BG2 Rewind: 2024 Year-End Highlights

00:00:00.000 | So the period 1820 to 1900, it only takes 87 years to double GDP.

00:00:07.140 | And then 1900 to 1950, only 60 years.

00:00:10.940 | And by the time you get to the year 2000, we're doubling global GDP every 20 to 30 years.

00:00:18.700 | What have earnings compounded at for tech versus non-tech?

00:00:37.980 | And what have stock prices compounded at at tech versus non-tech?

00:00:42.540 | Technology companies have compounded earnings at 13%.

00:00:45.740 | Their stock prices have compounded at 17%.

00:00:48.280 | So a little bit faster than earnings have compounded.

00:00:51.540 | But if you look at non-tech, they've only compounded earnings at about 6%.

00:00:56.460 | Their stock prices have grown at about 8%.

00:01:00.180 | Technology has gone from 5% of global GDP to 15% over the course of the last 15 years.

00:01:06.260 | When we have this conversation five or 10 years now, is tech going to be more or less

00:01:10.460 | than 15% of global GDP?

00:01:12.820 | By the grace of God, Apple was created here.

00:01:15.380 | Google was created here, Microsoft was created here, Meadow was created here.

00:01:18.580 | It is not a given that in 20 years from now, those companies will be created here.

00:01:22.100 | Why not just like seal that and lock that in as a monopoly and get all of that talent

00:01:26.660 | here?

00:01:27.660 | And the thing that is crazy to me is that US companies are currently employing all of

00:01:32.940 | the people we're talking about moving here.

00:01:34.740 | They're just employing them in a different country.

00:01:36.860 | Would you rather have an assistant with the intelligence of like Einstein, but they have

00:01:42.380 | no access to the internet and they don't know anything about your history?

00:01:46.620 | Or would you like to have an assistant that's just above average intelligence, knows everything

00:01:51.260 | about you, and they can use the internet?

00:01:54.060 | You would choose that.

00:01:55.060 | I think that is another 10x moment that's in front of us, which is we go from answers

00:02:00.320 | where I'm just asking for information to actions.

00:02:03.260 | And once it can start booking my hotel, booking, reserving my restaurant, and then I just say

00:02:09.340 | same thing, do it again.

00:02:11.740 | And over the course of the next four to five years, we'll have 2 trillion of data centers

00:02:16.700 | powering software around the world, and it will all be accelerated compute.

00:02:22.180 | So this is the AI data center build out.

00:02:25.500 | In blue is the new accelerated compute.

00:02:29.520 | In green is the replacement data center that we think will go to accelerated.

00:02:35.780 | And then in gray is the replacement that's non-accelerated compute.

00:02:39.560 | So this would be more like, you know, x86.

00:02:43.020 | I would make the argument that every company in Delaware has to move to a different domicile

00:02:49.900 | because they could be sued in a future derivative law suit for the risk they've taken by staying

00:02:56.300 | in Delaware.

00:02:57.300 | Oh my God, you're so right.

00:02:59.300 | A dollar of stock-based compensation is actually, not only should it not be ignored, it's the

00:03:05.100 | most single valuable expense.

00:03:07.500 | We even have a court criticizing Elon for taking an options package where he made no

00:03:13.640 | money unless he saved the company from bankruptcy.

00:03:17.220 | Meanwhile, the CEO of his arch rival, who created no shareholder value over a period

00:03:22.660 | of five years, is making tens of millions of dollars a year in RSUs.

00:03:27.000 | If you give one share of a restricted stock unit at a value of 20, then the recipient

00:03:33.660 | has $20 of value.

00:03:36.140 | Now if the stock happens to go down by 50%, the employee still has $10 of value.

00:03:42.100 | Whereas in the case of the stock option, if it goes down 1%, yet alone 50, it has zero

00:03:47.460 | value.

00:03:48.460 | I like to take a hot shower.

00:03:50.420 | But if the cost of that was $100,000, right, not many people would take hot showers.

00:03:57.000 | But he's saying that if you drive down the price of the cost of compute, then the reflexivity

00:04:03.360 | is people will consume a lot more of it.

00:04:05.940 | Now, this is also known as the Jevons paradox, right?

00:04:09.900 | As price goes down, we demand more of it.

00:04:13.660 | The aggregate amount of consumed, of the compute consumed, actually goes up, right?

00:04:20.380 | And that's really, you know, just a fancy way of talking about the elasticity of demand.

00:04:26.020 | If you're public with $100 million in revenue and a 10% growth rate, your valuation's not

00:04:32.660 | going to be all that great.

00:04:33.940 | But guess what?

00:04:34.940 | If you're private at $100 million in revenue with a 10% growth rate, it's not like you're

00:04:39.660 | better off, like you're just fooling yourself.

00:04:43.540 | Elon is building a much, much bigger cluster to train a much, much bigger model as is OpenAI,

00:04:50.660 | as is Zuckerberg.

00:04:51.660 | I mean...

00:04:52.660 | What Sam just said, the bigger models aren't the problem.

00:04:55.100 | Well, I mean, he may be doing the same game that everybody else is doing, Bill, and trying

00:04:58.980 | to throw everybody off the scent.

00:05:00.940 | Dario and Sam talk in these high-level platitudes about how this stuff's going to cure cancer

00:05:06.620 | and we're all going to not have to work anymore.

00:05:09.900 | And Zuck was down in the weeds in the meat.

00:05:12.220 | He's in the arena.

00:05:13.380 | Yeah, being super transparent.

00:05:15.420 | And I was just like, holy shit, maybe this guy's in charge now.

00:05:18.940 | Hey, man, great to see you.

00:05:20.900 | That was fun to take some of your money in Vegas this past weekend.

00:05:24.860 | I can't believe you're bringing that up.

00:05:27.040 | This is classic, you know, freshman economics.

00:05:30.540 | If someone's better at doing something than us, we should let them do it.

00:05:33.940 | We should buy it from them and we should send them what we're good at.

00:05:36.860 | It's interesting.

00:05:37.860 | We end up talking about AI every week, but I don't know what there is else to talk about

00:05:41.540 | because that's all anyone's talking about.

00:05:43.980 | And so we're gearing up for a battle royale.

00:05:48.620 | That's what it feels like.

00:05:49.780 | The forecast we have in here is simply the consensus Morgan Stanley and Goldman Sachs

00:05:54.180 | forecast for the balance of the year, which has been pretty accurate.

00:05:58.100 | The backdrop around inflation continues to be constructive.

00:06:02.580 | We had a couple of months in there where I think people got scared.

00:06:05.220 | Larry Summers said maybe the next rate move is up.

00:06:09.140 | Why does that matter so much, Bill?

00:06:11.140 | Just listen to the Berkshire Hathaway annual meeting.

00:06:14.520 | You know, Warren Buffett says, listen, I'm collecting my my interest payments every week.

00:06:18.820 | If I can earn five point four percent taking no risk, why would I take risks?

00:06:22.500 | So, you know, I still am in the camp that, you know, we're on a glide path.

00:06:27.780 | It's going to take a little bit longer.

00:06:30.240 | But I do suspect that rates are going to come down.

00:06:32.700 | Jay Powell said yesterday we're on hold for a couple more months.

00:06:35.940 | I think you're going to get a rate cut before the election.

00:06:38.640 | But I don't think the market actually needs a rate cut, Bill.

00:06:42.180 | What they need to know is that inflation is coming down and the Fed can give us a rate

00:06:46.180 | cut if they want to.

00:06:47.780 | It's very difficult to stay fit and efficient when you have a buffet of all options sitting

00:06:54.960 | in front of you and you can fund all of them.

00:06:57.340 | Scarcity breeds necessity, scarcity breeds innovation.

00:07:01.620 | Elon set out a vision for rockets that could land themselves and auto fleets that would

00:07:06.260 | be replaced by electric cars.

00:07:08.300 | At the time he said these things, they sounded totally outlandish.

00:07:11.720 | You have to will the future.

00:07:13.100 | You have to manifest the future.

00:07:14.500 | You have to describe the future.

00:07:16.020 | You have to motivate your employees.

00:07:17.860 | And importantly, you have to motivate sources of capital.

00:07:21.540 | People talk about the military industrial complex.

00:07:23.580 | We may have created a healthcare industrial complex that really can't stop maximizing

00:07:31.940 | profitability and not focus necessarily on the lowest cost, best, most preventative process.

00:07:39.980 | What do you do for yourself, your friends, and your family?

00:07:42.940 | Just tell me your care, standard of care for yourself, your friends, and your family.

00:07:48.500 | And 100% of them follow the same protocol, which is a calcium CT scan no later than 40.

00:07:55.500 | Five times as many women die of a heart attack each year in this country as die of breast

00:07:59.820 | cancer.

00:08:00.820 | Very, very few women do a calcium CT scan, which I might argue or I've heard argued is

00:08:07.300 | the mammogram for the heart.

00:08:09.260 | And instead of a $250 marginal cost to implement a rural broadband, we're going to get bulldozers

00:08:16.300 | out and drop fiber lines to a ranch in the middle of nowhere.

00:08:21.820 | It's stupid.

00:08:23.340 | That shot fired at Trump was one inch different.

00:08:27.340 | If Trump had been assassinated, can you imagine what would happen to the U.S. markets?

00:08:33.360 | So when markets are at all-time highs and you're just hanging out and there's not a

00:08:37.340 | lot of upside return and you got these other things you're concerned about, you got to

00:08:41.300 | take down those units of risk because there are always things like this that can happen.

00:08:45.340 | Take Tiablo Canyon.

00:08:46.700 | They built two nuclear reactors here.

00:08:48.980 | You know, this site was provisioned for six.

00:08:53.020 | I could imagine if you had four more reactors sitting here, you could have one reactor for

00:08:58.620 | Meta, one for Amazon, one for Microsoft, one for NVIDIA.

00:09:02.820 | You could have the data center sitting right next to them.

00:09:05.340 | I think this anti-Chinese mentality is a little nuts because if you look at low-cost EVs,

00:09:12.980 | if you look at their subway stations, if you look at, they're leading us in many areas.

00:09:19.380 | And we have this holier-than-thou attitude that America is the best and that we're the

00:09:24.700 | leader and we're being passed.

00:09:27.060 | We would all be better off if nuclear energy were cheaper for everyone on the planet.

00:09:31.220 | I'm no longer going to program computers with C++.

00:09:34.380 | I'm going to program AIs with prompting.

00:09:38.460 | Isn't that right?

00:09:39.460 | Now, this is no different than me talking to my, you know, this morning.

00:09:42.340 | I wrote a bunch of emails before I came here.

00:09:44.340 | I was prompting my teams, right?

00:09:47.420 | And I would describe the context.

00:09:49.100 | I would describe the fundamental constraints that I know of, and I would describe the mission

00:09:54.460 | for them.

00:09:55.460 | I would leave it sufficiently, I would be sufficiently directional so that they understand

00:10:00.580 | what I need.

00:10:01.580 | And I want to be clear about what the outcome should be, as clear as I can be.

00:10:05.220 | But I leave enough ambiguous space on, you know, a creativity space so they can surprise

00:10:09.580 | me.

00:10:10.580 | Isn't that right?

00:10:11.580 | Isn't that how I prompt an AI today?

00:10:13.180 | It's exactly how I prompt an AI.

00:10:15.220 | Companies are only limited by the size of the fishpond, you know.

00:10:19.380 | A goldfish can only be so big.

00:10:21.780 | What is our fishpond?

00:10:22.980 | What is our pond?

00:10:24.660 | And that requires a lot of imagination.

00:10:26.420 | And this is the reason why market makers think about that future, creating that new fishpond.

00:10:35.180 | It's hard to figure this out looking backwards and try to take share.

00:10:38.860 | The need to invent a new market to go serve it later is something that's very comfortable

00:10:44.020 | for us.

00:10:45.020 | Exactly, exactly.

00:10:46.020 | I've said before, the data center is now the unit of computing.

00:10:50.700 | To me, when I think about a computer, I'm not thinking about that chip.

00:10:53.820 | I'm thinking about this thing.

00:10:55.020 | That's my mental model and all the software and all the orchestration, all the machinery

00:10:58.820 | that's inside.

00:11:00.620 | That's my computer.

00:11:01.620 | And we're trying to build a new one every year.

00:11:05.820 | That's insane.

00:11:07.100 | Nobody has ever done that before.

00:11:08.100 | You may also be running up against the, even for the Mag7, the size of Capo X deployment

00:11:15.580 | where their CFOs start to talk at higher levels.

00:11:18.600 | You can have a thousand agents working together.

00:11:20.740 | You can have one that's making sure that the credit card charge is not too big.

00:11:23.900 | You can have another one to make sure that the address is right.

00:11:26.140 | You can have another one checking against your calendar.

00:11:28.460 | And so all of that's free.

00:11:29.980 | So I'm on the under and Brad, I'll even go under one year.

00:11:33.380 | Wow.

00:11:34.380 | Yeah.

00:11:35.380 | Wow.

00:11:36.380 | Yeah.

00:11:37.380 | That's crazy, Sonny.

00:11:38.380 | 10, 15 years ago, the funds were smaller.

00:11:41.460 | And as a GP, as an investor, you'll make money on that guaranteed portion, but not necessarily

00:11:47.140 | right.

00:11:48.140 | As I call it, get rich money.

00:11:49.140 | Bill, what were you making when you started, per year when you started in the business?

00:11:54.060 | Every founder and every company and every board will tell you, if we raise money, we're not

00:11:57.300 | going to spend it.

00:11:58.300 | We're going to stay frugal.

00:11:59.420 | That never happens.

00:12:00.620 | You might ask a founder, what are the three things that matter most this year that are

00:12:03.900 | critical for your success?

00:12:05.580 | If you only focused on those three things, you might maximize your chance of success.

00:12:10.340 | When you have lots of money, you're going to do six things at once, which means you

00:12:13.740 | are diluting the three things that matter.

00:12:15.740 | I think we're all actors stuck in the game.

00:12:19.020 | Right.

00:12:20.020 | You know, I'm not sure you can escape it.

00:12:22.740 | You know, a couple of observations.

00:12:24.220 | Is that the positive take on it?

00:12:25.780 | No.

00:12:26.780 | I know he wanted one.

00:12:28.060 | I apologize.

00:12:29.060 | If 2024 was the chat GPT moment for full self-driving, Bill, where we started seeing Waymo's on

00:12:35.380 | every street corner in San Francisco, if this was the chat GPT year, I think next year really

00:12:40.280 | is the year of achievement of a safety standard that allows RoboTaxi to go into action.

00:12:48.460 | The Department of Government Efficiency, the unofficial government department, which is

00:12:54.100 | run by Vivek and Elon, and is overseeing trying to reform government spending, government

00:13:02.140 | regulations, frankly, reducing the size of the federal government with the overall objective

00:13:08.100 | of getting control of our national deficit and our national debt, which we all agree

00:13:13.100 | is pretty egregious.

00:13:14.100 | And, you know, if you look at the federal revenues, just over $5 trillion expected this

00:13:19.500 | year, government spending $7 trillion.

00:13:22.020 | So we basically have a $2 trillion deficit.

00:13:25.020 | That spending is up dramatically, by the way, from the $4.5 trillion that we spent in 2019.

00:13:29.780 | So in COVID, we just lost our mind and we've never regained our mind.

00:13:33.260 | If you just grow revenues from here at 3% to 4% over the course of the next four years,

00:13:37.940 | Bill, and you cut costs in the first year by 5%, second year by 5%, third year by 2%,

00:13:44.060 | so all that does is get you back to trend line from 2019, as though we increased spending

00:13:50.260 | 3% a year from 2019, the budget balances.

00:13:54.860 | That's tiny, these are tiny changes in costs.

00:13:59.500 | And I shared that with Vivek and Elon, and they both had separate and independent reactions,

00:14:05.220 | which is not nearly ambitious enough.

00:14:07.900 | Invest America, Bill, you know, Invest America is a project and now piece of legislation

00:14:12.620 | that I've been working on for three years, and it's truly gaining a ton of traction.

00:14:18.080 | The federal government would cause to be created 3.7 million investment accounts a year for

00:14:23.740 | every child that's born in America.

00:14:25.860 | They would seed it with a very small amount of money that would be a roundy near to the

00:14:30.380 | federal government.

00:14:31.380 | It has a massive ROI, and we can achieve it while having less government, not more government.

00:14:38.900 | I think the notion that business applications exist, that's probably where they'll all collapse,

00:14:48.540 | right, in the agent era.

00:14:50.100 | Because if you think about it, right, they are essentially CRUD databases with a bunch

00:14:56.500 | of business logic.

00:14:58.820 | The business logic is all going to these agents.

00:15:03.980 | And these agents are going to be multi-repo CRUD, right, so they're not going to discriminate

00:15:09.780 | between what the backend is.

00:15:12.820 | They're going to update multiple databases and all the logic will be in the AI tier,

00:15:20.220 | so to speak.

00:15:21.560 | And once the AI tier becomes the place where all the logic is, then people will start replacing

00:15:27.300 | the backend.

00:15:28.300 | I think the company of this generation has already been created, which is open AI in

00:15:33.020 | some sense.

00:15:34.700 | It's kind of like the Google or the Microsoft or the meta of this era.

00:15:40.420 | I mean, I always say Google makes more money on Windows than all of Microsoft.

00:15:44.940 | I mean, literally.

00:15:45.940 | I mean, and I say, wow.

00:15:47.940 | Very few people could use GPT-2 and 3, right?

00:15:50.740 | A lot of people can use GPT-4.

00:15:52.780 | When we get to that quality of jump that we see for the next generation, the amount of

00:15:57.000 | people that can use it, the tasks that it can do, balloons out, and therefore the amount

00:16:01.100 | of sort of white-collar jobs that it can augment increased productivity on will grow, and therefore

00:16:05.380 | the market clearing price for that token will be very high.

00:16:07.660 | That's super interesting.

00:16:08.660 | If that's going to be dead, then why is Mark Zuckerberg building a two-gigawatt data center

00:16:11.220 | in Louisiana?

00:16:12.220 | Right.

00:16:13.220 | Why is Amazon building these multi-gigawatt data centers?

00:16:16.040 | Why is Google, why is Microsoft building multiple gigawatt data centers, plus buying billions

00:16:21.080 | and billions of dollars of fiber to connect them together because they think, hey, I need

00:16:24.980 | to win on scale, so let me just connect all the data centers together with super high

00:16:28.180 | bandwidth so then I can make them act like one data center, right, towards one job, right?

00:16:32.560 | So this whole, like, is scaling over narrative falls on its face when you see what the people

00:16:38.020 | who know the best are spending on.

00:16:39.940 | Until next year.

00:16:40.940 | Awesome.

00:16:41.940 | Thank you.

00:16:42.940 | Take care.

00:16:42.940 | As a reminder to everybody, just our opinions, not investment advice.

BG2 Rewind: 2024 Year-End Highlights

Chapters