but it also has been a source

00:02:16.440 | of a lot of consternation from Warren Buffett down the line.

00:02:21.060 | So we knew we wanted to talk about this, but in order to talk about it, we had to invite

00:02:25.120 | one of our old friends.

00:02:26.480 | Correct.

00:02:27.480 | Who may be, you know, the best expert on the topic.

00:02:32.100 | The legendary CFO of Priceline, Bob Mylod, who's served on tons of VC boards.

00:02:39.140 | And he's truly got, gets this from both the startup perspective and the public company

00:02:43.520 | perspective.

00:02:44.560 | So Bob, thanks for joining us.

00:02:46.480 | Gentlemen, thanks for having me.

00:02:48.040 | It's great to be here.

00:02:49.920 | Yeah.

00:02:50.920 | If any of my CEOs or CFOs are of the mindset, they want to get this right.

00:02:56.640 | There's no one on the planet I'd rather, you know, introduce them to and have them talk

00:03:00.480 | to them, Bob.

00:03:01.480 | Bob, I think you joined Priceline in 1999.

00:03:05.920 | You know, we must have met around 2000 or 2001.

00:03:09.760 | You were the CFO of the company from when I believe it was a billion dollar company

00:03:14.600 | all the way to it being worth, you know, upwards of a hundred billion, certainly on the board

00:03:21.480 | for that entire period of time, you're still on the board of booking and you know, it's

00:03:26.360 | become extremely valuable in the public market.

00:03:29.120 | There was a little bit of a, there was a big rollercoaster along that way, but that's generally

00:03:33.640 | the and I wasn't CFO the entire time, but yes, that's the general math of it.

00:03:38.600 | Bill, why don't you frame for us, why is this topic so important and importantly, a little

00:03:46.160 | bit about how we got here with stock-based compensation?

00:03:51.120 | Yeah.

00:03:52.120 | And I think everyone in Silicon Valley knows that equity participation is part of the model,

00:03:57.400 | always has been.

00:03:58.400 | Right.

00:04:00.160 | And in general, when it was started, venture capitalists felt strongly that it created

00:04:04.960 | alignment, right?

00:04:06.040 | You wanted managers and shareholders owning the same equity and having the same interest.

00:04:11.200 | And if you create one of these monster outsize, you know, wins, everyone's a winner.

00:04:17.320 | And so it just made sense.

00:04:18.440 | Right.

00:04:19.440 | And I think it's been since certainly since I got here, but way before I got to Silicon

00:04:22.640 | Valley, I think it's been a part of what people do.

00:04:27.360 | In the late 90s, you know, we went through the dot-com boom and bust.

00:04:32.420 | And prior to that time, I think the primary vehicle that was used to incent executives

00:04:40.540 | and employees were stock options, employee stock options.

00:04:44.120 | And they would be priced at the current price of the company.

00:04:48.320 | And you know, while a company's private, you had this, you know, evolving argument of how

00:04:53.780 | you value the options prior to being public, they were typically a discount to the preferred.

00:04:59.520 | Likpref and things could make that argument stand.

00:05:03.960 | Anyway, coming out of 1999, you know, you had a bunch of companies boom bust.

00:05:09.840 | You had a bunch of companies go out of business that people said were never a business.

00:05:14.520 | And you had a lot of skepticism coming in from, you know, people like writers for the

00:05:20.160 | Wall Street Journal, but even ISS, Institutional Shareholder Services.

00:05:24.760 | You may want to tell people quickly what ISS is.

00:05:27.080 | Yeah.

00:05:28.080 | I mean, you know, ISS is basically, you know, a body that helps advise owners, shareholders

00:05:36.680 | of public companies on how to vote on a variety of topics, including compensation, quite controversial.

00:05:44.720 | But they have a voice.

00:05:46.120 | But they have influence.

00:05:47.120 | That matters in this world.

00:05:48.800 | And people also were upset about repricing, backdating.

00:05:52.840 | You know, there were options that were backdated and then re-granting.

00:05:56.160 | So the stocks collapsed and then they just issued new options.

00:05:59.760 | And people didn't like that.

00:06:01.720 | And people begin to argue that options created too much risk.

00:06:06.680 | And so there was this idea, you know, that maybe we needed a different instrument.

00:06:11.920 | Along that same time, a very large investor in a very prominent internet company was worried

00:06:18.420 | about dilution.

00:06:19.420 | He was worried about the dilution from options.

00:06:21.680 | And for those that want to just use a rough framework, you know, when a startup would

00:06:26.000 | get started in early years, you might have as much as, say, 10% dilution from options.

00:06:31.440 | You might hire three executives during that year and it's a big grant.

00:06:35.200 | But then as you move towards being public, people would typically, and these are very

00:06:39.680 | gross rules of thumb, steer towards, you know, a 3% type dilution rate.

00:06:45.820 | This particular investor in this very large public internet company felt that if they

00:06:50.500 | moved to zero-based RSUs from options, and this gets into a technical detail around Black

00:06:57.120 | Scholes, that, you know, there's an argument from Black Scholes that a single RSU is worth

00:07:03.580 | about two and a half or three options.

00:07:05.440 | So if you could switch from these options to an RSU, this investor thought dilution

00:07:10.600 | might drop from, say, 4% to 1.2% or something like that.

00:07:15.640 | And there would be, quote, less dilution.

00:07:18.460 | And so that change happened at this one company, but then these zero-based RSUs became prevalent

00:07:25.240 | everywhere.

00:07:26.240 | Okay.

00:07:27.240 | So let's just pause there for a second because I think there's a potential we've lost half

00:07:31.080 | of the audience because it's a very technical and detailed topic.

00:07:36.160 | So if I just step back for a second, I want to invite you into the conversation, Bob.

00:07:40.480 | The reason, obviously, we care in the case of this large internet company, they're granting

00:07:47.020 | options.

00:07:48.020 | They want employees to be incented and aligned with them.

00:07:50.580 | They may not have a lot of cash in order to pay cash, but you said something critical.

00:07:55.140 | As they get toward the public market, right, there is—these things are obviously expensive

00:08:00.460 | to the potential shareholders.

00:08:01.940 | So folks who are going to buy it, Fidelity, Wellington, others in the public market, they

00:08:07.020 | care and are sensitive to that cost.

00:08:10.040 | Bob, Bill's describing the world that existed about the time, right, that you guys were

00:08:16.620 | going through this at Priceline.

00:08:19.700 | So tell us a little bit about kind of your experience in those early years with options

00:08:25.700 | and then kind of this transition that we struggle with to RSUs.

00:08:32.380 | Yeah, I think Bill does a very good job of describing what the world looked like in venture

00:08:37.820 | capital and private equity for that matter.

00:08:40.240 | That's what I did for 10 years before I joined Priceline.

00:08:44.020 | It was a very simple, straightforward model of investor makes investment at a certain

00:08:50.260 | price per share and sets aside an option pool to be granted to management at the same price.

00:08:57.520 | There wasn't some—there was no exercise of contorting oneself into figuring out how

00:09:03.600 | to make the strike price lower.

00:09:06.120 | It was just simply as investors reap rewards from the outcome of the company growing, employees

00:09:16.840 | share in those rewards.

00:09:19.400 | So by the way, I became kind of battlefield promoted during the dot-com bust.

00:09:24.400 | So I became CFO of Priceline in the fall of 2000 and this was—Brad, as you mentioned,

00:09:32.600 | I joined the company just after it was founded, less than a year or almost exactly on the

00:09:39.880 | one year anniversary of the founding, the company went public.

00:09:45.720 | A few months after that, we had a market value of about $35 billion.

00:09:50.440 | We were a tiny little company but this was the absolute heyday of the dot-com bubble

00:09:57.520 | and then the dot-com bubble burst and essentially our stock went to zero.

00:10:03.000 | If you go back and look at the stock price run of Priceline, we were all the way down

00:10:07.480 | to $5 a share which is a post-reverse split adjusted value.

00:10:13.720 | I became CFO at that point and I wasn't too much worried about stock, stock options.

00:10:21.720 | Our stock had just plummeted so much that all of the stock options that were sitting

00:10:26.280 | in the hands of whether it was myself or every single employee, they were under water.

00:10:31.280 | So the only possible way for us to compensate people at that point was in cash which really

00:10:39.200 | sowed the seeds for some of the ultimate successes.

00:10:42.640 | When you start to think of your business in the form of you're paying people based upon

00:10:47.160 | the cash that you have in your bank and you have no more ability to raise money, it's

00:10:51.880 | very, very clarifying about what you can afford to pay people.

00:10:57.840 | That was the model that we operated for a couple of years, operating Priceline and then

00:11:03.000 | what I will define as one of the great unintended consequences of an accounting change occurred

00:11:11.320 | right around the 2002 to 2004 timeframe and there were two things.

00:11:16.280 | First, there was the adoption of Sarbanes-Oxley where people like me and Jeff Boyd are – this

00:11:21.840 | CEO, we had to sign our names on the financial statements under penalty of criminal prosecution

00:11:29.520 | and there was the adoption of FAS 123 which required that historically, any granting of

00:11:36.940 | stock options to employees did not come with a charge.

00:11:40.600 | The theory historically was it's hard to value the stock options because they really only

00:11:45.980 | have value if a stock goes up and therefore, we won't try and do that even though there

00:11:52.320 | is such a thing as a Black-Scholes model and the things do have value.

00:11:56.400 | Historically, a stock option granted resulted in no charge.

00:12:02.320 | The SEC and the FASB were trying to get to a place where companies had to recognize that

00:12:09.560 | granting of a stock option did represent the transfer of economic value from the employer

00:12:17.960 | to the employee and that needed to be accounted for.

00:12:20.520 | So FAS 123 came along and every public company had to expense the granting of a stock option.

00:12:28.620 | Well, as you can imagine for somebody like me with a hugely volatile – being the CFO

00:12:34.720 | of a company with a hugely volatile stock and I'm a University of Chicago grad and

00:12:41.880 | I studied the Black-Scholes model and had a very clear understanding that there is a

00:12:47.160 | lot of inputs into the calculation of a Black-Scholes calculation that there's some subjectivity

00:12:54.600 | to it and that made CFOs like me very uncomfortable as to signing your name to financial statements

00:13:03.440 | with the stock option charges in it.

00:13:06.760 | Not to mention the fact obviously that when we're sitting down talking to employees

00:13:11.900 | about their compensation and you say to them, "Hey, here is your stock option and by the

00:13:17.120 | way, Fisher Black and Myron Scholes says it's worth – they say it's worth $100,000,"

00:13:25.920 | and you're looking at it and saying, "Well, the stock is trading for $20 a share and you're

00:13:30.080 | telling me my strike price is $20 per share.

00:13:33.440 | So in my mind, it's worthless."

00:13:35.080 | In fact, it is Bob, unless the stock goes up in value.

00:13:40.780 | But the reason it has some intrinsic value according to Black-Scholes is they look at

00:13:45.120 | the volatility and other measures related to the stock and it's effectively a probability

00:13:49.560 | weighted estimation of what this thing is worth today, right?

00:13:54.680 | So from an – go ahead, Bill.

00:13:58.520 | I believe – and Bob could correct me because I didn't go to the University of Chicago.

00:14:01.880 | But I believe Black-Scholes was backed into by looking at – you can create it once synthetically

00:14:07.440 | with the short and a call option and these kind of things and –

00:14:10.520 | Explain that to an employee.

00:14:12.760 | But I understand –

00:14:13.760 | Yeah.

00:14:14.760 | But you said it didn't have value.

00:14:17.080 | At the money call option has value.

00:14:18.800 | No, for sure.

00:14:19.800 | They trade on the public markets today.

00:14:20.800 | That's my only point.

00:14:22.800 | Yeah.

00:14:23.800 | Absolutely.

00:14:24.800 | And they do have value.

00:14:25.800 | Again, which is why for the dawn of the venture capital era, stock options were granted and

00:14:33.200 | they did have value and they created huge value.

00:14:36.480 | But again, only in the event that value was created for shareholders.

00:14:42.160 | As the 2000s moved along, there were more and more companies – and by the way, this

00:14:46.880 | includes Priceline.

00:14:48.600 | We moved away from stock option grants to RSU grants.

00:14:54.120 | And let's just explain again to folks what some of these things are.

00:14:59.360 | So stock option, I get a security.

00:15:02.200 | It says the strike price is $20 that you referenced.

00:15:06.640 | Our stock today is trading at $20.

00:15:09.720 | According to Black-Scholes, that would have some intrinsic value today.

00:15:13.280 | You would have to recognize and expense that over time.

00:15:15.680 | But for me as the employee, I'm just looking at this and saying this is only going to be

00:15:20.040 | valuable to me if this stock goes up over time.

00:15:22.680 | In fact, if I quit the company three years later and it's still at $20 a share, then

00:15:27.760 | I don't walk away with anything.

00:15:29.880 | The move to RSU's restricted stock units means that I'm going to own a security and zero

00:15:37.400 | basis RSU's.

00:15:38.600 | I'm going to own – this is just like get somebody handing you a share of stock and

00:15:44.000 | when I leave three years later, whatever I vested in that share of stock, I can cash

00:15:48.840 | that out and I can walk away with that cash value.

00:15:52.060 | So it works very differently from an incentive and alignment perspective than options were

00:15:59.120 | working.

00:16:00.120 | Sorry to interrupt you, Bob.

00:16:01.120 | Why don't we –

00:16:02.120 | Bob Gourley Yeah.

00:16:03.120 | No, that's exactly right.

00:16:04.120 | So if you give one share of a restricted stock unit at a value of 20, then the recipient

00:16:11.240 | has $20 of value.

00:16:13.400 | Now, if the stock happens to go down by 50 percent, the employee still has $10 of value,

00:16:21.680 | whereas in the case of the stock option, if it goes down one percent, let alone 50, it

00:16:26.560 | has zero value.

00:16:27.560 | Here's where we get into questions of alignment with shareholders because you could theoretically

00:16:36.000 | have a management team that leads a company over a year to a 20 percent decline and pick

00:16:43.240 | up 80 percent of the value of their equity award, whereas shareholders are down 20.

00:16:50.020 | You as a large public owner of shares, you would be down 20 and management gets 80 percent

00:16:54.460 | of their equity value.

00:16:55.460 | Trevor Burrus So Bob, why did you go from options to RSUs?

00:16:58.060 | Bob Gourley Well, again, I will just say that before

00:17:00.460 | the adoption of FAS 123, if you were to grant an RSU, the full cost of that RSU would have

00:17:10.640 | to be expensed in the P&L, in the income statement of a public company.

00:17:14.180 | Trevor Burrus So treat it like cash, same as cash.

00:17:16.580 | Bob Gourley It was treated as cash and – by the way,

00:17:19.740 | there were some companies that did it, but there was – it was very clearly understood

00:17:24.500 | by all people that that was essentially no different than cash.

00:17:30.180 | It would be no different than issuing a share of stock to the public for $20 and then turning

00:17:37.740 | around and taking that $20 and handing it to the employee.

00:17:41.540 | So that was always expensed.

00:17:43.900 | It was just when the stock option also got expensed that issuers looked at that and said,

00:17:50.060 | "OK.

00:17:51.060 | Well, if I now have to expense stock options and I always had to expense RSUs, I might

00:17:56.500 | as well just do RSUs."

00:17:59.980 | It was at this time also that companies got into the practice of adding up all the charges

00:18:07.020 | related to this stock-based compensation and proforming the effects of those out of the

00:18:13.180 | income statement so that they could effectively compare current results to what the results

00:18:18.460 | would have looked like before FAS 123 was adopted.

00:18:21.220 | Trevor Burrus OK.

00:18:22.220 | So hold on a second here.

00:18:24.580 | So you shifted from stock options, which as a shareholder, I like because nobody makes

00:18:33.180 | money unless I make money, to RSUs where people made money even if I lost, right?

00:18:42.060 | And then because that was all now in the P&L, you had to expense that like cash, which Warren

00:18:49.020 | Buffett and many others had been pushing for since '97, '98.

00:18:54.140 | But in order to get around that, everybody started adjusting what they – their earnings,

00:19:01.060 | EBITDA.

00:19:02.060 | They started adjusting the stock-based compensation expense out of EBITDA, OK?

00:19:07.880 | So as though the expense no longer counted.

00:19:12.100 | Did you start that?

00:19:13.100 | Warren Buffett I don't know that I started it.

00:19:17.260 | But we did – so we did do the proforming.

00:19:21.340 | But the reason that we did the proforming wasn't because we didn't view that every

00:19:27.220 | single dollar of RSUs that we were handing out had real, very significant value.

00:19:33.620 | It was more that we were going through this transition period where some of the compensation

00:19:38.000 | we had handed out historically were stock options and now we were moving to a world

00:19:43.900 | of RSUs.

00:19:47.100 | So there was a multiyear period where some of the stock-based compensation was option-based.

00:19:52.260 | Some of the stock-based compensation was RSU-based.

00:19:55.480 | So that would have created even more confusion around how to think about core earnings.

00:20:04.160 | So we did proforma them and obviously almost all companies proforma them.

00:20:12.720 | I think the big difference is that we, as a company, we always viewed one dollar of

00:20:20.640 | an RSU as fully being worth one dollar.

00:20:24.600 | In fact, if you really are bullish on your company, which we were, we viewed that a dollar

00:20:30.160 | of RSUs was actually worth more than a dollar because we certainly had a view hopefully

00:20:34.360 | that if we did our jobs, that the stock was going to go up.

00:20:37.560 | So it was our most valuable form of currency that we used to pay employees and we certainly

00:20:46.240 | thought about that all along the way.

00:20:48.680 | We never – it never occurred to us that it shouldn't count or that it didn't count

00:20:56.560 | and we always looked at it in terms of the dollars associated with employee compensation,

00:21:02.400 | period.

00:21:03.400 | We never looked at it in terms of what dilution – what do these share grants represent as

00:21:09.480 | a percentage of the total outstanding shares or "dilution"?

00:21:15.700 | I assure you, as somebody who sat across the table paying people their compensation, I

00:21:23.800 | wouldn't sit across the table at year end and say, "Congratulations, we've had a

00:21:27.760 | good year.

00:21:29.020 | Your salary is X and your bonus is Y and we're giving you a stock grant of percentage points

00:21:38.240 | of dilution."

00:21:39.880 | Percentage points of dilution do not show up in a W2.

00:21:44.860 | Percentage points of dilution are not what you pay taxes on.

00:21:50.840 | People on the other side of the table are looking at dollars.

00:21:54.580 | So we always looked at dollars and I think where some other companies have strayed is

00:22:01.760 | they've forgotten that those dollars are dollars.

00:22:04.480 | For a tech company, oftentimes stock-based compensation happens to be the single largest

00:22:12.860 | expense component of a tech company and yet somehow, inexplicably, even though it's

00:22:22.200 | the largest single expense component of that company, it is not considered to be even an

00:22:28.240 | expense in some companies and in some boardrooms.

00:22:33.080 | The result of that is that you've seen runaway stock-based compensation on an almost

00:22:40.560 | uninterrupted basis for the better part of 10 to 15 years.

00:22:45.440 | As this happened, exactly what Bob talked about, and as it became way more prevalent,

00:22:51.520 | we were entering that ZERP period and so there was spend at all costs.

00:22:56.320 | There wasn't a lot of attention to profitability or even kind of paying attention on the expense

00:23:02.960 | side because that window happened.

00:23:05.720 | Coming out of the ZERP window, people like yourselves are finally raising their hand,

00:23:10.680 | I'm not blaming this on you, and saying, "Hey, wait a minute.

00:23:14.920 | This isn't right."

00:23:15.920 | Not only is there a non-alignment problem, but these numbers are massive and this drumbeat

00:23:22.080 | has just started to grow.

00:23:25.000 | If we boil it way down, because I think we all have benefited from the magic of stock

00:23:31.400 | options as incentive and alignment for all the value creation that's occurred in Silicon

00:23:36.760 | Valley over the course of the last 20 years, I think what we're trying to point to and

00:23:41.400 | I want to drive toward is kind of what are those best practices if you're a founder,

00:23:45.720 | if you're on the comp committee on a board, if you're an investor.

00:23:51.000 | To me, when you pro forma something out, when you stop treating it like cash, it leads to

00:23:59.480 | a misallocation of resources because you're effectively saying this doesn't exist.

00:24:06.900 | What happens when the cost of something disappears?

00:24:09.600 | You get a lot more of it.

00:24:11.440 | Let me give you a very specific example that I heard the other day.

00:24:16.200 | Take for example the recent competition for AI engineers in places like OpenAI and Google

00:24:22.800 | because venture-backed private companies have to compete in an ecosystem against public

00:24:29.160 | companies.

00:24:31.080 | There are rumors that lead engineers are offered millions of dollars a year.

00:24:36.120 | But instead of cash, for example, Google's offering $500,000 in cash plus a million and

00:24:43.560 | a half dollars in RSUs vesting monthly to these engineers.

00:24:51.580 | Every single month, those RSUs are immediately available for sale.

00:24:56.160 | And then as a shareholder, you have to ask the question, how much of this expense hits

00:25:00.880 | the book at the end of the year?

00:25:04.040 | If you're adjusting out the cost of those RSUs, then you're pretending that the cost

00:25:09.440 | of that engineer is only $500,000.

00:25:15.120 | That to me is what leads to this spiral in inflation.

00:25:19.440 | Now, you saw this in the early 2000s and you said when you had to treat it as cash, you

00:25:26.600 | got a lot more sober about what everybody was paid.

00:25:29.280 | Now the difference in that period of time is the whole world deflated.

00:25:34.800 | Everybody kind of came down together.

00:25:36.120 | But today, we have public companies that are pro-forming this out.

00:25:40.360 | They're still spending a lot on these engineers.

00:25:44.000 | For me, the problem with this is when we – it leads to excess in the public markets which

00:25:50.400 | then misallocates resources away from startups and I think leads to less innovation.

00:25:56.760 | It's – we're now – whatever, we're 20 years since the adoption of FAS 123.

00:26:05.080 | I especially thought when Zerp disappeared that people were going to wake up and say,

00:26:10.480 | you know, we're actually not counting profits correctly here.

00:26:17.000 | These tech companies are far less profitable than we actually think they are.

00:26:21.840 | I've been on – I've been involved in so many companies now as a venture investor

00:26:25.640 | that have gone public.

00:26:26.960 | I've sat on multiple compensation committees of both private and public companies and Brad,

00:26:32.520 | exactly to your point, I can't tell you the number of conversations I've had with

00:26:37.640 | the leadership of a company who – I can think of a specific conversation where CEO

00:26:44.080 | called me and said, "I'm trying to hire this really talented software engineer.

00:26:50.440 | But that person wants $400,000 annual salary which really screws up our compensation here

00:26:58.040 | because that will have that person making $100,000 more per year than anybody else at

00:27:03.800 | our company.

00:27:05.720 | What do you think?

00:27:06.720 | Should we stretch for this person?"

00:27:10.080 | I said, "Well, what's the RSU grant?"

00:27:14.880 | The answer was $10 million.

00:27:18.360 | I'm sitting here like, "Wait, I just got a phone call from somebody worrying about

00:27:22.600 | the $100,000 salary."

00:27:24.600 | Of course, the CEO should worry about the $100,000 salary because if he's playing

00:27:31.160 | the game where that is the number that counts and the other one doesn't, then they're

00:27:36.360 | going to focus on the $100,000.

00:27:38.700 | Of course, I'm looking at it going, "No, you really want to focus on the $10 million."

00:27:43.440 | Absolutely.

00:27:44.440 | It seems to me, just back to why I care so much, obviously, when I think about companies

00:27:53.200 | getting fit, companies dealing with the post-SERP hangover where we had excess hiring, excess

00:28:02.360 | pay, they got out of shape, let's make it very concrete.

00:28:07.280 | If a company has $20 a share in earnings but has 10% dilution every year, then my claim

00:28:17.800 | on that earnings as a shareholder is going down.

00:28:20.520 | Absolutely.

00:28:21.520 | Right?

00:28:22.520 | If they produce $20 in earnings with no dilution, they're not issuing stock at all.

00:28:28.520 | They're paying it all in cash, then I have a claim on that $20 in earnings.

00:28:33.720 | I think the same thing in venture-backed companies.

00:28:35.840 | I think what some people find as shocking is they get to the end of the journey, Bob.

00:28:42.080 | They start off owning – you're a Series A investor, a Series B investor in a company.

00:28:47.160 | You think you own a certain percentage of the company.

00:28:50.920 | But if you're not paying attention to that 10 million share grant, by the end of the

00:28:55.400 | time, like your ownership claim – by the way, founders should care because their ownership

00:29:00.680 | claim as founder is going down.

00:29:03.320 | But unfortunately, they've had to compete in this arms race where the public companies

00:29:09.180 | are all pretending this isn't an expense.

00:29:13.680 | And look, I think competition is real.

00:29:16.000 | I don't want to be seen as this investor type that's saying, "Oh, the owner should

00:29:20.760 | keep everything."

00:29:22.760 | You're out there competing for employees and talent every day, and the price has gone

00:29:27.200 | up, and you are competing with large companies.

00:29:30.200 | Listen to these numbers.

00:29:31.680 | Big companies have an advantage from my point of view.

00:29:34.320 | Microsoft's SBC is $2.8 billion a quarter, so that's over $11 billion a year.

00:29:40.840 | And this morning, we woke up and read about them taking some employees from a hot AI company

00:29:45.300 | in the Valley.

00:29:46.480 | And when you're handing out $11 billion a year in RSUs, you've got a lot to play

00:29:51.920 | with.

00:29:52.920 | Now, that $11.2 billion happens to only be 10% of free cash flow and 35 basis points

00:30:00.920 | of dilution.

00:30:01.920 | It's one-third of one percent because they're worth, what, $3 trillion?

00:30:06.520 | You can't be a big company unless you sort of – in my opinion, unless you establish

00:30:10.720 | much better hygiene when you're the smaller company or even before you're a public company.

00:30:15.480 | And I want to be very clear, like booking holdings – I'm chair of the board of booking

00:30:22.000 | holdings and I'm on the comp committee.

00:30:23.840 | And I want to be clear, like I view that our employees – I know that our employees have

00:30:32.960 | been very, very generously rewarded over a multi-decade period, partly because we live

00:30:41.040 | in a competitive marketplace and we have to make competitive stock grants.

00:30:44.960 | But the biggest reason is because our stock has appreciated over that period so rapidly.

00:30:54.600 | Has gone from $10 a share to $3,400 a share.

00:30:59.880 | That is the appreciation that has driven returns for not only the employees, but driven returns

00:31:07.880 | for all the shareholders of Booking.com.

00:31:09.920 | Correct.

00:31:10.920 | And at least from our perspective, our culture has been, hey, the way to get rich, the way

00:31:17.840 | to make a lot of money is for – to create huge shareholder value.

00:31:24.160 | And the senior management and the board of directors along the way is going to try to

00:31:28.480 | put our thumb on the scale in the form of smartly using our capital to further reward

00:31:36.040 | shareholders.

00:31:37.040 | Well, again, like you're exactly right.

00:31:39.080 | If you look at the top – certainly the NASDAQ 100 companies, but absolutely tech

00:31:44.600 | companies below that market cap, what you will generally see is a year-over-year increase

00:31:52.400 | in share count, literally the expansion of the number of slices of pie because of these

00:32:01.560 | stock granting practices.

00:32:03.800 | You will see some companies that do share buybacks and they will literally state, "I'm

00:32:08.000 | doing the share buyback to offset dilution associated with employee stock-based compensation,"

00:32:16.000 | and thereby doing that, taking a huge cash expense in the form of stock buybacks and

00:32:22.560 | putting it below the line so that when people look at operating cash flow or free cash flow,

00:32:28.680 | it looks like one number.

00:32:30.400 | But yet below the line, there's all this money being spent just to buy back the stock

00:32:34.560 | that employees are selling when their stock vests.

00:32:40.400 | We have taken the view that we want to be very, very fair.

00:32:45.560 | But again, back to the thumb on the scale, we take our excess cash flow or we have tried

00:32:50.780 | to take our excess cash flow and buy our stock back.

00:32:54.120 | So if you were to look at our share count, say 10 years ago, I think we probably had

00:33:00.560 | 50, 52 million shares outstanding.

00:33:04.760 | Today we have 35 million shares outstanding, meaning that during the last 10 years, if

00:33:12.800 | the value of our company, if the enterprise value has doubled, the share price has tripled.

00:33:20.720 | The biggest beneficiaries of that are the people who have gotten the grants that we

00:33:24.880 | gave in the prior periods.

00:33:27.240 | So to me, creating that virtuous cycle where you really are counting that stock as an expense,

00:33:36.120 | that's the way to do it.

00:33:37.120 | A long, long time ago, we adopted the practice of not pro-forming out the cost of stock-based

00:33:44.040 | compensation because it's a very, very real cost.

00:33:48.400 | We do reference cash flow and we do reference free cash flow.

00:33:54.400 | But in our most recent proxy statement, we actually pointed out that the thing that we

00:33:58.400 | really look at is how much of our free cash flow is being spent on stock-based compensation.

00:34:05.040 | Maybe that's a segue, Brett, to what are the solutions to get this under control?

00:34:10.240 | But we have always had booking.

00:34:13.360 | And certainly with all the companies that I'm involved with, I try to get everybody

00:34:16.300 | to focus on the fact that a dollar of stock-based compensation is actually, not only should

00:34:22.280 | it not be ignored, it's the most single valuable expense.

00:34:26.920 | You had a chart and in the chart you compared booking holdings, and we'll show this, the

00:34:32.600 | percentage of operating cash flow that various tech companies spent on stock-based comp.

00:34:40.680 | And so again, I've been on comp committees on public companies.

00:34:45.080 | So if I'm on the comp committee of a public company board, generally what I do at the

00:34:49.520 | start of every year when I'm negotiating the options or the RSU packages for the employees,

00:34:55.440 | we bring in a comp consultant, right?

00:34:59.520 | And this comp consultant makes some recommendations to us based upon the peer groups that look

00:35:05.680 | like this particular company.

00:35:09.040 | Not surprisingly, because everybody has adopted this practice, you know, it's a little bit

00:35:13.440 | of **** in, **** out.

00:35:15.280 | And so if, you know, if everybody's, you know, following a dumb comp practice, that is your

00:35:21.440 | peer group, you know, compare.

00:35:24.420 | But tell us a little bit about this chart and, you know, if I'm, we're on a public company

00:35:30.000 | comp board, and I'm going to, Bill, ask you the same thing as it relates to venture.

00:35:33.800 | But if I'm on that public company comp board, what is my takeaway in terms of what the gold

00:35:38.580 | standard is?

00:35:39.580 | What, is this the way that I should be measuring it?

00:35:41.960 | Or should I be paying attention to Compensia and these other comp consultants that are

00:35:46.120 | coming my way?

00:35:47.120 | I don't know that there's a gold standard.

00:35:48.760 | To me, the gold standard is to just count it.

00:35:51.200 | Again, as I said, like you would never think of not counting expenses that are denominated

00:35:56.040 | in euros if you were, if you had a European subsidiary.

00:36:02.840 | Just the dollars count.

00:36:04.160 | I guess what I'm saying is, is like, it shouldn't be excluded to begin with.

00:36:09.200 | So, one easy way of looking at the practices of various companies is to look at, okay,

00:36:16.240 | what is the dollar value of stock-based compensation expressed as a percentage of OCF or FCF?

00:36:24.080 | That is a blunt way of looking at it.

00:36:25.720 | And by the way, it will shock you when you look at just how much of a company's free

00:36:33.760 | cash flow goes directly back to stock-based compensation.

00:36:39.200 | I think we've done that analysis, Bob.

00:36:42.320 | You've done that analysis.

00:36:43.320 | You know, I think on a chart of maybe 40 or 50 names, we have maybe 10 names that are

00:36:50.720 | under 20 percent of their free cash flow as a percentage of stock-based compensation.

00:36:56.880 | So it is – like the thing that jumps off the page at you is just for tech companies,

00:37:03.600 | which is what we're here talking about, what we care about, it's a massive, if not the

00:37:08.760 | largest single expense for them.

00:37:11.760 | And it's the one thing that way too many of them are not paying attention to.

00:37:15.880 | I think to give some credit where credit is due, I do think some of the adjustments, for

00:37:21.000 | example, that you've seen Meta make both in terms of headcount and now how they're talking

00:37:25.960 | about SBC as a real expense, Uber is talking about SBC as a real expense.

00:37:31.240 | Like there has – partly due to your leadership, partly due to this move to GetFit, people

00:37:37.640 | are paying attention and are talking about it.

00:37:40.080 | I think from our perspective, it's not that you should stop giving – using stock.

00:37:46.360 | But to your point, you should discount the damn thing and treat it as cash and we should

00:37:50.880 | have an honest conversation on an all-in basis what these various companies are earning.

00:37:57.360 | Yeah, and look, let me talk – first of all, I do think shareholders are starting to care.

00:38:02.880 | Like it's – the conversations are happening.

00:38:06.320 | I think most of them oddly happen in one-on-one conversations and as was said earlier by Bob,

00:38:12.840 | the industry seems to kind of accept this adjusted EBITDA thing.

00:38:17.360 | It shows up in the sell-side research.

00:38:20.400 | It's in the earnings releases from the companies.

00:38:22.720 | There are many companies who only talk about that on the earnings call and never talk about

00:38:27.360 | the net income.

00:38:28.360 | So it is – but I think the shareholders are paying attention.

00:38:32.400 | I think they're very unhappy with where things are.

00:38:36.080 | A couple of things Bob hit on.

00:38:37.720 | You know, if you're in a boardroom where people are saying we should buy back the exact

00:38:42.680 | number of shares to offset the SBC, this is going to sound like I'm really – but you're

00:38:49.840 | dealing with financially ignorant people and they just don't understand and those same

00:38:56.040 | people I think are being widely disrespectful to people in your practice, Brad, of owning

00:39:03.760 | shares.

00:39:04.760 | You're not that stupid to where you would say, "Oh, if you buy it back, then it never

00:39:08.400 | happened."

00:39:09.400 | Like –

00:39:10.400 | I've had this argument many times.

00:39:11.880 | Right.

00:39:12.880 | Right.

00:39:13.880 | You're taking my earnings that we could either use to invest in new projects, invest

00:39:18.920 | in new countries –

00:39:20.320 | Do a dividend.

00:39:22.240 | Or just hand – give back by way of dividends and instead you're buying back the shares,

00:39:27.400 | which effectively just acknowledges that this in fact was the cash that we said it was in

00:39:32.200 | the first place.

00:39:33.200 | By the way, when the RSUs invest, they sell – so you're selling and buying at the

00:39:36.600 | same time.

00:39:37.600 | Right.

00:39:38.600 | What's the point of that?

00:39:39.600 | But anyway, I also agree with the comment that was made that I think management teams

00:39:45.520 | treat cash – I thought that conversation about the hiring example you mentioned.

00:39:50.320 | But management gets very concerned about spending a dollar of cash and they're – and many

00:39:55.520 | of these companies, they almost don't care about spending the equity, which is a weird,

00:40:00.120 | bizarre thing because they're actually the same thing in these cases.

00:40:03.400 | Well, it's not weird and bizarre if at the end of the year, remember, a lot of their

00:40:07.840 | bonuses are based on adjusted EBITDA.

00:40:10.920 | Okay?

00:40:11.960 | So again, to Bob's point, if we adjust this out, then it gets back to Charlie Munger,

00:40:17.640 | show me the incentives and I'll show you the behavior.

00:40:21.080 | Let me jump off of that.

00:40:22.680 | So having been in many public comp committees, here's what happens.

00:40:29.080 | There's no one in that committee that's thinking about things from a shareholder alignment

00:40:34.400 | perspective.

00:40:35.400 | They're all thinking about what are the common practices.

00:40:39.120 | So you go hire a comp consultant and the comp consultant is going to tell you what your

00:40:44.280 | peers are doing.

00:40:45.280 | In fact, you create a peer list, you study what the peers are doing and you build your

00:40:49.520 | program based on what the peers are doing.

00:40:51.840 | And if enough people choose the 75th percentile, then you move – what happens there?

00:40:56.560 | You move even higher.

00:40:58.800 | But nowhere does anyone – I don't think the comp consultant or anyone stand up and

00:41:03.440 | say, "Hey, is this a good program for the shareholders?"

00:41:07.240 | And when I've tried to do that in the board meetings, I get pushback.

00:41:10.960 | And there's a great –

00:41:12.360 | Ironic given that there's a fiduciary responsibility to the shareholders, but I agree with you.

00:41:18.480 | No doubt.

00:41:19.480 | And I love this quote.

00:41:20.960 | Warren Buffett said, "I'd rather stick a viper down my shirt than hire a comp consultant."

00:41:25.840 | And my big takeaway from that is the best and the brightest companies have to have the

00:41:31.800 | courage to lead and go in their own direction and create a compensation program that is

00:41:37.800 | unique to them and that is financially rigorous and understands these things.

00:41:42.480 | We've heard a great example of booking.

00:41:44.520 | I think another one is Netflix.

00:41:46.640 | So Netflix today has a remarkably unique compensation structure.

00:41:51.820 | No one does anything like them.

00:41:53.440 | But today, if you're an employee, you get to choose either options or cash.

00:41:59.460 | And there's a slider and you get to pick where you want to be on that slider.

00:42:03.120 | They don't have RSUs, none.

00:42:05.480 | And in a very recent comp committee consultant conversation, I said, "I'd rather use options."

00:42:10.760 | They said, "I wouldn't recommend it.

00:42:12.160 | No one else is doing it."

00:42:13.160 | Yeah, no one else is doing it.

00:42:14.840 | No one else is doing it.

00:42:15.840 | Unbelievable.

00:42:16.840 | And so anyway –

00:42:17.840 | But this gets back to the conversation you and I had last week, right?

00:42:21.280 | We even have a court criticizing Elon for taking an options package where he made no

00:42:27.420 | money unless he saved the company from bankruptcy.

00:42:31.000 | Meanwhile, the CEO of his arch rival, who created no shareholder value over a period

00:42:36.440 | of five years, is making tens of millions of dollars a year in RSUs.

00:42:40.760 | And by the way, I'll say it again, I said it before, I would offer any CEO I work with

00:42:45.460 | the Elon package and I'm certain none of them would take it.

00:42:50.460 | And two, if Delaware doesn't get this right, they really need to understand that shareholder

00:42:55.520 | alignment and executive compensation is a very, very important piece of making these

00:43:01.560 | markets work.

00:43:03.560 | But I'll just – again, Bill, I'll agree with you that – because I've been on both

00:43:08.040 | sides of it obviously when I was on the management side putting together these compensation plans

00:43:13.740 | and presenting them to our compensation committee and to our compensation consultant.

00:43:19.880 | And of course, as a director of many public companies and many comp committees, the dynamic

00:43:28.400 | is kind of almost always the same.

00:43:30.360 | A company establishes its peer group and no board ever wants to think of their management

00:43:37.800 | team as being anything other than above the mean.

00:43:42.560 | And you have 100 percent of all those companies that are thinking of their management teams

00:43:45.960 | as being above the mean, so you have this constantly upward pressure on these – on

00:43:52.680 | the granting practices.

00:43:54.320 | It takes a mindset of like how do we as executives and how do we as leaders of the company want

00:44:01.840 | all of us to make money?

00:44:04.600 | And I think one of the – back to sort of what are solutions.

00:44:09.880 | Consultants, comp committees, venture capitalists, even investors, they're very far too focused

00:44:18.080 | on the term annual dilution percentage.

00:44:24.080 | Hey, this is a reasonable granting practice because it's only 2 percent or it's only

00:44:30.680 | 3 percent or whatever.

00:44:32.880 | But really, all anyone is saying when they say that is they're making a commentary on

00:44:39.440 | the valuation of the company because if I'm a company making a billion dollars of free

00:44:45.600 | cash flow but I have a 100 billion dollar market cap, right, and I've got a 1 percent

00:44:54.360 | dilution rate, well, I theoretically look like I'm managing the company well, it's

00:45:01.720 | 1 percent even though I just wiped out all my free cash flow.

00:45:06.320 | You just have to focus on dollars, ignore dilution percentage and if you want, yeah,

00:45:13.640 | you could look at stock based compensation expressed as a percentage of free cash flow.

00:45:19.320 | But again, Brad, to your point on that slide, it's a shocking number.

00:45:24.560 | The percentage of free cash flow that goes right back in to stock is a huge number and

00:45:31.680 | when you do that over and over and over again, if it's 25, 30, 35 percent every single year

00:45:38.120 | which that's kind of the math of it for many, many years.

00:45:42.560 | The dilution is actually – if you dilute free cash flow by 30 to 35 percent every single

00:45:48.440 | year for 20 years, guess what?

00:45:51.160 | Your dilution is 35 percent.

00:45:52.960 | It's not a half of – it's not 50 basis points.

00:45:57.160 | It's not 1 percent.

00:45:58.160 | It's not 2 percent.

00:45:59.160 | It's literally a third of the company.

00:46:02.520 | I think you make such a good point, Bob.

00:46:05.800 | I go back to this Buffett interview he did on CNBC because this gets pretty esoteric.

00:46:13.220 | You sit in these comp committee meetings.

00:46:15.120 | I see the eyes glaze over.

00:46:16.920 | There's a social dynamic at play.

00:46:19.400 | You sit on the board.

00:46:20.400 | You're in the room.

00:46:21.640 | You often know these people, right?

00:46:23.900 | You want to treat them fairly.

00:46:26.720 | But Buffett said this and we'll show this quote.

00:46:28.680 | He said, "I think the best way to compensate people is with cash.

00:46:31.840 | If you want to give them a bonus for exceptional performance, give them cash.

00:46:35.600 | It's simple.

00:46:36.600 | It's straightforward.

00:46:37.600 | It's clean."

00:46:38.600 | He said, "I don't believe in using stock options as a form of compensation.

00:46:42.120 | I think it's just a way to get around paying out cash.

00:46:44.800 | I think it's a way to sort of camouflage what's really going on."

00:46:48.560 | Now, I disagree with Buffett on the idea of whether or not we should use options.

00:46:54.080 | I actually think there's a tremendous amount of alignment and incentives that are created

00:47:00.400 | through that.

00:47:01.400 | But if you ask me, Bob, what do you think is the best thing if you're sitting on a comp

00:47:05.600 | committee, the conversation that I would force is let's have a discussion about cash, right?

00:47:13.320 | What is the value of what we're giving up, right, in order to achieve our objectives?

00:47:20.400 | Does it make sense?

00:47:22.160 | Because if not, then it obscures bad business models.

00:47:27.820 | It obscures sloppy performance, right?

00:47:30.560 | To me, it comes back to this question of getting fit.

00:47:34.840 | You want to know, "Am I really fit?"

00:47:38.200 | If you're going to work out and you're not measuring your weight or your body fat or

00:47:43.480 | anything else, you're kidding yourself.

00:47:48.360 | To me, the measure of whether or not a company is really performing is the cash equivalent

00:47:54.640 | focus.

00:47:56.200 | Bill, on this, the thing I wanted to ask you is what has been the rule of thumb in Silicon

00:48:05.120 | Valley?

00:48:06.120 | If you're a founder, you used to invest almost exclusively in Series A companies, first institutional

00:48:12.440 | money in.

00:48:14.240 | I don't know.

00:48:15.240 | The rule of thumb is there would be a 20% or a 30% stock option pool in addition to

00:48:20.840 | what the founders owned in the business.

00:48:23.360 | That would go down over time, I think.

00:48:26.640 | As you march to the public market, you would have to focus on, again, treating this as

00:48:31.760 | cash and getting profitable.

00:48:35.360 | What are the rules of thumb today in Series B, Series C, Series D?

00:48:40.120 | Where do you think you're seeing some of the sloppiness?

00:48:43.680 | Well, look, the thing I'd start with in answering that question is mostly startups never get

00:48:49.440 | to this place that we're talking about.

00:48:54.140 | A lot of the commentary in Silicon Valley is always about the 10 or 20 or 50 companies

00:49:00.360 | at the top, and the vast majority of these companies never get to that place.

00:49:06.200 | The general rules of thumb, as I started with, are like 10% headed down to three.

00:49:11.360 | It is done as a dilution perspective.

00:49:16.040 | Profitability first, I think, is if you have a company, especially one that burns a lot

00:49:20.240 | of cash, and if new fundraising and capital is expensive, I think one of the reasons people

00:49:26.520 | start with this adjusted EBITDA number is it does represent a point in time at which

00:49:32.360 | you no longer have to raise capital from the markets.

00:49:37.520 | You could argue over whether that matters or not if you're issuing stock, but it does.

00:49:41.500 | It may matter, especially if capital is hard to come by.

00:49:45.840 | Then I would just say some of the same things we've been saying.

00:49:48.640 | Don't fool yourself.

00:49:50.600 | Don't assume cash is cheaper or more expensive than equity.

00:49:54.720 | They're really the same thing, and eventually it's going to matter.

00:49:57.880 | I do think Wall Street figures it out.

00:50:00.280 | I don't think there are any secrets.

00:50:03.280 | I think alignment is really, really important.

00:50:05.760 | I encourage companies to, and I do this when I'm recruiting, to look at scenarios of the

00:50:12.960 | stock being up 20%, 40%, 80%, 100%, and look at the payouts that people are going to make

00:50:19.280 | on that, and see if you have alignment.

00:50:21.880 | I think if you're only doing zero basis RSUs, you've got a horrible alignment problem.

00:50:27.940 | Don't be mad.

00:50:29.200 | When you think you're EBITDA profitable, and your stock doesn't have the same multiple

00:50:33.360 | as the other company, and I guarantee if you look under the hood, they've got a lower SBC

00:50:39.600 | expense, and they're not pissing away their free cash flow.

00:50:44.920 | I don't think there are any secrets.

00:50:46.400 | I think competition is a real issue.

00:50:48.320 | One of the things I wanted to go to with Bob, I saw this really interesting presentation

00:50:55.360 | from Gary Steele, who turned around Splunk in a similar way to what happened at Meta.

00:51:01.920 | Gary crushed it at proof point, and a huge value creator.

00:51:06.640 | He said, in a very short window, he said he took stock participation from 87% to 15%.

00:51:15.920 | I saw so many CEOs in the audience, their face just went into shock.

00:51:20.800 | It has become almost ritual in Silicon Valley to assume you have to give everybody stock,

00:51:27.760 | including the chef.

00:51:30.160 | It literally is part of, "Oh, we're cool like everybody else."

00:51:33.960 | I think that's being re-examined.

00:51:36.600 | The other thing he said was, they moved a lot of employees out of Silicon Valley.

00:51:42.760 | That's an interesting thing, and I'm hearing that from more and more CEOs.

00:51:47.200 | They say, "Silicon Valley is a wonderful place to start a company.

00:51:50.040 | It's a horrible place to scale a company."

00:51:53.000 | My goal isn't to come on this podcast and say, "Hey, everybody is getting overpaid."

00:51:57.960 | Far from it.

00:51:58.960 | It's more just how I think people – we need to have a better model by which people

00:52:05.120 | get paid.

00:52:06.120 | At least at Priceline Booking, we always had the mantra of a dollar has to be worth a dollar,

00:52:16.440 | whether it's in cash or comp.

00:52:20.080 | If we can't convince somebody that a dollar worth of stock is worth a dollar, then you

00:52:26.120 | actually shouldn't grant it.

00:52:27.320 | You should just give a dollar of cash because they will actually value it greater.

00:52:33.840 | I think over time, because we've been so careful about this and because we have been

00:52:39.360 | able to create shareholder value through many ways, mostly through the incredible work of

00:52:47.200 | the employee base, but also, as I said, through smart capital management that has allowed

00:52:52.560 | us to reduce our share count, that's an incredible benefit that employees have gotten

00:52:58.760 | over the past many years.

00:53:01.440 | When the share count is getting reduced, theoretically, the stock price is supposed to go up.

00:53:06.560 | That benefits not only shareholders, but the RSU recipient.

00:53:10.320 | Ryan: Hey, Bob, do you have any sense how many employee millionaires that Booking Priceline

00:53:16.480 | has made over the last 20 years?

00:53:19.080 | Bob: No, I don't.

00:53:21.320 | But it's – it is certainly –

00:53:23.680 | Ryan: Extraordinary number.

00:53:25.280 | Bob: Yeah.

00:53:26.280 | Ryan: Extraordinary number.

00:53:28.360 | Maybe this is a good place to wrap.

00:53:32.240 | I think your leadership on this issue, right, with this fundamental thing – this is not

00:53:38.120 | about undermining founders' and employees' opportunity to make an extraordinary return,

00:53:47.080 | make millions of dollars in these startups.

00:53:50.160 | In fact, it's about protecting that.

00:53:52.640 | And it starts in the public markets.

00:53:55.400 | Because if the public markets all pretend that this doesn't exist, then they're going

00:53:59.800 | to pay excessive amounts, which deprives startups and the venture capital community of being

00:54:05.400 | able to hire those employees and compete.

00:54:07.480 | That's the example you just gave us just now of Microsoft.

00:54:13.080 | And so I think that accountability, treating stock as cash, as public shareholders – and

00:54:18.800 | the good news is, like I said, there's a lot of evidence that this is starting to be

00:54:26.320 | a real matter of conversation on Twitter, on CNBC, in annual letters, folks like Bob

00:54:32.760 | leading the way.

00:54:34.360 | And then I think we need to begin telling the stories again in Silicon Valley.

00:54:40.240 | The reason the Michael Dells of the world own as much of their companies at the end

00:54:44.680 | of the day as they do is because from day one, they were concerned about giving shares

00:54:49.820 | away where they didn't need to give shares away.

00:54:52.560 | And so that discipline cascades down.

00:54:55.440 | I think all companies are better when their DNA from the get-go is just honest.

00:55:02.080 | They treat all of these things as expenses.

00:55:04.880 | It forces you to build a better business model.

00:55:07.920 | We were lucky to be partnered with Bob and investors from the early days.

00:55:12.320 | And they were forced to really reboot and think about the business model that they were

00:55:18.000 | going to focus on.

00:55:19.000 | If they kept pretending that this didn't matter, I think it would have been a very

00:55:23.520 | different outcome for that business.

00:55:25.280 | I think you would have produced far fewer millionaires than the number that you've

00:55:29.120 | produced.

00:55:30.120 | And again, in Silicon Valley today, I do see this also cascading down into the companies

00:55:36.120 | that we're investing in.

00:55:37.120 | Certainly the companies that were on the comp committees of the boardrooms that we're in,

00:55:42.000 | this is a matter of conversation.

00:55:45.040 | I would say during the heyday of Zerp, I was seeing some really crazy stuff in terms of

00:55:52.720 | annual dilution.

00:55:53.880 | And now it's a tough conversation at each of those boards because I think a lot of venture

00:55:57.440 | capital firms and founders woke up and said, "Wow, I just gave away a lot of the business.

00:56:03.600 | I thought that was okay because the valuations had gone from $500 million to $5 billion.

00:56:08.840 | But now that I know I'm really only worth a billion dollars, the consequence of that

00:56:14.080 | dilution was way more dramatic than I thought."

00:56:16.480 | And I would just close by once again going back to this notion of independent leadership.

00:56:22.280 | The only companies I see kind of getting out of this box are ones where you have people

00:56:28.160 | that are thinking from a first principles standpoint, like Bob did at booking.

00:56:33.520 | In the Netflix example I gave, they're only diluting 1.5% a year, which is pretty left

00:56:40.680 | rail kind of thing if you look at them on one of these charts we're going to post.

00:56:45.480 | And so it takes someone that really understands this stuff at a fundamental level to be willing

00:56:53.240 | to step outside the box because once again, the comp consultants, maybe even ISS are going

00:56:58.080 | to tell you to do what everybody else is doing.

00:57:00.840 | And I don't think that's going to get you to the finish line.

00:57:03.720 | Right.

00:57:04.720 | Thanks for being with us, Bob.

00:57:07.240 | Well, Brad, I thought maybe we'd hit two or three more topics today.

00:57:13.000 | And wrapping it up, I just thank you once again to Bob.

00:57:15.920 | And I wanted everyone to know that we're going to put a bunch of resources in the show notes.

00:57:20.240 | We've taken snapshots of 50, 100 different companies, looked at SBC as a percentage of

00:57:26.840 | free cash flow, which Brad and his team believe is the most important, but also revenue and

00:57:31.880 | market cap.

00:57:32.880 | I think those are all useful ways to kind of get a sense of how companies are doing

00:57:37.240 | on this issue.

00:57:38.360 | And then lastly, my good friend, Mike Mobison, who many people in the internet finance world

00:57:44.580 | know about, he put out a really great piece on stock-based comp and reading it put a lot

00:57:50.640 | of thoughts in my head as we went through this.

00:57:53.000 | So I'll put a pointer out there to that as well.

00:57:55.760 | OK, so there's a lot of new stuff happening, Brad, since we talked last.

00:58:02.280 | One of them that's kind of interesting, and I think I saw a photo on the internet of Sundar

00:58:08.960 | and Tim Cook at a restaurant that may have kicked this off.

00:58:11.920 | But there's a rumor that Apple's talking to Google about leveraging Gemini, I believe.

00:58:17.960 | And tell me what you think of this.

00:58:19.800 | Yeah, I mean, I think the news broke yesterday.

00:58:22.840 | Both our stocks were up a bunch on it.

00:58:24.520 | This idea that-- I think the implication of the rumor or the tweet was that Apple was

00:58:32.040 | somehow maybe outsourcing or partnering with Google on Gemini.

00:58:36.920 | And I tweeted in response to this that I put the probability of them, quote unquote, outsourcing

00:58:44.800 | AI to Gemini at next to zero.

00:58:47.640 | We talked about this on the pod a few weeks ago.

00:58:50.680 | And I think that we-- I go back to that analysis, which is this.

00:58:57.480 | Apple is in this enviable position to be able to build the killer personal assistant that

00:59:05.800 | has memory, that has persistent, that can do transactional thing for us, whether it's

00:59:10.660 | make a phone call, send a text, book a restaurant reservation, book a hotel, et cetera.

00:59:16.480 | So we suggested that they're going to work on their own native small language model that

00:59:21.600 | really doesn't need to be Einstein.

00:59:23.880 | It just needs to serve me really well as my personal assistant.

00:59:27.500 | I think what's likely being commingled here is I presuppose that Apple is also talking

00:59:34.800 | to Google, to Meta, to OpenAI, and others about bringing those models in as part of

00:59:41.920 | a generative search application that will sit on the deck as well.

00:59:45.820 | So think of that maybe similar to perplexity.

00:59:47.960 | And then, of course, they have their search deal with Google.

00:59:52.960 | So to me, I don't think anything really changes in my view about Apple.

00:59:58.800 | I think they continue to be in this really interesting position.

01:00:02.160 | I'm sure they're going to talk about it in June at their developers conference.

01:00:07.840 | I think it's unlikely that you're going to see some transformative thing get shipped

01:00:12.880 | this fall.

01:00:13.880 | I think you may very well see a generative search application that gets shipped this

01:00:19.420 | fall that may, in fact, have multiple of these answer engines embedded inside it, including

01:00:25.960 | Gemini.

01:00:26.960 | But the final thing I would just say on this is I've been, over the course of the last

01:00:33.580 | couple months, pretty openly critical in my critique that Google needs to face up to this

01:00:40.940 | innovators dilemma, that needs to make some real changes in terms of how they're going

01:00:46.380 | to market.

01:00:48.100 | And you and I have those debates all the time.

01:00:52.460 | And I think there is a group of people on Twitter and otherwise who assume that I must

01:00:59.060 | somehow-- that must mean that I'm short Google.

01:01:03.700 | And we've said many times, this show is not investment advice.

01:01:08.300 | Don't presuppose you know where we stand on a particular stock.

01:01:12.220 | I'm often asking tough questions of stocks that we own.

01:01:15.660 | I mean, go back to Meta.

01:01:18.180 | And in fact, we own Google.

01:01:19.740 | We weren't shorted this week when this news came out.

01:01:24.700 | And so I think it's an important reminder, much like Stan Druckenmiller.

01:01:30.460 | At this moment in time, when you're in the middle of these phase shifts, mental flexibility

01:01:34.660 | is incredibly important.

01:01:37.420 | And we are going to follow the facts, and we may change our opinions on a particular

01:01:43.100 | stock.

01:01:44.100 | I think ultimately for Google, the challenge remains the same.

01:01:48.100 | The challenge for Apple remains the same.

01:01:50.020 | You should imagine that everybody is talking to everybody right now as they try to sort

01:01:56.180 | out where they're going to compete.

01:01:58.940 | But I don't know.

01:01:59.940 | What was your reaction to it, Bill?

01:02:01.540 | Well, it's funny, because I think you can have two reactions that are exactly opposed

01:02:06.820 | to one another.

01:02:07.820 | Right.

01:02:08.820 | So you could view it-- you said both stocks were up.

01:02:10.980 | Yes.

01:02:11.980 | So that's the interpretation, that this is positive for Apple, that they're maybe getting

01:02:16.780 | their AI game ready to go.

01:02:19.740 | You could view it as weakness on Apple, that in order to get AI right, they need to go

01:02:24.600 | talk to Google, and they can't do it on their own.

01:02:27.420 | As we have said before, I think the handset is a remarkable asset in the long-term AI

01:02:33.140 | race.

01:02:34.140 | I mean, you know, one issue that could come out of this is regulatory.

01:02:38.020 | I mean, if you combine the two of them, that's nearly every handset.

01:02:41.340 | So if the same AI code were running on both, I don't think that'd be great.

01:02:46.460 | I do believe that the handset is so valuable here.

01:02:51.620 | If I were Apple and wanted to prove to the world that they had their AI game together,

01:02:56.620 | and I know this is a bit of a broken record with people, but fix Siri.

01:03:00.740 | Like get Siri involved and release-- But that is the small language model that

01:03:03.860 | I'm talking about, right?

01:03:04.920 | That is going to be the new back end for Siri.

01:03:07.660 | Like there's no way they're outsourcing that to Google.

01:03:09.580 | And by the way, on that topic, I think voice AI may be more interesting than, say, Sora

01:03:16.620 | or the video stuff that OpenAI released.

01:03:19.260 | I think more and more people are realizing, because LLMs are language models, that people

01:03:25.580 | would love to just talk to them.

01:03:27.580 | You brought up the point that Meta may be more focused on the sunglasses with an earpiece,

01:03:33.620 | right?

01:03:34.620 | Right.

01:03:35.620 | And you want to be able to talk to that.

01:03:36.620 | And so for me, the irony of this is, I think voice AI may be a new frontier that you may

01:03:44.540 | see a bunch of these people race after.

01:03:47.340 | And getting that to work appropriately so that you feel comfortable-- I always like

01:03:52.660 | to reference the movie Her-- just talking to this thing and not having any errors, I

01:03:57.260 | think that's super interesting.

01:03:58.620 | So we'll see how this all plays out.

01:04:01.500 | Let's move on to the next topic.

01:04:02.820 | So there's been a lot of conversation in the past few weeks about TikTok and ByteDance.

01:04:10.580 | And obviously, it very quickly moved to the center of almost all discussion, because it

01:04:19.300 | involved the House and the Senate.

01:04:21.220 | And it's on the Sunday morning talk shows.

01:04:23.580 | It's in every newspaper.

01:04:25.700 | And it's been discussed quite a bit.

01:04:27.380 | But I'd love to hear your thoughts, Brad, on this situation.

01:04:30.700 | Yeah, well, I've said publicly before, we were early shareholders in ByteDance.

01:04:35.660 | I think we've been involved almost a decade in the company.

01:04:39.660 | So take everything I share with that grain of salt.

01:04:43.540 | First, I think everybody knows ByteDance is a super large company.

01:04:49.580 | There's lots of numbers that are rumored out there.

01:04:51.860 | But the reporting suggests that over 90% of its revenues are outside the United States,

01:04:57.020 | and over 100% of its profits are outside the United States.

01:05:00.900 | What's that mean?

01:05:01.900 | Well, I think I've seen other numbers out there that they're doing somewhere between

01:05:04.940 | $8 billion and $10 billion in TikTok US revenue, and that they're losing money on TikTok US.

01:05:10.620 | So I think that it's important just to understand where that sits in the context of the entire

01:05:15.200 | business.

01:05:16.200 | Secondly, with respect to myself, I'm a dad, and certainly an American first.

01:05:21.820 | And my kids use TikTok.

01:05:23.700 | They create on TikTok, et cetera.

01:05:25.540 | But I read all the same things everybody else reads.

01:05:29.380 | And I think it's impossible for me to prove or disprove.

01:05:33.300 | And I think there's legitimate debate as to what's going on.

01:05:37.500 | I think there's also lots of legitimate debate about the House bill itself and whether or

01:05:43.060 | not that's narrowly tailored enough, whether the language needs to be improved, et cetera.

01:05:48.700 | But clearly, it passed the House by a wide majority.

01:05:53.660 | I think there's a real question as to whether or not that will pass the Senate.

01:05:57.460 | But I think that misses the point a little bit.

01:06:00.460 | Set all of that aside, what's become clear to me over the course of the past 12 months,

01:06:07.220 | and certainly over the course of the past 30 days, is irrespective of the merits of

01:06:12.380 | the matter, right?

01:06:13.460 | This is now in the sovereign domain, right?

01:06:17.220 | AI, over the course of the last 12 months, has gone from an interesting technology to

01:06:24.020 | literally the leading edge of where we're battling with global sovereigns over trade,

01:06:32.260 | national security, national economic interest.

01:06:35.220 | So given that, I think it's almost inconceivable for me now to see how TikTok US has a path

01:06:41.840 | forward under kind of the status quo.

01:06:45.240 | So where does that leave us?

01:06:47.140 | That leaves us, on the one hand, with the proposal that I think President Trump had

01:06:53.260 | made toward the end of his term, and which I hear bantered about by lots of senators

01:06:58.280 | and House members, which is some sort of spin to US shareholders, where maybe ByteDance

01:07:03.640 | swaps some of its ownership or sells some of its ownership to somebody like a Microsoft

01:07:08.020 | or an Oracle, and the app can continue running in the United States, OK?

01:07:12.420 | I think there are a lot of US regulators that would find that an acceptable outcome.

01:07:18.060 | But I'm increasingly of the mind that the Chinese regulators would not find that an

01:07:22.100 | acceptable outcome, OK?

01:07:24.660 | And the US regulators won't allow the app to continue as is.

01:07:28.940 | So I think there's a real chance that when we're looking at this 12 to 18 months from

01:07:33.340 | now, that it's an absolute no-go, that they can't find a middle ground, and that the app

01:07:38.300 | is shut down, which, of course, we saw President Trump tweet, like, the implications of what

01:07:44.380 | that makes Meta more valuable or more powerful, as he suggests, or perhaps these other US

01:07:51.900 | companies.

01:07:52.900 | But the thing, I guess, that interests me the most is what is—you know, that's just

01:07:56.340 | the first step, Bill, right?

01:07:58.500 | Like that's the first move on the AI chessboard between these two sovereigns, if you will.

01:08:04.940 | It's hard for me to see a world where TikTok US is shut down or banned, where there's not

01:08:11.020 | some response by China.

01:08:13.220 | And I think we just had this conversation about Apple putting Gemini AI on the Apple

01:08:19.260 | phone.

01:08:20.260 | I'm certain, you know, Apple China is a monstrous business, is a huge business in China.

01:08:26.020 | And I imagine that now that AI is going to be embedded in these phones, that there's

01:08:30.840 | probably a little bit different conversation that's being had in China about, you know,

01:08:36.340 | about that there as well.

01:08:37.340 | So I think we're seeing just the early innings, unfortunately, right?

01:08:42.780 | I tend to think that global trade is a good thing for human prosperity.

01:08:47.780 | I think it's driven a lot of our global GDP and prosperity over the course of the last

01:08:51.980 | 20 years.

01:08:52.980 | So I don't celebrate that we're entering what appears to be a global AI Cold War on

01:08:59.020 | the economic front.

01:09:01.820 | And I'm certainly not critical of the fact that US regulators have these concerns.

01:09:07.620 | But unfortunately, I don't see much of a way out here.

01:09:09.660 | I hope I'm wrong.

01:09:10.780 | I hope that, you know, when President Xi was here a few months ago, he said, I don't want

01:09:16.300 | a Cold War, I don't want a hot war.

01:09:17.940 | I want a partnership.

01:09:19.060 | But that's not what it looks like from my vantage point today.

01:09:22.100 | Yeah.

01:09:23.100 | I'll give you a couple of my thoughts.

01:09:24.860 | First of all, I thought our friends over at All In did a really great job of tearing this

01:09:29.180 | apart.

01:09:30.180 | And even where they disagreed, I think every single one of them agreed that the data probably

01:09:35.660 | should not be going to the CCP.

01:09:37.260 | Like I haven't seen anyone said, oh, I'm completely cool with data going to the CCP.

01:09:42.260 | So if everyone's in agreement on that, and I do understand the arguments that maybe the

01:09:47.220 | House bill is too flexible or too vague or too broad, and you could tighten it up and

01:09:52.380 | achieve the same things.

01:09:54.740 | As I've told you, I feel like that the company, and the company I'll say is either US TikTok

01:10:02.620 | or ByteDance, could have done a better job of proving data independence.

01:10:07.340 | And I think they could have hired an auditor, you know, KPMG or Pricewaterhouse or whatever,

01:10:13.180 | rather than simply saying, oh, it's in Oracle in Texas.

01:10:17.340 | To me, there's not much certainty in that comment.

01:10:20.340 | I would also say relative to your Apple comment, I do believe Apple has to keep all their data

01:10:25.140 | locally over there and prove it to the CCP.

01:10:27.660 | So I think-

01:10:28.660 | But that may not be enough.

01:10:29.660 | No, it may not.

01:10:30.660 | In this new context.

01:10:31.660 | But I agree with you.

01:10:32.660 | But I'm just saying from a prosody standpoint, I think they're already doing what is being

01:10:36.540 | asked here.

01:10:38.380 | And as we all know, and we've said, the social networks here aren't allowed to operate there.

01:10:42.740 | So you could have a, I think you have a pro-China mindset and just still hold the line and say

01:10:49.060 | fairness is fairness.

01:10:51.300 | So we will see how it plays out.

01:10:52.500 | I'm hopeful that there's a way to get it together.

01:10:56.100 | But who knows?

01:10:57.340 | We will see.

01:10:59.820 | Anything else on that?

01:11:00.820 | No, I mean, I think that it's, once you start going down this slippery slope, right, you

01:11:11.300 | know, we have a lot of US companies that have a big footprint in China.

01:11:16.540 | And so-

01:11:17.540 | But you're already down the slippery slope.

01:11:19.100 | Well, I mean, we haven't banned it yet.

01:11:22.260 | So I think if and when this effectively, you know, if this were to get banned in the United

01:11:26.780 | States, again, I would like to think that that would be, you know, where this ends.

01:11:31.100 | All I'm suggesting is that ChatGPT only came out 18 months ago.

01:11:36.780 | And you can't use it in China.

01:11:40.740 | The wheels of international relations, I think, turn slowly.

01:11:45.620 | And when we look at, it's first started with silicon, right, with restraints on China's

01:11:52.220 | ability to get Nvidia chips.

01:11:54.220 | And again, I don't disagree with, you know, our ability to keep our best stuff to ourselves.

01:12:00.020 | But it's hard to look at the facts on the field and say that this is not, we're not

01:12:05.220 | entering a new phase of global international trade relations, you know, vis-a-vis China.

01:12:11.540 | Now, one might say that we're just, you know, creating a more fair and balanced playing

01:12:16.020 | field, right, relative to the restrictions that have been imposed on U.S. companies.

01:12:20.820 | But I think it would be naive of us to think that there's not going to be, you know, counter

01:12:24.980 | responses out of China.

01:12:26.180 | I hope not, but-

01:12:27.180 | I take the other side of it simply because we're not even getting to the point of equality

01:12:31.740 | here.

01:12:32.740 | So anyway, we'll leave it at that.

01:12:35.420 | And then this morning, just this morning, we rarely have news that broke just this morning,

01:12:39.780 | a very unusual thing has played out with Microsoft and inflection.

01:12:44.460 | Why don't you tell people what went down?

01:12:46.780 | Well, inflection, you know, is, you know, just as a reminder, was one of the early LLMs

01:12:58.060 | to be trained on H100s.

01:13:00.620 | They have a consumer product called PI, which stands for personal intelligence, founded

01:13:04.860 | by Mustafa Suleyman, one of the co-founders of DeepMind, you know, and a friend.

01:13:12.300 | And you know, what they did was, you know, as one of the-

01:13:16.180 | And by the way, just to put in, I think there's another way to frame this.

01:13:19.420 | There were three LLM companies in Silicon Valley that raised over a billion dollars,

01:13:26.260 | although some of that was with credits, but there were three, Anthropic, OpenAI, and this

01:13:31.580 | one.

01:13:32.580 | Correct.

01:13:33.580 | And, you know, this is a large startup, but what they were going after was really think

01:13:38.940 | the chat GPT consumer market with PI.

01:13:42.660 | And, you know, today we heard news that Microsoft has effectively hired the employees out of

01:13:51.180 | inflection.

01:13:52.180 | We don't know all the details, you know, of the deal here.

01:13:57.220 | But you know, we looked at investing inflection, as many others did, and, you know, one of

01:14:02.860 | the reasons that strategic investors have displaced a lot of venture capital investors

01:14:08.580 | in the LLM game, and we talked about this, I think, on, you know, episode two, is the

01:14:13.900 | quantum of capital it takes to compete in the LLM space is very high.

01:14:20.220 | And on the other side, understanding the durability of the business model, like in this case,

01:14:25.260 | getting consumers to pay you, you know, for PI is far from clear, particularly in a world

01:14:31.720 | where you got to go compete against Apple, you got to compete against Meta and others

01:14:37.020 | who are going to build a personal assistant.

01:14:39.500 | And so, I think it was a very difficult venture bet to underwrite.

01:14:42.860 | I think it was a logical bet for folks like Microsoft, you know, to underwrite because,

01:14:49.780 | you know, they want to be in the consumer answer engine space, the consumer chatbot

01:14:55.260 | space, if you will.

01:14:56.260 | I think the most interesting thing out of this, to me, and if you read Satya's tweet

01:15:01.400 | that he sent out this afternoon, is he said he's excited to have them build consumer AI,

01:15:07.220 | which, you know, he said, like the copilot, that is loved and benefits people from around

01:15:13.660 | the world.

01:15:14.660 | So, just step back from a second here.

01:15:16.100 | He owns 49% of OpenAI, ChatGPT is charging really hard to be, you know, kind of my personal

01:15:23.740 | copilot, if you will.

01:15:25.820 | And now they have another, you know, another team in the hunt.

01:15:29.740 | So, you know, I think this is smart by Satya and Microsoft.

01:15:34.460 | Some might view it as hedging their bets a little bit against ChatGPT, but I think it's

01:15:40.460 | great for the consumer.

01:15:41.980 | They're going to be a lot of at-scale players that are vying to be your personal assistant,

01:15:46.700 | that are vying to answer your questions.

01:15:48.860 | We just talked about Apple, we talked about Meta AI, Google's certainly going to be in

01:15:53.580 | that hunt.

01:15:54.580 | And now it looks like Microsoft is doubling down on the consumer copilot, which will bring

01:16:00.020 | even more competition to the table.

01:16:01.860 | What was your takeaway?

01:16:02.860 | Well, look, I mean, I got to tell you, I was floored a bit when I saw this, mainly because

01:16:10.340 | of the peculiarity of it all.

01:16:12.380 | So to have a company that is one of the three, let's just call it biggest backed AI companies

01:16:21.100 | in Silicon Valley, not get bought, like if they got bought, that means, "Oh, Microsoft

01:16:26.700 | thought this was strategic."

01:16:28.220 | Microsoft, you remember when Facebook brought WhatsApp, it was 10% of market cap.

01:16:35.220 | So 10% of Microsoft's market cap is $300 billion.

01:16:39.620 | So they could have paid $10 billion, and the last round here was at $4 billion.

01:16:44.060 | And you'd go, "Oh, great, you know, everyone loves AI, investors are getting paid, everything

01:16:49.740 | worked out."

01:16:50.740 | But this isn't everything worked out.

01:16:52.660 | This is employees, all the key employees leaving the company, and the blog post suggests the

01:16:59.620 | company goes on with a new CEO.

01:17:02.740 | And to me, I just had to look this up because I don't watch racing, but the yellow flag

01:17:08.140 | comes out when there's a hazard on the track and everyone needs to slow down.

01:17:13.420 | And you look at this Series A here, a well-known venture firm, Greylock, 260 or 250 at a billion

01:17:20.580 | pre.

01:17:21.580 | These are these AI deals that have remained like Zerp-like, right?

01:17:27.220 | And this is a warning sign that you can't just pay any price for any deal in the AI

01:17:33.500 | sector and get paid.

01:17:35.340 | And because I don't, you know, I don't see this now, I don't see the current status as

01:17:41.460 | a company that's going to have a venture-like return from here.

01:17:46.340 | There's a lot of details, there's a lot of these credit things that are all tied up,

01:17:49.860 | there's stuff we don't know.

01:17:51.740 | But it's unusual, and I view it as a yellow flag that we should slow down on the track,

01:17:58.260 | on the AI track, and maybe be a little more careful.

01:18:01.180 | Well, one of the things that it brings up too is a conversation you and I were having,

01:18:06.580 | you know, just about LLMs generally, you know.

01:18:09.460 | I spent the last two days down at NVIDIA, you know, GTC, their Global Developers Conference.

01:18:16.260 | It's just extraordinary.

01:18:18.940 | The ecosystem of AI that they are consuming, right?

01:18:22.660 | This is not a chips company.

01:18:24.020 | This is not just a systems company or a supercomputer company.

01:18:27.580 | They are now expanding on the CUDA stack with a totally new inference software cluster called

01:18:34.100 | NIMS.

01:18:35.460 | And in NIMS, you know, they have embedded, you know, open-source models that are, you

01:18:42.420 | know, highly and tightly integrated with their overall software stack.

01:18:46.980 | And one of the things we've talked about and debated the last 12 months is just the durability.

01:18:52.860 | Like, what's the business model for the LLM?

01:18:57.060 | Is it going to, you know, you have these open-source models by well-financed companies like Meta

01:19:02.740 | that don't seem intent on charging for them.

01:19:06.700 | And when you have companies that are valued at $10, $20, $90 billion, right, around that,

01:19:13.220 | the real question was about the durability of revenue.

01:19:16.980 | And you know, we have a vibrant debate within Altimeter on this, right?

01:19:23.140 | So the acronym we all know when it comes to software companies like these LLMs is ARR.

01:19:30.560 | So everybody wants to know what's their revenue.

01:19:33.380 | ARR stands for Annual Recurring Revenue.

01:19:37.380 | And recurring revenue, like a monopoly cable subscription, gets a really high multiple.

01:19:43.740 | That's why softwares have gotten high multiples.

01:19:45.540 | They're very predictable.

01:19:47.860 | But when we look at what's going on in AI, I've banned the use of the term ARR within

01:19:55.380 | Altimeter when we talk about AI revenues, because they're certainly not recurring like

01:20:01.140 | a monopoly.

01:20:02.140 | They're consumption-based.

01:20:03.140 | And they have some amount of cogs tied to them.

01:20:04.820 | Right.

01:20:05.820 | So we call them ERR, Experimental Run Rate Revenues, right?

01:20:12.780 | And so when I look at these, you know, one of the things that inflection and these other

01:20:19.500 | models that it's been very challenging for us to get our arms around backing, even though

01:20:24.440 | they have extraordinary founders and are building extraordinary things, is just if you have

01:20:29.740 | to pay a high price, then you need to have confidence in the durability of those revenues

01:20:34.640 | and profits on a go-forward basis.

01:20:37.420 | And I think you have some thoughts on this.

01:20:38.900 | Yeah, look, there's nothing that I've spent more time on recently, just trying to learn

01:20:44.460 | and ask questions, but, you know, both trying to understand how these things work, but also

01:20:48.780 | how they're being used.

01:20:50.460 | And one thing you constantly hear from developers on LLM, and let's separate this maybe from

01:20:55.340 | the consumer game that I guess—and then there were two or three, right?

01:20:59.740 | You got Microsoft.

01:21:00.740 | I guess you got Google perplexing OpenAI, $20 a month.

01:21:04.220 | That's a whole separate conversation.

01:21:05.500 | But on this B2B side, the model that just seems so prevalent with all the companies

01:21:11.700 | I talk to, you might test your thesis against one of the premier models, but then the minute

01:21:18.980 | you go to run it, you then run it on four or five others and see which one—at different

01:21:25.160 | price points and see which one is cost-effective for you.

01:21:28.740 | And the promiscuity is unbelievably high.

01:21:34.980 | Another way of saying that, one of the things that leads to high valuation multiples is

01:21:38.180 | switching costs.

01:21:39.540 | And today, you know, as it stands, I would say if you were to rate these LLMs on their

01:21:45.240 | switching costs with zero maybe being none and 100 being perfect, I'd say they're at

01:21:51.620 | two.

01:21:52.620 | Like, it's remarkably low.

01:21:54.780 | And there's another huge irony.

01:21:56.540 | Some people believe—and I won't try and make this argument, I'll just say I know

01:22:00.260 | some smart people have said it—that as the context window gets bigger, there'll be less

01:22:05.860 | training and even less fine-tuning, which means you're just dumping data into the prompt.

01:22:12.340 | And this data is not being dumped in in a structured way.

01:22:15.580 | It's just PDFs or CSV files or whatever.

01:22:19.260 | There is no lock-in.

01:22:20.260 | You just dump it in the next one.

01:22:22.420 | So I think one thing that's imperative in any of these LLM models, if they want to be

01:22:27.180 | successful in the long run in terms of creating equity value, they've got to find a way

01:22:31.020 | to get to switching costs.

01:22:32.780 | And I don't know what that is exactly right now.

01:22:35.740 | I think some element of fine-tuning or embedding data or memory like we've talked about.

01:22:40.600 | But I don't think they have them today.

01:22:42.380 | And maybe this is a sign of that.

01:22:44.580 | I don't know.

01:22:45.580 | It'll be interesting to watch.

01:22:46.580 | Well, I think along those lines, I had dinner this week with one of the heads of data and

01:22:52.900 | AI at one of the largest banks in the United States.

01:22:56.460 | And they did what I think a lot of other people did after they saw ChatGPT.

01:23:00.980 | So remember, ChatGPT comes out in Q4 of '22.

01:23:05.100 | The CEOs of all these banks and every other enterprise went home, and that's what the

01:23:08.940 | conversation was around the holiday table.

01:23:11.180 | So the CEOs or the board came back in Q1 and they said, "Hey, what are we doing in AI?"

01:23:15.480 | And they turned to their tech team and they said, "Let's get something done."

01:23:18.380 | So in Q2 of 2023, you saw all of these new customers sign up the Azure OpenAI API, right?

01:23:27.380 | Or do it directly with OpenAI because it was the only game in town.

01:23:31.760 | But quickly, all of the other data platforms, whether you were AWS, whether you're Google,

01:23:39.980 | Snowflake, Databricks, et cetera, they all spun up their own AI capabilities.

01:23:45.060 | Now they all embed free models, whether it's Lama, Mistral, et cetera.

01:23:48.840 | Even Microsoft now has competitive open models.

01:23:51.940 | And my hunch is that OpenAI will also be found on some of these data platforms, not just

01:23:59.540 | on Microsoft.

01:24:02.100 | And so it seems to me the switching cost is not with the LLM.

01:24:06.060 | I agree with you on this.

01:24:07.620 | The switching cost is really around where your data resides, right?

01:24:11.960 | We've heard this talked about, data gravity.

01:24:14.480 | So this particular bank, which had spun up the OpenAI API in Q2, has been playing with

01:24:21.440 | it on a big rag over the course of the last year.

01:24:24.640 | But I asked them what their plans were going forward.

01:24:27.160 | They said, "Oh, we're moving.

01:24:28.820 | We're going to run all of our AI in AWS, right, on open models, on our production data," right?

01:24:36.360 | So it seems to me that the AI is finding its way back to the data.

01:24:41.480 | And the only way, I think, for Azure or anybody else to keep this experimental revenue when

01:24:46.460 | people run to use these models is you have to deliver something sufficiently differentiated

01:24:52.400 | that it gets somebody to pull all of their data to you.

01:24:55.700 | And that's a hard road to hoe.

01:24:58.000 | And I'm reminded – I don't know if I've used this metaphor before, but there was a

01:25:02.460 | time before digital music where you just couldn't – there was so much music in the world,

01:25:07.140 | you couldn't imagine it.

01:25:08.140 | And I remember one day back in the Napster days where someone tapped me on the shoulder

01:25:12.680 | and he said, "Do you see this?"

01:25:13.680 | And he was holding a hard drive and he said, "This is all the music I've ever recorded,"

01:25:18.060 | and like he just had it.

01:25:19.780 | And to a certain extent, you know, the LLM, it may be – and I've heard this argument

01:25:25.300 | from some academics – it may be that most of the data, think Wikipedia, everything we

01:25:30.860 | know is already in there.

01:25:32.580 | And like it's gotten in, you know, small very quickly.

01:25:36.940 | And what – the reason I bring that up relative to what you said is it may be that the open

01:25:42.180 | source LLMs are good enough and that putting them proximate to your data is a smarter play

01:25:49.500 | than trying to put your data proximate to the leading proprietary LLM.

01:25:54.780 | And if that plays out, that'll be super interesting.

01:25:59.060 | And one thing's for sure, the open source LLM horse is out of the barn, right?

01:26:05.820 | It's running wild and free.

01:26:07.180 | I mean, and you've got to love the competition.

01:26:11.060 | And you know, again, it's a – you and I have talked about some of the concerns about

01:26:16.900 | regulatory capture as, you know, in preventing this competition.

01:26:22.460 | You know, going back to the conversation about China, it's so important that we protect

01:26:26.880 | this competition among these models because it's driving down the cost to all enterprises.

01:26:31.460 | It's driving down the cost to all inference, to consumers.

01:26:36.020 | And you know, all I know is it's going to be an exciting few months ahead.

01:26:39.140 | Always.

01:26:40.140 | Always.

01:26:41.140 | So fun to have you here.

01:26:42.140 | Fun to do this again.

01:26:43.140 | All right, man.

01:26:44.140 | And until next time.

01:26:45.140 | Take care.

01:26:45.140 | Bye-bye.

01:26:46.140 | Bye-bye.

01:26:47.140 | Bye-bye.

01:26:48.140 | Bye-bye.

01:26:49.140 | Bye-bye.

01:26:50.140 | Bye-bye.

01:26:51.140 | Bye-bye.

01:26:51.140 | (dramatic music)

01:26:53.900 | you

Ep5. Stock Comp, AI Cold War, Valuations for LLM | BG2 with Bill Gurley, Brad Gerstner, & Bob Mylod

Chapters