E169: Elon sues OpenAI, Apple's decline, TikTok ban, Bitcoin $100K?, Science corner: Microplastics

00:00:00.000 | - Chamath, who are you giggling with?

00:00:01.700 | Are you with your kids?

00:00:02.540 | What's going on here?

00:00:03.540 | You're making googly face.

00:00:04.540 | - I got a message from that.

00:00:06.420 | - Oh no, is it Hellmuth?

00:00:08.120 | - No, no, I got a message from that.

00:00:09.860 | She's so funny.

00:00:10.820 | So I was like, blah, blah, blah.

00:00:14.500 | She did a Uranus joke into my chat and I lost it.

00:00:19.500 | - It is like Friedberg always says,

00:00:22.960 | Uranus is right there at the edge of the universe.

00:00:26.760 | No, she's like, are you hungry?

00:00:28.200 | I could eat a Uranus.

00:00:30.080 | Oh no, we could go to a great restaurant, Uranus.

00:00:33.880 | They serve great meals there.

00:00:36.520 | They have a great chocolate lava cake.

00:00:38.540 | (upbeat music)

00:00:41.120 | - All right, everybody.

00:00:56.080 | Welcome back to your favorite podcast of all time.

00:00:59.800 | It's episode 169 of All In with me again,

00:01:04.040 | the chairman dictator, Chamath Palihapitiya,

00:01:06.980 | David Friedberg, your Sultan of Science

00:01:09.480 | and the Rain Man, yeah.

00:01:12.280 | Definitely David Sacks has his Montclair hat back again.

00:01:15.680 | I guess they're back in stock.

00:01:16.980 | How we doing boys?

00:01:17.820 | Welcome back to the show.

00:01:19.080 | Let's get to the docket here.

00:01:20.920 | Issue one, Elon has sued OpenAI, begun the meme wars have.

00:01:26.020 | After we finished the recording last week,

00:01:28.800 | the memes are incredible.

00:01:30.520 | After we finished recording last week,

00:01:31.760 | Elon sued Sam Waltman, Greg Brockman

00:01:33.960 | and the OpenAI organization.

00:01:36.020 | He's suing for breach of contract,

00:01:38.240 | fiduciary duty and unfair competition.

00:01:40.420 | And his argument is basically OpenAI started

00:01:42.760 | as an open source nonprofit.

00:01:43.880 | As you know, he gave them something like 50 or 75 million

00:01:47.120 | of both those numbers quoted.

00:01:49.320 | And then of course, everybody knows OpenAI

00:01:51.560 | turned into closed AI.

00:01:52.840 | They became a closed source for profit venture

00:01:55.280 | after the tech was deployed.

00:01:57.420 | This enabled them obviously to benefit

00:02:00.260 | from nonprofit tax breaks while building the tech,

00:02:03.820 | but they now have two corporate entities.

00:02:07.260 | There's this nonprofit called OpenAI.

00:02:09.240 | There's a for-profit, it's called OpenAI Global LLC.

00:02:11.640 | That was created in 2019.

00:02:13.900 | And there's all this funky relationship between the two.

00:02:16.320 | We'll get into that

00:02:17.160 | 'cause it's kind of interesting actually.

00:02:18.520 | Elon said that if OpenAI is allowed to do this,

00:02:21.020 | then it should be the standard

00:02:22.020 | for every company going forward.

00:02:23.160 | That's an interesting point.

00:02:24.400 | You can start by donating money to a nonprofit

00:02:26.240 | and then make a for-profit.

00:02:27.160 | It seems pretty capital efficient.

00:02:28.640 | So what does Elon want to get out of this?

00:02:30.620 | According to the lawsuit,

00:02:31.680 | he wants OpenAI to open source their models.

00:02:34.380 | By the way, that's what Facebook and Apple

00:02:35.760 | are doing right now.

00:02:36.600 | So that seems more than reasonable

00:02:37.920 | since that's how the company was formed.

00:02:41.280 | And he wants to make sure that shareholders

00:02:42.720 | receive no financial benefit from OpenAI.

00:02:45.080 | And we can get more into a bunch of the OpenAI

00:02:50.260 | nonprofit status and their structure.

00:02:52.400 | It's super convoluted.

00:02:54.880 | But I want to just get your initial reaction, Sax.

00:02:58.080 | I haven't read the legal filings, Jason.

00:03:00.280 | My knowledge of the case

00:03:01.240 | is just coming from following what's on Twitter.

00:03:03.480 | And the last time I commented on one of Elon's lawsuits

00:03:07.320 | on this pod, I ended up in six hours of deposition.

00:03:10.080 | So remember that was the whole Twitter lawsuit.

00:03:15.080 | At any event, my understanding of the case

00:03:17.060 | from the tweets that are flying back and forth

00:03:21.920 | is I think Elon is making two points.

00:03:24.260 | One is that when OpenAI was set up,

00:03:28.540 | it was set up to be a nonprofit

00:03:31.260 | and to promote AI as an open source technology

00:03:35.160 | that everyone could benefit from

00:03:36.840 | and no one large tech company would control.

00:03:40.540 | In that case, Elon was primarily concerned about Google.

00:03:43.320 | I think now he's more concerned about Microsoft.

00:03:45.560 | Nonetheless, the idea was that this would be open source.

00:03:49.380 | So I think point number one is Elon feels

00:03:52.020 | like the rug was pulled out from under him

00:03:54.060 | after he donated 40 something million dollars to this.

00:03:58.040 | They completely changed what it was gonna be.

00:04:00.380 | And I think he's used the word swindled before.

00:04:02.620 | He feels like he was cheated.

00:04:05.120 | And I think usually when there's a lawsuit like this,

00:04:07.620 | it's usually because somebody does feel

00:04:09.420 | fundamentally cheated by what happened.

00:04:12.140 | So I think that's the first point.

00:04:13.700 | The second is that I think that Elon's making is,

00:04:16.320 | well, wait a second, if you can start a company

00:04:20.020 | as a nonprofit using pre-tax dollars

00:04:23.160 | and then all of a sudden convert to for-profit,

00:04:25.640 | then why wouldn't everybody do this

00:04:27.080 | in order to circumvent paying taxes?

00:04:29.160 | And I think that's another, I'd say,

00:04:30.880 | valid and interesting point that he's made.

00:04:33.020 | Now, OpenAI has responded to this

00:04:35.520 | by publishing some of Elon's emails to them.

00:04:38.220 | And they think that they have a smoking gun here

00:04:40.300 | because apparently Elon told them

00:04:43.440 | they need to raise far more money

00:04:45.280 | in order to have a chance of taking on Google/DeepMind.

00:04:49.540 | I just don't know whether that's the smoking gun

00:04:51.160 | they claim it is, but that's basically

00:04:52.940 | what they're putting out there.

00:04:54.580 | And then to add fuel to the fire,

00:04:57.460 | you've got VCs coming over the top

00:04:59.080 | debating different aspects of this.

00:05:01.060 | You've got sort of Vinod defending OpenAI.

00:05:04.020 | He was the first VC investor in it.

00:05:06.240 | And so he's been out there attacking Elon.

00:05:08.860 | And then I think Marc Andreessen is responding to Vinod.

00:05:15.060 | I don't know if he's exactly defending Elon.

00:05:17.260 | Anyway, it's turned into a whole maelstrom on X,

00:05:20.560 | and then everyone else is starting

00:05:22.340 | to put out a whole bunch of crazy memes.

00:05:25.780 | So I don't know who's gonna win this case in court,

00:05:28.380 | but the memes are definitely lit.

00:05:31.500 | And maybe the funniest one is that Elon has said

00:05:36.020 | that if OpenAI will simply change their name

00:05:38.420 | to ClosedAI, he'll drop the lawsuit.

00:05:40.800 | - I think he's serious about that.

00:05:42.280 | - I think he is serious about it,

00:05:43.660 | but I think he's making his point,

00:05:45.040 | which is you ended up doing something different

00:05:48.900 | than what you told me when you got me

00:05:51.620 | to write this big donation,

00:05:52.780 | and when we co-founded this thing together.

00:05:55.620 | So I think that sort of sums up Elon's position on this.

00:05:59.720 | - Chamath, a year from now,

00:06:02.460 | where do you think we'll be with this?

00:06:04.340 | - Well, I think there's one more thing,

00:06:06.500 | which is that they used his name pretty aggressively

00:06:11.500 | to get more money.

00:06:13.120 | And he was very instrumental

00:06:15.380 | in getting some of the early critical hires,

00:06:18.340 | particularly Ilya, which has been, I think,

00:06:20.420 | well-documented to leave Google, right?

00:06:24.300 | So that's probably sacks where also some of this feeling

00:06:28.020 | that he was bamboozled comes from.

00:06:33.180 | But I think that all of those emotions matter less

00:06:37.540 | than the rule of law, which is his second point,

00:06:39.800 | which is the important point,

00:06:41.540 | irrespective of whatever one email says or another email.

00:06:44.880 | It's not great for the US tax system

00:06:47.720 | if all of a sudden a big gaping loophole

00:06:50.480 | is identified and taken advantage of.

00:06:52.620 | I think that there's a huge economic incentive

00:06:55.680 | for the state of California,

00:06:58.100 | every other state where these open AI employees lived

00:07:00.960 | and have gotten equity and now have gotten paid.

00:07:03.740 | There's an incentive for treasury.

00:07:06.000 | There's an incentive for the IRS.

00:07:09.020 | It touches a lot of aspects of very complicated tax law.

00:07:13.480 | And so I think that's why there will be a resolution

00:07:16.740 | to this case, because I think that that answer

00:07:19.820 | is very important.

00:07:20.780 | And I think it will, as he has correctly identified,

00:07:24.660 | motivate a lot of people's actions going forward.

00:07:27.800 | So independent of what the emotional situation is

00:07:31.460 | amongst all the actors in this play,

00:07:34.340 | the reality is he's identified a loophole

00:07:36.540 | and that loophole needs to get fixed.

00:07:37.860 | I'll give you a different example of this

00:07:40.100 | that's much more benign,

00:07:41.140 | but it dragged on for eight or nine years.

00:07:44.360 | At Facebook, there was a period where,

00:07:46.740 | and this was public, so I can tell you,

00:07:49.180 | we transferred a lot of our IP to Ireland at one moment

00:07:54.100 | and there was a transfer payment and whatnot.

00:07:56.620 | And then a few years afterwards when everybody realized,

00:08:00.260 | including us, and we had miscalculated,

00:08:02.160 | but when everybody realized the value of Facebook,

00:08:05.140 | that transfer payment did not seem correct.

00:08:07.380 | And it was a huge tax arb that we had facilitated, right?

00:08:12.140 | Because all of that IP sitting inside of Ireland

00:08:15.140 | gets taxed at a very different rate

00:08:18.140 | than that IP would have gotten taxed in the United States.

00:08:20.440 | And other companies had copied this.

00:08:22.580 | Long story short, the IRS sued.

00:08:24.420 | Similar to you, David, I was in years of depositions

00:08:28.380 | and interviews and all of this stuff.

00:08:30.460 | So the point is that the government really cares

00:08:33.060 | about these kinds of things

00:08:34.240 | because so much money is on the line.

00:08:36.960 | And if OpenAI turns out to be

00:08:38.980 | this multi-hundred billion dollar behemoth,

00:08:41.780 | this will get figured out in court

00:08:43.860 | because there's just too much money at stake.

00:08:46.240 | - Freeberg, any thoughts and maybe Steelman,

00:08:50.060 | if you feel like it, OpenAI's position?

00:08:53.140 | - I think we're all ignoring the documents

00:08:57.000 | that OpenAI put out yesterday,

00:08:58.700 | showing emails and interactions with Elon,

00:09:02.100 | where Elon acknowledged and recognized the necessity

00:09:05.100 | of having a for-profit subsidiary

00:09:08.120 | that could pursue the interests of the foundation

00:09:12.460 | by raising billions of dollars of outside capital.

00:09:16.000 | I think it's a really interesting set of facts

00:09:20.080 | that provides a different light on the story.

00:09:23.300 | And it was really important that OpenAI released it.

00:09:25.860 | I'm gonna share a-

00:09:27.080 | - Hold on, I didn't mention that.

00:09:29.220 | - Freeberg, I mentioned that too.

00:09:30.560 | My point is it doesn't allow you to break the law.

00:09:32.720 | - Yeah, so let me just tell you guys a little bit

00:09:35.580 | about another example.

00:09:37.720 | My sister works at a nonprofit

00:09:39.820 | called the Cystic Fibrosis Foundation.

00:09:41.980 | Cystic Fibrosis is an inherited disorder,

00:09:44.860 | affects the lungs and digestive tract.

00:09:46.760 | It's debilitating, really affects children.

00:09:50.660 | There was a nonprofit called

00:09:53.760 | the Cystic Fibrosis Foundation established in 1955.

00:09:56.360 | They did tons of work on drug development research.

00:09:58.540 | And it was a nonprofit for years until they realized

00:10:00.860 | that there needed to be a market-based system

00:10:02.680 | to create the necessary incentives

00:10:04.240 | to drive the capital needed to find a drug

00:10:07.000 | that could cure cystic fibrosis.

00:10:09.300 | And in the year 2000,

00:10:11.480 | they invested out of their foundation endowment,

00:10:14.060 | $40 million in a company called Aurora Biosciences.

00:10:18.140 | Subsequently, it was acquired by Vertex

00:10:20.280 | and they continued to invest another $60 million,

00:10:22.560 | so a total of 100 million bucks in the lifetime

00:10:25.100 | of the development of a drug that could cure this disease.

00:10:28.740 | In 2005, a drug was discovered,

00:10:32.300 | and in 2012, the FDA approved it.

00:10:35.180 | The nonprofit then sold their rights,

00:10:37.560 | their interest in this for-profit entity for $3.3 billion.

00:10:42.560 | It was an incredible return.

00:10:44.860 | It was the largest investment return

00:10:47.500 | by a nonprofit in history.

00:10:49.660 | - Who do they sell it to, back to the company?

00:10:52.700 | - I actually think they sold the rights to royalty pharma,

00:10:55.260 | who you and I know well.

00:10:57.340 | And that $3.3 billion opened up this ability

00:11:01.180 | for them to continue to make investments.

00:11:02.620 | They now fund biotech VCs

00:11:04.620 | and they make direct investments and other things.

00:11:06.680 | But it really set the benchmark for this concept

00:11:08.680 | of what's called venture philanthropy,

00:11:10.500 | where a nonprofit parent company can make investments

00:11:13.260 | in for-profits that can raise additional capital

00:11:17.060 | that's needed to pursue the broad, difficult interests

00:11:20.400 | of the nonprofit.

00:11:21.860 | And I think that this argument really kind of ties

00:11:24.180 | into what happened with OpenAI.

00:11:25.500 | And you can see it in the email exchanges with Elon,

00:11:28.060 | where he was so prescient

00:11:29.460 | that I give Elon extraordinary credit for this,

00:11:31.720 | that he saw that this is going to take billions of dollars

00:11:33.740 | a year of investment to realize the pursuit

00:11:36.260 | that OpenAI was going after.

00:11:37.620 | And there was no way that Elon was going to be able

00:11:39.420 | to generate that cash himself or that Reid or others

00:11:42.060 | were going to just be able to pony up that money.

00:11:44.060 | They needed to have some sort of for-profit vehicle

00:11:46.600 | that would allow the market to work

00:11:48.100 | and allow capitalists to find their capital

00:11:50.660 | into this organization to make this investment interest.

00:11:53.540 | So I think the real question with respect

00:11:56.340 | to is OpenAI in trouble as a foundation

00:11:59.580 | is does the nonprofit own a meaningful piece

00:12:03.100 | of the for-profit entity?

00:12:04.420 | And I don't know the answer to that.

00:12:05.420 | I don't know if you guys do.

00:12:06.640 | - Sounds like it.

00:12:07.480 | I mean, yeah, it must, right?

00:12:08.820 | - So if they do, then they have an investment interest.

00:12:10.900 | The second question is does the nonprofit parent

00:12:14.180 | still do charitable stuff?

00:12:16.280 | Because if it doesn't, then there's a real question

00:12:18.900 | on the nonprofit status.

00:12:20.220 | I think you have to have a certain amount of your assets

00:12:22.220 | deployed every year in order to qualify for 501(c)(3).

00:12:25.860 | So I think the test and the courts will likely end up being,

00:12:28.600 | look, it's totally reasonable to have a for-profit entity,

00:12:31.260 | to fund a for-profit entity.

00:12:32.420 | Other nonprofits have done it,

00:12:33.820 | particularly when you need to attract billions of dollars

00:12:36.100 | of outside capital to make it work.

00:12:37.780 | The real question is does the nonprofit

00:12:39.300 | still do nonprofit stuff?

00:12:40.540 | - That's an interesting question because,

00:12:42.560 | Freiburg, do you think that when they booted off

00:12:46.780 | the nonprofit experts from the board,

00:12:51.020 | that they may have crossed a trip wire there?

00:12:53.860 | - Yes.

00:12:54.700 | - Yeah, I'm the lawyer.

00:12:55.540 | I'm the lawyer.

00:12:56.360 | That's a great question.

00:12:57.200 | - I did a little journalism,

00:12:58.100 | and I talked to a couple of people who are in and around

00:13:02.140 | two other of these scenarios.

00:13:03.420 | One of the key issues here is that the IP,

00:13:06.180 | in your example, Freiburg,

00:13:08.020 | it wasn't like that cystic fibrosis organization

00:13:11.380 | did the drug discovery, right?

00:13:12.860 | They didn't do it, correct?

00:13:15.740 | Well, they funded it.

00:13:16.580 | - I think they handed over some research

00:13:18.720 | that was open to the public at the time,

00:13:20.300 | but they funded an independent for-profit.

00:13:23.180 | - And so what happened here was you had the IP

00:13:26.220 | and the employees of the nonprofit gifted,

00:13:29.680 | they were transferred over to the for-profit organization.

00:13:33.460 | And when the Mozilla Foundation did this

00:13:36.260 | and another company called Samasource,

00:13:38.900 | and these both have two structures,

00:13:43.060 | a nonprofit and a for-profit,

00:13:44.720 | but in both of their cases, they set up separate boards.

00:13:47.940 | And the two different boards,

00:13:49.860 | the answer to the for-profit concerns

00:13:51.740 | and the nonprofit concerns,

00:13:53.120 | they also didn't give huge chunks of equity

00:13:56.060 | to all of the employees, which then sold them in secondary.

00:14:00.180 | So what's happened here is all of the IP in all likelihood

00:14:04.220 | has been, and the wealth transfer

00:14:05.940 | has gone to all these employees.

00:14:07.660 | They've now done the secondary.

00:14:09.440 | And this actually triggers a lot of IRS investigations.

00:14:14.140 | The Mozilla Foundation, which was making the Netscape

00:14:17.060 | browser, which many of you probably have used

00:14:18.700 | for the years, they were making hundreds of millions

00:14:20.820 | of dollars a year, David, from advertising with Google.

00:14:23.780 | The IRS said, "Hey, wait a second,

00:14:24.900 | "you look like a for-profit to us."

00:14:26.700 | So they made a for-profit entity.

00:14:29.060 | All of the, and Mitchell Baker and the team over there

00:14:33.740 | set this up so all the money went back

00:14:35.700 | into the Mozilla Foundation in their nonprofit efforts.

00:14:40.460 | And they just paid above market salaries.

00:14:44.540 | But it was very important that these things be separated

00:14:47.160 | and that the employees not have equity

00:14:48.780 | and that all that money flowed back up.

00:14:50.580 | And I think this is where OpenAI

00:14:51.700 | is gonna get super tripped up.

00:14:53.320 | And I think the IRS is gonna have a field day here.

00:14:55.780 | - Here's the, Nick, can you please throw the image up?

00:14:57.820 | This is from the OpenAI website

00:15:00.380 | that explains their structure.

00:15:02.140 | So let's just speculate for a second

00:15:04.520 | 'cause we're just guessing.

00:15:05.940 | But it has been reported that Microsoft owns 49%, right?

00:15:10.940 | That's the number?

00:15:12.480 | - Of the for-profit.

00:15:14.180 | In this example, Microsoft, where it says minority

00:15:17.280 | economic interest, that number there would be 49

00:15:20.000 | and that's what they own of this OpenAI Global LLC.

00:15:24.400 | And the majority owner is this holding company.

00:15:27.040 | Okay, so that means 51% is there.

00:15:29.660 | But it says here that that 51% is owned

00:15:32.400 | by a combination of the nonprofit,

00:15:35.180 | employees and investors.

00:15:37.540 | So the question is, back to Sax's question,

00:15:40.220 | we can know very precisely how much the nonprofit owns

00:15:44.880 | by just Xing out the investors and employees.

00:15:48.900 | And I'm gonna guess it's at least 30 or 40%.

00:15:52.900 | I think that that's probably a pretty reasonable guess.

00:15:55.700 | So it means that the OpenAI Foundation

00:15:57.720 | probably owns somewhere between five to 20%.

00:16:01.940 | Is that fairly reasonable guess at the maximum?

00:16:06.420 | - That seems right, yeah.

00:16:08.220 | And so I think that they're gonna have to show

00:16:10.720 | that ownership structure and decompose all these entities

00:16:14.040 | and show the actual percentages.

00:16:15.860 | If it's a lot less than that, if it's say like a 1% thing,

00:16:21.180 | then it's a little bit more sketch, I think.

00:16:23.220 | But if it's really like 15 or 20%,

00:16:26.480 | they're still gonna have to prove that all of this was done

00:16:29.960 | in a clean way because the big thing that these guys did

00:16:33.940 | was because by creating this LPGP structure

00:16:36.460 | and these cap profits,

00:16:38.660 | it's clear that they were trying to avoid something.

00:16:41.300 | And the question is in the discovery,

00:16:43.980 | will everybody learn what it is

00:16:46.540 | that they were trying to avoid?

00:16:47.660 | Because I've been through this,

00:16:49.500 | I think Sax, you've been through it,

00:16:50.780 | Freeburg, you were through it.

00:16:52.220 | When you set up these large vehicles

00:16:55.420 | and pooled amounts of money,

00:16:57.100 | folks come out of the woodwork

00:16:58.340 | with all kinds of convoluted structures.

00:17:00.380 | And the discipline is always to tell these lawyers

00:17:02.580 | and accountants, no.

00:17:04.820 | And they'll be like, let's set up a master feeder

00:17:06.740 | and you'll go through Bermuda and this and that.

00:17:09.980 | And it's so easy to say yes because it's very seductive

00:17:12.860 | what they're selling you,

00:17:14.380 | but it's always about trying to avoid something.

00:17:16.320 | So the real question is why this convoluted structure,

00:17:19.480 | what was it trying to avoid?

00:17:21.260 | - And where did this IP come from?

00:17:23.700 | Like all that IP, Sax, came from a nonprofit.

00:17:26.900 | All those employees worked for the nonprofit.

00:17:29.260 | Then they decided to go from an open source model to closed.

00:17:32.380 | And then by doing that, they capture all the value

00:17:35.940 | and then it's gifted to these employees

00:17:38.760 | and this investor base.

00:17:39.860 | That's where I think something doesn't seem right.

00:17:43.380 | - Well, it's important to remember

00:17:44.380 | the majority of the IP that exists at OpenAI today

00:17:48.220 | was generated after the for-profit was set up.

00:17:50.660 | The real question, was there a fair transfer at the time

00:17:53.340 | when it was set up of value into this for-profit entity?

00:17:56.680 | And my guess is that GP box that you just looked at,

00:17:59.800 | the compensation that they're earning through that GP box

00:18:02.460 | that the nonprofit earns,

00:18:04.080 | is likely meant to be kind of a fair amount

00:18:08.100 | relative to what was contributed at the time.

00:18:10.280 | The real question for me still remains,

00:18:12.420 | what other nonprofit and charitable work

00:18:15.480 | does the OpenAI 501(c)(3) actually do?

00:18:18.740 | And if the answer is nothing

00:18:20.600 | and it's just a shell nonprofit

00:18:23.000 | and under it the only thing that it owns

00:18:24.560 | is a for-profit interest,

00:18:26.340 | then I think there's a real question on what's the activity.

00:18:28.820 | And I don't mean to speak out of turn on all this stuff.

00:18:30.900 | I don't know enough,

00:18:32.760 | but I think that's the only sort of lens

00:18:34.340 | I would kind of look at this stuff through.

00:18:35.940 | - Sax, you wanna say something?

00:18:37.100 | - Well, a couple of points.

00:18:38.200 | One is, one of my contentions has been

00:18:41.220 | never innovate on structure.

00:18:43.060 | All you do is create legal problems.

00:18:44.960 | There's a tried and true way of creating a startup,

00:18:48.160 | which is a C corp.

00:18:49.860 | When you try to innovate on legal instead of product,

00:18:53.060 | it almost always backfires.

00:18:55.560 | The second point here is,

00:18:57.180 | I think OpenAI is in a little bit of a

00:18:59.620 | damned if you do, damned if you don't situation.

00:19:02.220 | To go back to Freeberg's point about the nonprofits

00:19:05.500 | sitting at the top there,

00:19:07.000 | remember what created all the problems

00:19:08.980 | when Sam briefly got fired as CEO and then came back,

00:19:12.560 | is they had these directors from the nonprofit world

00:19:16.340 | who didn't really seem to understand how startups work.

00:19:19.140 | And so either they fired Sam for no reason,

00:19:23.260 | which was kind of incompetent,

00:19:24.620 | or they had a really good reason

00:19:26.220 | but didn't communicate it properly,

00:19:27.920 | which was also incompetent.

00:19:29.340 | Either way, that whole thing was a spectacle.

00:19:32.740 | And it just kind of showed

00:19:34.180 | that there was this culture clash going on

00:19:36.580 | between the for-profit company

00:19:39.140 | and the standard way of running

00:19:41.220 | a Silicon Valley startup to maximize the outcome,

00:19:43.900 | and then this nonprofit board

00:19:45.460 | that was sitting on top of the whole thing.

00:19:47.560 | So I think that now they're in the situation

00:19:50.900 | where they've changed that nonprofit board,

00:19:54.680 | they've booted off the nonprofit people,

00:19:59.160 | and that may have been the right thing

00:20:01.140 | for the for-profit entity,

00:20:03.260 | but now it might get them in trouble

00:20:05.100 | because it lends credence to Elon's lawsuit

00:20:08.900 | that they've completely changed the original mission

00:20:12.240 | of this organization.

00:20:14.760 | Right, it was supposed to be-- - Yeah, and the mission.

00:20:16.100 | - It was supposed to be open source,

00:20:18.220 | and it was supposed to be nonprofit,

00:20:20.700 | and now it's for-profit and closed source.

00:20:22.580 | - Think about this, and again,

00:20:24.140 | we are speculating here,

00:20:25.500 | but if the mission was to be open source,

00:20:28.880 | and then you realize you've got something super valuable,

00:20:31.720 | and you close source it, create a for-profit,

00:20:34.740 | and then take all the employees and all the IP

00:20:37.380 | and put it into the for-profit, lock it down,

00:20:39.940 | break the original mission that,

00:20:41.220 | hey, this was gonna help humanity

00:20:42.920 | because we're gonna make it open source.

00:20:44.600 | So if you donate money here and we get the tax exemption,

00:20:47.560 | humanity benefits because everybody gets

00:20:50.020 | to look at that code at the same time

00:20:51.600 | and work on it and iterate.

00:20:53.400 | Now they've just basically stolen all that

00:20:56.400 | in order to enrich themselves.

00:20:57.820 | And I think then you add to it,

00:20:59.280 | Sam maybe doing some deal-making,

00:21:01.360 | which seems to be one of the triggers according to reports,

00:21:03.660 | and this OpenAI Venture Fund, et cetera,

00:21:06.480 | maybe that deal-making made the nonprofit people say,

00:21:08.520 | hey, listen, you're doing even more for-profit stuff

00:21:10.940 | with the OpenAI name.

00:21:12.360 | We're supposed to be a nonprofit here.

00:21:13.900 | - What's the venture fund thing?

00:21:15.260 | - He started a venture fund to invest in companies

00:21:17.740 | called the OpenAI Ventures Fund,

00:21:19.580 | and he was the sole owner of it,

00:21:21.540 | which they're saying now it's a clerical mistake

00:21:23.800 | or something, but he's invested in a bunch of startups

00:21:26.720 | that have unique access to the OpenAI,

00:21:30.200 | you know, I think infrastructure.

00:21:32.880 | - He's the GP on behalf of OpenAI.

00:21:34.500 | - Yes, the OpenAI.fund, yes.

00:21:36.780 | - No, but he's the GP on behalf of himself

00:21:38.820 | or on behalf of OpenAI.

00:21:40.160 | - The way the paperwork was done,

00:21:41.480 | it was on behalf of himself.

00:21:42.720 | They're now saying that that was done in error.

00:21:45.640 | So, okay, fine, maybe it was, maybe it wasn't, who knows?

00:21:47.880 | - Whoa, whoa, whoa, wait, the OpenAI fund,

00:21:51.500 | yes.

00:21:52.820 | - Whose money is that?

00:21:54.100 | - Sam raised the money to create an OpenAI fund

00:21:58.620 | to invest in companies using OpenAI software.

00:22:02.160 | - Okay, so that's a different kind of issue.

00:22:04.820 | That's called a corporate opportunity issue

00:22:08.060 | where OpenAI's fund-- - Self-dealing, self-dealing.

00:22:11.140 | - It's basically a conflict of interest

00:22:13.700 | where the economic opportunity

00:22:16.180 | of the OpenAI Startup Fund should belong to OpenAI

00:22:18.580 | if Sam created a separate fund with separate LPs

00:22:22.080 | that he's the GP of and gets economics in that,

00:22:24.760 | then that's a potential usurpation

00:22:26.380 | of a corporate opportunity.

00:22:27.640 | I don't know the truth of that matter.

00:22:30.760 | I'm just saying that's a slightly different issue.

00:22:33.520 | To go back to the point-- - And it uses the OpenAI name.

00:22:35.900 | - Yeah, that's where the corporate opportunity

00:22:38.400 | becomes really explicit.

00:22:40.760 | To go back to the point you were making a minute ago

00:22:43.200 | about the employees, I just think,

00:22:45.240 | I mean, you used the word stealing.

00:22:46.580 | I think that's a really strong word.

00:22:48.080 | I mean, I think we have to complement the employees

00:22:51.940 | of this company, if it is a company,

00:22:54.780 | including Sam, including Greg, including Ilya,

00:22:57.480 | for creating an amazing product, right?

00:23:00.460 | - Something incredible.

00:23:01.300 | - They've created something incredible.

00:23:02.140 | - Sure, absolutely.

00:23:02.960 | - And they've created, I think, an amazing ecosystem,

00:23:05.860 | and there's a lot of developers building on top of this,

00:23:08.160 | and I think no one is attacking their work.

00:23:11.660 | And I think it's a little too harsh to say

00:23:13.280 | that anyone's stealing anything

00:23:14.300 | 'cause they have created all the value.

00:23:16.480 | The question, and I think you have to include Sam in that

00:23:18.820 | and say that he's done a great job, okay, as CEO,

00:23:21.740 | because they've built something amazing.

00:23:23.260 | - And, Sax, to your point, it's not stealing

00:23:25.400 | because very smart investors

00:23:27.100 | are putting in money eyes wide open.

00:23:29.740 | - Yeah, I think that the criticism here

00:23:32.160 | is related to structure, from my standpoint.

00:23:34.820 | It's related to structure,

00:23:36.260 | and whether Elon was told something at the beginning

00:23:39.620 | that's different, it got changed,

00:23:41.700 | and clearly it got changed in a way

00:23:44.440 | that was not consistent with the terms

00:23:47.240 | under which he initially contributed all this money.

00:23:50.560 | - Yeah, mine is, I'm stating what is

00:23:55.220 | the most cynical interpretation of what happened here.

00:23:57.820 | There is the most benign and benevolent interpretation

00:24:01.820 | you could have as well,

00:24:02.660 | and maybe the truth is in between the two,

00:24:04.540 | but the series of events that occurred

00:24:06.780 | does not look good when you have a mission

00:24:08.840 | to give this intellectual property to the world.

00:24:12.300 | So, no one person benefits from it.

00:24:14.540 | All of humanity is supposed to benefit from it.

00:24:16.640 | That was the point of being open source.

00:24:20.120 | - Right. - Then you close it.

00:24:21.720 | Now, who gets the benefit if it's closed?

00:24:23.940 | - Right, that's a good point. - The employees.

00:24:25.980 | - Right, that's a good point.

00:24:26.800 | - And that's gonna be the crux of this,

00:24:28.240 | and if you really wanna-- - Or the investors.

00:24:30.800 | The investors. - The investors.

00:24:32.200 | - But that's how they were able

00:24:33.220 | to attract billions of dollars,

00:24:34.500 | 'cause you can't get billions of dollars

00:24:36.640 | to invest in something that's gonna be open sourced,

00:24:38.640 | and there's no return. - And then the employees

00:24:40.720 | sell two billion in secondary.

00:24:42.460 | So now you, if you really wanna take

00:24:45.000 | the most cynical approach to this or interpretation,

00:24:48.300 | and again, this is just one interpretation,

00:24:50.840 | they took an open source project, they closed it,

00:24:54.320 | they raised money, and then within the next two years

00:24:57.520 | on this incredible innovation,

00:24:59.500 | they sold $2 billion and put that in their pockets.

00:25:02.720 | Now, those employees, if it was a for-profit,

00:25:05.220 | obviously no problem with that,

00:25:07.280 | but they took the stated mission

00:25:09.540 | of giving this to all of humanity,

00:25:11.500 | and then took 3% or 4% of that gain

00:25:13.760 | and put it in their pockets.

00:25:14.640 | That's where I think this whole thing

00:25:16.240 | is gonna get really crazy,

00:25:18.120 | and this is where the IRS-- - I think that's a good point.

00:25:20.060 | - Yeah, and the IRS also interpreted Mozilla,

00:25:24.080 | which was just $100 million,

00:25:25.680 | and none of the employees took any of it.

00:25:27.640 | The 100 million that came out of Mozilla,

00:25:30.100 | they were just saying,

00:25:30.940 | "I think you should pay tax on that.

00:25:32.720 | "Should you pay tax on that?"

00:25:33.780 | And then they were investigated for years.

00:25:35.940 | So I think the IRS is gonna be on this like crazy

00:25:38.900 | based upon what happened to Mozilla, which was--

00:25:41.360 | - Jason, do you think individual employee sellers

00:25:45.020 | are gonna get audited by the IRS because of this?

00:25:48.400 | - I don't, I mean, we're in uncharted territory here.

00:25:51.380 | When I did some tweets and said,

00:25:52.720 | "Hey, can everybody contact me

00:25:53.800 | "who has information on these structures?"

00:25:55.780 | I was able to find the Samasource example,

00:25:59.520 | and they did it very clean,

00:26:01.080 | and there was no IP transfer

00:26:03.760 | or employees enriching themselves,

00:26:06.440 | or a God King like Sam doing all kinds of deals

00:26:09.480 | and enriching himself.

00:26:11.480 | It was a very clean structure.

00:26:12.500 | And then Mitchell Baker who set up Mozilla

00:26:14.200 | also did the same thing.

00:26:15.220 | Hey, we'll just pay the employees an extra 50K

00:26:17.240 | on their salaries,

00:26:18.560 | and they'll be a little overpaid in Silicon Valley,

00:26:20.260 | but they're not getting billions of dollars in equity,

00:26:22.920 | which again, I have no problem

00:26:24.580 | with people getting billions of dollars in equity

00:26:26.760 | in a for-profit,

00:26:27.860 | but this was supposed to benefit humanity.

00:26:30.940 | It was supposed to be open.

00:26:31.960 | If this was open source and they took the two billion,

00:26:34.520 | I would actually not have much of a problem with it

00:26:36.280 | because we could all be looking at that mission

00:26:38.960 | and then Saks and myself and Friedberg

00:26:41.000 | and anybody who wanted to use that code

00:26:42.840 | could go in there and adapt it.

00:26:44.680 | And you know what?

00:26:45.520 | There are for-profits, Apple and Meta,

00:26:48.640 | which are producing open source software.

00:26:50.640 | So if they're producing, Apple's pretty competent

00:26:53.080 | and Meta's super competent in this respect.

00:26:55.860 | They're able to do open source,

00:26:57.520 | but Sam is claiming or OpenAI is claiming

00:27:00.340 | it's too dangerous to show us the open source code.

00:27:02.480 | I call bull (beep) on that.

00:27:04.320 | Their code is not too dangerous for us to sit.

00:27:06.720 | That's complete utter bull (beep)

00:27:08.120 | - So you think that there is a motivation

00:27:10.200 | for profit, basically.

00:27:11.220 | What you're saying is essentially

00:27:12.640 | that this was highly motivated by profit.

00:27:15.140 | - I think. - Versus like even,

00:27:18.120 | like some of the emails,

00:27:20.240 | the emails try to paint a picture

00:27:23.480 | of very industrious people on both sides

00:27:27.400 | who are trying to solve a very hard problem

00:27:29.840 | and dealing with a conundrum

00:27:31.300 | that's really around capital, right?

00:27:33.700 | - Yeah.

00:27:34.540 | - We need to do all this training,

00:27:35.740 | it's gonna be so expensive.

00:27:37.520 | But I think what you're saying,

00:27:38.500 | which is a deviation from that,

00:27:40.420 | is actually, come on guys,

00:27:41.380 | these are all very smart people

00:27:42.920 | and those folks saw an opportunity

00:27:44.560 | to make a ton of money and they took it.

00:27:47.300 | - I mean, that seems like what happened here to me.

00:27:50.120 | I mean, it's Occam's Razor kind of situation.

00:27:52.420 | I think they, if you look at it,

00:27:54.700 | I think they probably regretted making this a nonprofit

00:27:57.340 | and then tried to figure out a way to reverse it.

00:27:59.200 | That's actually what I think is going on here.

00:28:00.920 | And I do think there is part of it, Chamath, you're right,

00:28:02.880 | that they needed servers and they needed capacity,

00:28:05.820 | but they could have done that

00:28:07.560 | without giving the employees tons of equity,

00:28:09.440 | without selling billions of dollars,

00:28:11.020 | keeping it open source and then telling Microsoft,

00:28:14.460 | hey, if you want access to this open source or whatever,

00:28:17.140 | and they could give them-

00:28:19.580 | - You're making a very, very good point.

00:28:21.660 | And the reason is that if it was open source,

00:28:26.420 | Microsoft could have just taken it

00:28:28.420 | and done its own training and just paid for it.

00:28:31.100 | - Yeah.

00:28:32.060 | - Or make a donation to the foundation.

00:28:34.060 | - This is interesting 'cause-

00:28:36.000 | - It's a very good point, Jacob.

00:28:36.840 | - Because OpenAI, yeah, OpenAI dropped a few of the emails

00:28:41.080 | that they had with Elon.

00:28:42.200 | Obviously, those are cherry picked

00:28:43.460 | to help their case the most.

00:28:45.180 | What we don't know is what are the other emails show?

00:28:47.960 | Like, what were their alternatives

00:28:49.320 | to this particular structure?

00:28:50.740 | Could they have raised the necessary funds

00:28:52.040 | in a way that was more consistent

00:28:53.620 | with the original mission of the company?

00:28:55.480 | Could they have remained open source, for example?

00:28:59.020 | Also, are there any emails where employees talk

00:29:02.740 | about the potential benefit to them of going-

00:29:06.900 | - Private.

00:29:07.740 | - Going for profit, right?

00:29:09.260 | - Those emails have already been leaked.

00:29:10.940 | There was this thing where,

00:29:12.860 | in the OpenAI emails that were leaked,

00:29:15.020 | they talked about being able to swap this phantom equity

00:29:18.700 | in OpenAI into either YC equity,

00:29:21.100 | and then Elon said, "Oh, maybe SpaceX too,

00:29:23.140 | "but I'll have to figure that out."

00:29:24.540 | - Well, then also like Microsoft coming in and buying 49%,

00:29:28.700 | and then it's closed.

00:29:29.540 | This is the other piece that I would wanna see

00:29:31.700 | in discovery as well.

00:29:32.580 | And again, I don't have any problem with any of the people.

00:29:34.900 | I love Sam Allman, I think he's great.

00:29:36.740 | I think all the people there are great.

00:29:37.740 | I think what they've done for humanity is great.

00:29:39.700 | I just think they should have kept this thing open source,

00:29:41.640 | which was the mission.

00:29:43.060 | But then they closed source at Chamath,

00:29:45.900 | and then give 49% of it, all the weights,

00:29:48.460 | all the source code to Microsoft.

00:29:50.900 | So that, to me, was like a really...

00:29:53.420 | Like, you wanna talk about taking this nonprofit's IP,

00:29:57.560 | and then some amount of that bag

00:30:00.040 | gets given to the employees for billions of dollars,

00:30:01.900 | and then Microsoft gets 49% of all that nonprofit's effort

00:30:05.340 | to then go commercialize,

00:30:06.900 | and Microsoft has added, what, $500 billion in market cap

00:30:10.240 | since this whole thing has been announced?

00:30:12.880 | So now all that profit has been aggregated

00:30:15.680 | into Microsoft stock.

00:30:16.940 | - And I think this is where you gotta understand

00:30:18.860 | why Elon feels swindled, is because not only are we going

00:30:22.840 | from nonprofit to for-profit and open source

00:30:25.820 | to closed source, he was specifically concerned

00:30:28.060 | about all the benefits of AI crewing

00:30:31.160 | to one powerful, big tech company.

00:30:33.400 | Now, at that time, he thought it was Google.

00:30:35.400 | Now it's Microsoft.

00:30:36.700 | There's not really that much of a difference.

00:30:38.580 | He never wanted all the benefits of AI

00:30:41.740 | in the hands of one really powerful tech company.

00:30:45.780 | And Microsoft is the most, it's now, what,

00:30:48.520 | the biggest company in the world by market cap?

00:30:50.300 | So this is like the opposite of what he intended.

00:30:52.620 | - They gifted a trillion dollars to Microsoft, probably.

00:30:54.800 | There's an easy solution.

00:30:56.420 | If they are people of good faith and they're doing this

00:30:59.460 | for the right reasons, open source it.

00:31:02.160 | Just go back and open source it.

00:31:04.020 | - Jason, but to your earlier point,

00:31:05.620 | that may solve the lawsuit, and Elon may drop the lawsuit,

00:31:08.300 | but it's opened a can of worms with respect to tax

00:31:11.800 | and structuring that's much bigger than just open AI.

00:31:14.820 | - Yeah, I don't think it solves that.

00:31:15.660 | - It's the entire nonprofit industry.

00:31:16.820 | It's a tax problem for every new company.

00:31:19.300 | It's every other entrepreneur that studies this model

00:31:21.700 | and tries to replicate it for their own personal gain,

00:31:24.380 | even if that wasn't intentional here.

00:31:26.760 | So there's a whole set of issues

00:31:29.240 | that we've really cracked the egg here.

00:31:32.560 | We gotta figure out how to unscramble it.

00:31:34.600 | - There's the IRS issues, then there's what's morally right,

00:31:37.340 | and then there's Elon's beef.

00:31:38.540 | And if it is a for-profit company and Elon put in 50 million

00:31:41.180 | when it was a seed round, what would he own, Sachs?

00:31:44.900 | What would his ownership in this for-profit be?

00:31:47.540 | - Oh my God, I mean, probably--

00:31:49.300 | - 98%.

00:31:50.820 | - Okay, and so--

00:31:52.260 | - Yeah, we need to know the total size of that round.

00:31:54.360 | He put in the first 40 million.

00:31:55.580 | Do we know what other people put in?

00:31:57.240 | It was most of the money, right?

00:31:58.640 | - Yeah, it was most of the money.

00:31:59.480 | So, I mean, if you put,

00:32:00.860 | let's just put a crazy valuation on it, 500 million.

00:32:04.220 | Okay, he owns whatever, 10%--

00:32:05.620 | - No, no, at that point, it would,

00:32:09.460 | AI was not what it is today.

00:32:11.060 | - Of course not.

00:32:11.900 | - So you would've raised that like,

00:32:13.060 | you would've raised 150 on 50 pre.

00:32:15.980 | - Okay.

00:32:17.220 | - That's the way that a very capex-intensive deal

00:32:19.540 | in a space that wasn't thought to yield big outcomes,

00:32:22.860 | that's unfortunately the cost of capital.

00:32:23.700 | - Elon would own half the company, at least.

00:32:25.900 | - At least.

00:32:26.740 | - All right, so give him 20% of the company,

00:32:28.060 | it's $20 billion since even half

00:32:30.120 | of what you're saying, Zach, right?

00:32:31.260 | - We talked about on the show when that whole fracas

00:32:34.180 | with these non-profit directors went down,

00:32:37.060 | is we said, go back and restructure the whole thing

00:32:40.660 | to make it what it always should've been,

00:32:42.620 | which is just a clean for-profit entity,

00:32:45.060 | give Elon his equity, give Sam his equity,

00:32:48.180 | 'cause it never made sense that Sam had no equity either.

00:32:49.900 | - Sam has no equity, but he's got a venture firm,

00:32:52.120 | it's all so convoluted.

00:32:53.160 | - It's a weird form of compensation in which

00:32:55.980 | they're giving him corporate opportunities

00:32:59.440 | in effect as like a type of compensation,

00:33:01.800 | when really he should just have compensation

00:33:03.720 | in the corporation, and then the corporation

00:33:06.420 | should own all of its opportunities.

00:33:08.680 | - Okay, that's a really interesting thing you just said.

00:33:10.700 | So basically, yeah, it's like he famously has no equity,

00:33:14.560 | but then he has this retained optionality

00:33:16.800 | to monetize the ecosystem.

00:33:18.440 | So even though he's not monetizing the thing,

00:33:20.720 | he gets to monetize everything around him.

00:33:23.400 | - And it seems like the board,

00:33:24.980 | or at least the new board's okay with that,

00:33:26.580 | because he is being undercompensated

00:33:28.860 | with respect to the main thing.

00:33:30.440 | So then he gets the side things,

00:33:31.820 | but that's not really the way it should work either.

00:33:34.280 | - You know, Bill Gates said to us famously

00:33:35.940 | when we were at, when I was at Facebook,

00:33:38.280 | the value of an ecosystem is when the economic value

00:33:41.620 | generated by the ecosystem exceeds that of the platform.

00:33:44.820 | Now, in this case, you'd actually rather have

00:33:48.340 | 50 basis points of the ecosystem than 5% of open AI,

00:33:51.580 | if that's true, 'cause if this thing

00:33:52.900 | could be so revolutionary, you're talking

00:33:55.340 | 10, 20, 30 trillion dollars.

00:33:57.060 | - What a mess, oy oy oy.

00:34:00.460 | - Sam should just be given like a huge option grant

00:34:03.380 | in open AI, but then open AI should own

00:34:05.900 | its own venture fund.

00:34:07.000 | - Yeah, and the SEC is looking into all this stuff.

00:34:11.020 | You know, they look into a lot of things in fairness,

00:34:12.860 | but they're looking into it, what a mess,

00:34:14.700 | and we'll keep track of it.

00:34:16.420 | The other thing that's crazy, I don't know

00:34:18.580 | if you guys know this, but Nick, just as we close here,

00:34:21.960 | the most insane part of open AI's LP investment agreement,

00:34:24.660 | which is on their website, you can just search

00:34:25.980 | for open AI LP agreement, is this part.

00:34:29.180 | The partnership exists to advance open AI Inc.'s mission

00:34:31.740 | of ensuring that safe artificial and general intentions

00:34:33.900 | is deployed and benefits all of humanity.

00:34:35.860 | The general partners' duty to this mission

00:34:37.780 | and principles advance in open AI's ink chart,

00:34:39.900 | yada yada yada, take precedent over the obligation

00:34:42.220 | to generate a profit.

00:34:43.980 | The partnership may never make a profit

00:34:45.180 | and the general partner is under no obligation to do so.

00:34:46.940 | The general partner is free to reinvest any

00:34:49.300 | and all of the operating entities' cash flow

00:34:51.660 | into research and development activities

00:34:53.200 | and/or related expenses without any obligation

00:34:55.760 | to the limited partners.

00:34:57.140 | And so they basically told everybody,

00:35:00.140 | Vinod and employees or whatever,

00:35:01.640 | we can just basically wipe your equity out

00:35:04.220 | and we can do whatever we want with the profits

00:35:06.020 | and you're probably gonna lose your money.

00:35:07.380 | So this structure is weird.

00:35:10.340 | - Do you guys think that when the investors came in,

00:35:13.580 | especially in this latest round,

00:35:15.100 | the $86 billion round,

00:35:16.560 | do you think they underwrote the legal risk

00:35:19.300 | inherent in the structure?

00:35:20.920 | Or are they just sort of hand-waved over it and said--

00:35:24.500 | - No, typically what happens in these deals

00:35:27.500 | is you hire somebody like KPMG or Deloitte

00:35:32.500 | or Ernst & Young to do the full financial diligence packet.

00:35:35.820 | And that typically tends to be how a lot

00:35:38.980 | of late-stage organizations document that they've done

00:35:43.080 | and manage their fiduciary responsibilities

00:35:45.240 | on behalf of their limited partners.

00:35:47.060 | So what happens is you will do a deal,

00:35:49.160 | you'll sign a term sheet, you'll turn it over to Deloitte,

00:35:52.120 | you'll turn it over to KPMG and say,

00:35:54.500 | "Please go and run this down."

00:35:56.740 | And then what they do is they will furnish a report

00:35:58.740 | that says, "Yes, this meets all the customary expectations."

00:36:01.960 | I suspect that if these folks were doing a decent job

00:36:06.960 | of running late-stage money,

00:36:09.920 | they probably sent it to those folks.

00:36:12.380 | And those folks probably produced something

00:36:14.420 | that said, "This looks fine."

00:36:15.820 | Now, if you read Buffett's letter this year,

00:36:18.380 | he has a really great commentary on Deloitte and KPMG

00:36:22.760 | and these sorts of letters,

00:36:24.740 | which is not exactly the most supportive

00:36:28.360 | is the best way to say it.

00:36:29.400 | Nick, you can find that mentioned.

00:36:31.380 | In his case, it's Deloitte and Touche.

00:36:33.000 | But typically, SACS, that's what they do.

00:36:34.360 | They go to a KPMG or a Deloitte or a Ernst & Young and say...

00:36:37.660 | And by the way, 99% of the time,

00:36:41.840 | I mean, they do really good work,

00:36:43.360 | but it's a standard structure.

00:36:45.220 | And they just wanna make sure

00:36:48.160 | that nothing nefarious is a muck.

00:36:49.760 | In this case, I doubt anything was a muck anyways.

00:36:52.160 | I doubt though that they looked at the structure

00:36:54.660 | and then also elevated the litigation risk

00:36:57.660 | of this getting unwound.

00:36:58.800 | I just don't see that that diligence report,

00:37:01.240 | and I've seen enough of them,

00:37:02.640 | typically has that section in it.

00:37:06.840 | - The other thing that's completely hypocritical here

00:37:08.660 | is they said when they hit AGI

00:37:11.080 | and they're gonna be like a sentient,

00:37:13.520 | artificial intelligent going on here,

00:37:16.060 | Friedberg, that they would wrap up shop

00:37:18.400 | and they're gonna no longer be a nonprofit, et cetera.

00:37:21.760 | But they're claiming they haven't hit that,

00:37:24.200 | but they closed the software.

00:37:26.960 | So it should be open source if they haven't hit AGI

00:37:29.560 | and you don't think they've hit general intelligence, right?

00:37:34.320 | Friedberg, for anything close to it.

00:37:36.600 | Maybe you could educate the audience on what that is

00:37:39.080 | and that claim that they have to-

00:37:40.840 | - No, I don't think they have. - Should have shut off

00:37:41.680 | the for-profit.

00:37:42.840 | Yeah, they haven't.

00:37:43.800 | - But I think we keep repeating this concept

00:37:48.400 | of the models should be open source versus closed source.

00:37:52.000 | Making AI for the benefit of humanity

00:37:54.000 | can be interpreted in a lot of ways.

00:37:56.020 | There may have been some anecdotal conversation

00:37:58.240 | at some point with Elon or others

00:38:00.620 | about we're gonna make the models open source.

00:38:03.020 | But there was a reason that that change was made

00:38:07.140 | along the way, which was to attract dollars.

00:38:10.040 | And those dollars need to have some return of capital

00:38:15.040 | available to them because they're private investor dollars.

00:38:18.720 | And so I don't think that that was necessarily,

00:38:20.880 | and correct me if I'm wrong,

00:38:21.720 | I don't think that's in the mission,

00:38:23.020 | that the open AI software models will be open source.

00:38:26.280 | Making AI for the benefit of humanity

00:38:28.920 | could probably be interpreted in a lot of different ways

00:38:31.040 | and we'll see.

00:38:33.400 | But no, I don't think that anyone has achieved

00:38:36.800 | this holy grail of general intelligence.

00:38:41.800 | - Yeah.

00:38:43.840 | - I think one of the more interesting

00:38:46.360 | and kind of wacky things about open AI

00:38:49.320 | is that their mission is explicitly to create EGI,

00:38:52.800 | which most people would associate

00:38:56.280 | with some sort of sci-fi dystopian outcome.

00:38:59.480 | And I think this has raised the fear factor around AI

00:39:04.480 | because they're explicitly trying to create

00:39:06.500 | the sentence that's gonna replace humanity.

00:39:09.820 | Now, I think they defined AGI in a different way.

00:39:12.240 | They say it's something that can replace 80% of the jobs,

00:39:15.960 | but I think we all kind of know what it really is.

00:39:18.980 | So I just wonder if--

00:39:20.240 | - I don't know, I think that that's like,

00:39:22.780 | that assumes a steady state in the world.

00:39:24.920 | So if you end up with a system

00:39:28.880 | and the system has all the capabilities

00:39:33.020 | of a bunch of really highly qualified knowledge workers,

00:39:36.560 | and I can sit in front of a computer terminal

00:39:39.200 | and I can say, "Let's design a mission to Mars."

00:39:41.800 | A mission to Mars could be a 20-year engineering project

00:39:45.120 | with hundreds of people involved to design the buildings,

00:39:47.520 | to design the flight path, to figure out the fuel needs,

00:39:49.840 | to figure out how you would then be able to terraform Mars.

00:39:52.920 | And what if one person could interact with a computer

00:39:55.680 | and design a plan to go and inhabit Mars?

00:39:58.960 | All of the technical detail docs could be produced,

00:40:01.400 | all of the engineering specifications could be generated,

00:40:04.260 | all of the operating plans, all of the dates,

00:40:06.620 | the amount of labor needed,

00:40:07.820 | the amount of production needed,

00:40:08.860 | the amount of capital needed.

00:40:10.220 | What would otherwise take NASA or some international

00:40:13.900 | or well-funded private company many, many decades to do,

00:40:17.620 | a piece of software could do in a very short order.

00:40:20.260 | I think that's like a really,

00:40:22.140 | like for me, poignant example of the potential

00:40:25.140 | of having these tools broadly available,

00:40:27.860 | that the potential of humanity

00:40:29.660 | starts to become much broader.

00:40:31.060 | We could say, I want to develop a city underneath the ocean

00:40:34.820 | because I want to explore more of the earth.

00:40:36.680 | I think humans need to go solve cancer,

00:40:39.840 | figure out the biologic drugs

00:40:41.200 | and the combination of biologic drugs

00:40:42.600 | that would be needed to solve cancer

00:40:44.120 | based on this patient's genotype.

00:40:46.940 | The extensibility of highly knowledgeable

00:40:50.320 | or what other people might call

00:40:51.380 | general intelligence type tooling is extraordinary.

00:40:54.100 | That one individual starts to have an entire cohort

00:40:59.240 | of knowledge workers available at their disposal

00:41:01.580 | to do things that we can't even imagine today.

00:41:04.580 | So I don't think that it is nefarious.

00:41:06.140 | It's nefarious because we assume

00:41:07.400 | a steady state of the world today, that nothing changes.

00:41:09.640 | Therefore, a piece of software replaces all of us.

00:41:12.280 | But the potential of humanity starts to stretch

00:41:14.360 | into a new era that we're not really comfortable with

00:41:17.200 | because we don't really know it or understand it yet.

00:41:19.440 | - I'm not saying it's nefarious to want to develop AI

00:41:22.960 | because I agree with you

00:41:24.720 | about all the extraordinary potential of it.

00:41:26.940 | I'm saying there's something a little bit cultish

00:41:29.040 | and weird about explicitly devoting yourself to AGI,

00:41:33.200 | which I think in common parlance means Skynet.

00:41:37.340 | - Yeah, it means something sentient smarter than humans.

00:41:41.740 | - Maybe that parlance is what needs to be addressed,

00:41:44.160 | which is AGI effectively enables equivalence

00:41:47.200 | to a human knowledge worker.

00:41:49.300 | And that can kind of unleash a new kind of

00:41:52.640 | set of opportunities.

00:41:53.480 | - So you think that's what it is?

00:41:54.640 | - My definition for AGI is smarter

00:41:56.620 | than the smartest human being who ever lived.

00:41:58.540 | - Yeah, I was talking to somebody this week

00:41:59.880 | who's in a position who said the definition of AGI

00:42:03.000 | is very fuzzy, that there isn't a clear definition.

00:42:06.020 | And therefore, it allows every side to kind of anchor

00:42:09.240 | on their interpretation of what that term means

00:42:12.000 | and therefore kind of justifies their position.

00:42:14.280 | So I don't really feel great about like just saying,

00:42:16.800 | are we at AGI?

00:42:18.180 | We don't have a clear sense of what it means.

00:42:20.080 | I do think if you look at some of the work that was done

00:42:22.500 | by Anthropic and published in the Cloud3 model this week,

00:42:25.660 | did any of you guys see the demos that were done

00:42:27.600 | of the output of that model?

00:42:28.960 | There was a guy who wrote a thesis in quantum physics

00:42:32.240 | on a very esoteric, complicated problem set.

00:42:34.820 | And he asked Cloud3 to solve this problem set

00:42:37.500 | and it came up with his thesis.

00:42:39.480 | It was really like extraordinary.

00:42:40.880 | And this is something he's like,

00:42:41.820 | no one in the world knows this stuff.

00:42:43.180 | And he's like, I can't believe this model

00:42:45.380 | like came up with my thesis.

00:42:47.840 | And that's the sort of thing that very few people on earth

00:42:50.280 | even read or understand.

00:42:52.020 | And the Cloud3 model was able to kind of recreate

00:42:54.760 | the basis, the buildup and then the output of his thesis

00:42:58.360 | was really extraordinary.

00:42:59.200 | - It'd be like somebody writing a screenplay

00:43:00.280 | and then giving it the first two acts

00:43:01.800 | and say, guess the third act.

00:43:02.840 | And it's like, oh yeah, this is the third act,

00:43:04.320 | here's what happens.

00:43:05.480 | Like it's pretty impressive, the reasoning ability.

00:43:07.680 | - Yeah, complicated.

00:43:08.520 | - Let me ask you, like deep down,

00:43:09.900 | when these guys say they're gonna create AGI,

00:43:12.260 | what do you think they really mean

00:43:14.460 | in their heart of hearts?

00:43:16.180 | - Oh, they mean the Terminator.

00:43:17.320 | Yeah, they mean the sentient God.

00:43:19.480 | - You guys are propagating some bad (beep)

00:43:22.080 | you guys shouldn't be saying that.

00:43:23.400 | - I think that's what they think.

00:43:24.560 | - Remember what Larry Page said to Elon,

00:43:28.760 | don't be speciesist.

00:43:30.600 | - Yeah.

00:43:31.440 | - I think there's a meaningful number of people

00:43:34.000 | in the tech community who deliberately wanna give rise

00:43:37.920 | to the super intelligence.

00:43:40.000 | - There's another point of view of super intelligence

00:43:41.760 | where super intelligence means that the software

00:43:44.640 | is now more intelligent than all humans.

00:43:46.860 | And as a result, the software may have its own motivations

00:43:49.680 | to figure out how to supersede humans on earth.

00:43:53.320 | Now, the Larry Page statement,

00:43:55.840 | which I don't know firsthand,

00:43:57.600 | I read the same article you did,

00:43:59.540 | is one that a group of people might say,

00:44:01.880 | evolution is evolution.

00:44:03.260 | You know, there are people-

00:44:05.280 | - Yeah, they would say that.

00:44:06.400 | - That's right.

00:44:07.380 | And I know Elon's taking this point of view

00:44:09.200 | that like, you know, we need to maintain human supremacy,

00:44:13.720 | but my favorite test is, you know,

00:44:17.560 | there's the Turing test,

00:44:18.520 | which like you can't tell if it's a human

00:44:20.400 | or if it's a robot,

00:44:21.720 | but Mustafa came up with the modern Turing test,

00:44:25.120 | which is an AI model is given $100,000

00:44:27.600 | and has to obtain 1 million, go.

00:44:30.520 | - Kind of interesting.

00:44:31.360 | - That's a good idea.

00:44:32.760 | - The other interesting one was Gary Marcus.

00:44:35.040 | He said the IKEA test, the flat pack furniture test,

00:44:39.680 | and AI views the parts

00:44:40.840 | and instructions of an IKEA flat pack product

00:44:43.120 | that controls a robot to assemble the furniture correctly.

00:44:46.480 | - Oh, look, what's up, Sonny?

00:44:48.680 | It's now time to do just a quick little congratulations

00:44:52.440 | to our dear friend, Sandeep Madra.

00:44:54.760 | We call him Sonny, that's his nickname.

00:44:56.660 | And he's one of our poker buddies.

00:44:57.960 | He keeps building great companies.

00:44:59.980 | And in this case,

00:45:01.760 | Sandeep is the first person to collect all four besties.

00:45:04.360 | We all invested in his company, Definitive.

00:45:06.900 | And Definitive was working in AI,

00:45:10.000 | but we got some great news this week

00:45:12.640 | and we thought we would give him his flowers.

00:45:14.640 | Sonny, you want to tell us what happened this week

00:45:16.660 | with our investment in your company,

00:45:19.180 | Definitive Intelligence at definitive.io,

00:45:21.840 | I believe is your domain.

00:45:23.340 | - Yeah, well, you know, with your guys' support

00:45:25.420 | and, you know, we've been growing our company

00:45:27.500 | and we saw a really great opportunity

00:45:29.660 | to work together with Grok

00:45:30.900 | and we've been working with them for a couple of months

00:45:32.880 | and all the hype that you've seen has been built on

00:45:35.920 | the collaborations that we've done,

00:45:37.640 | building the cloud offering, the API offerings.

00:45:40.280 | And so, you know, we've decided to merge with them

00:45:43.180 | and we're super excited.

00:45:44.240 | And all the besties are now not only shareholders

00:45:48.120 | in Definitive previously, but now shareholders in Grok.

00:45:50.720 | - This is the first investment, I think,

00:45:52.660 | where we're all on the cap table, right?

00:45:54.440 | So we're all rooting in the same direction.

00:45:56.480 | What's happening in the developer side?

00:45:58.160 | How's the momentum going?

00:45:59.000 | - Well, the momentum's incredible.

00:46:00.860 | You know, we have now 16,000 plus developers

00:46:04.520 | in our self-serve playground.

00:46:07.400 | There's well over like a thousand apps

00:46:09.800 | that people have developed using the API

00:46:12.320 | and all kinds of new functionality.

00:46:14.240 | The API allows people to get a higher rate

00:46:17.120 | of throughput on tokens and low latency.

00:46:19.320 | So there's all kinds of new applications

00:46:20.760 | from voice to real-time translation of web pages.

00:46:25.240 | We're collecting them on our Discord.

00:46:26.720 | We have 3,000 people in the Discord as well.

00:46:30.360 | It is a real community that's come together

00:46:32.320 | building around Grok.

00:46:33.240 | And what I will say is sort of the same jump

00:46:35.960 | that developers saw when we went from dial-up internet

00:46:38.240 | to broadband, they're seeing that now

00:46:40.040 | and from using like traditional APIs for LLMs

00:46:42.960 | to using, you know, the ones that we offer.

00:46:45.760 | - You guys support the latest and profit models

00:46:48.400 | that they just launched that seemed pretty kick-ass.

00:46:51.240 | - No, we don't have those yet, but we're in just, you know,

00:46:53.820 | we're having discussions with everyone out there

00:46:56.000 | and we want to support their models.

00:46:57.960 | Right now, what we've done given all the demand

00:46:59.940 | is we've kind of limited it to Llama 2, 70B and Mixtrol.

00:47:03.320 | And we actually have a bunch of other models

00:47:04.840 | that we make available in private mode for folks.

00:47:07.080 | We're pretty excited.

00:47:08.240 | But if there's anyone out there

00:47:09.320 | that wants to have us, you know, give us a call

00:47:11.400 | and we'll get you going on our systems.

00:47:13.840 | - Well, Sonny, congrats to you.

00:47:16.420 | You're an incredible entrepreneur.

00:47:19.920 | And it's always fun to kind of be on the journey with you.

00:47:23.020 | This is my fourth business that I've done with Sonny Madra.

00:47:27.080 | - That's amazing.

00:47:27.920 | Extreme Lens. - Yeah.

00:47:29.080 | - Then there was a company in between.

00:47:30.520 | - No, no, no, there was a company in between.

00:47:32.500 | You have to understand, Sonny and I met

00:47:33.800 | because Sonny went to school where I grew up.

00:47:35.800 | He went to the University of Ottawa.

00:47:37.080 | I grew up there and we met through a mutual friend

00:47:39.960 | and his first company was called Spongefish,

00:47:42.200 | which I backed, I think in 2006 maybe.

00:47:46.820 | I mean, I had no money.

00:47:47.880 | I may have written a $10,000 check.

00:47:49.360 | - Would you give me $8,000?

00:47:50.920 | - 10K maybe. - Maybe less.

00:47:52.480 | - 10K check? - Maybe less, 5K.

00:47:53.760 | Yeah, I mean, whatever I had, I didn't have much.

00:47:55.880 | - I like it, scraped together.

00:47:57.440 | - And it's our fourth business.

00:47:58.680 | - Sonny's a variable rabbit's foot.

00:48:00.680 | You're a rabbit's foot, Sonny.

00:48:02.480 | - It is true, he is.

00:48:03.320 | He's a luck bot.

00:48:04.320 | - Luck prefers the prepared.

00:48:06.760 | - Well, I mean, I've been in two of Sundeep's companies

00:48:10.560 | and I thank him for that.

00:48:12.020 | And then Extreme Labs, actually, to thank you again.

00:48:14.920 | You sponsored a lot of my events 10 years ago

00:48:17.360 | when I was coming up, so I appreciate that.

00:48:18.740 | So it's just great to see a nice guy win

00:48:21.100 | as opposed to the three miserable people

00:48:23.400 | who I work with on this podcast keep winning.

00:48:25.720 | So that's good.

00:48:26.560 | Now, finally, a good guy wins

00:48:27.560 | as opposed to the four of us.

00:48:28.720 | - Well, I appreciate all your guys' support.

00:48:30.580 | It means a lot to us, you guys getting the developers

00:48:33.520 | and keep putting it out there and keep us honest as well.

00:48:37.360 | So if we're not doing something right, let us know.

00:48:39.080 | Really appreciate you guys all.

00:48:40.840 | - Sax, you wanna say something motivational to Sonny here?

00:48:44.080 | I know you always have a great motivational word

00:48:46.380 | for your friends, something you say that just gives people

00:48:49.540 | that thrill of being in the game.

00:48:51.460 | Go ahead, Sax, you always have something kind to say.

00:48:53.560 | - We're gonna ship 20 million into your safe note.

00:48:56.400 | I know you closed it, but we're gonna pry it open

00:48:58.360 | and get some allocation.

00:49:01.960 | - Wait a second, if you're prying open that safe,

00:49:04.680 | the safe's, if there's a wedge,

00:49:06.800 | I know a guy with a podcast who's got 800,000 followers,

00:49:09.640 | maybe I could slip in a quick 500,000.

00:49:11.360 | Is there possible?

00:49:12.360 | What do you think, Sandeep?

00:49:13.280 | Let us know right now.

00:49:14.120 | - I don't know, we did a really big announcement

00:49:15.720 | with the Saudis this week, so that price might be up.

00:49:18.480 | - So let us ship something into that safe note

00:49:20.340 | so we actually have some real skin in the game.

00:49:21.840 | - David, text me.

00:49:22.960 | - All right, we'll do.

00:49:23.800 | - Well, no, but seriously, hey, Sandeep,

00:49:25.240 | wait a second, if he gets in--

00:49:26.060 | - I mean, if you're gonna let him text you--

00:49:26.900 | - I gotta get a second chip in the 500,000.

00:49:28.400 | - Jason wants in for 500, by that he means $500,

00:49:31.760 | so we'll let you into it.

00:49:32.600 | - No, I got a gift card for you from Starbucks.

00:49:34.840 | I got a $500 gift card for you and the team.

00:49:37.380 | I think I got like 150,000 in the brief credits.

00:49:39.220 | - Jacob's gonna redeem that Starbucks gift card.

00:49:42.040 | He'll sell it for 20 cents on the dollar

00:49:43.580 | and then ship it into the--

00:49:44.420 | - Yeah, let me zip it in.

00:49:45.240 | Let me zip it in, maybe I can 100X it.

00:49:46.360 | No, seriously, 500K from your boy, Jay Cal.

00:49:48.760 | All right, everybody, thanks to Sandeep for jumping on.

00:49:51.920 | Congratulations on the murder.

00:49:53.040 | - Thanks, guys.

00:49:53.880 | - All right, issue two, issue two.

00:49:55.800 | Apple is battling two major iOS developers

00:49:59.080 | and regulators are siding with the devs.

00:50:01.960 | You may know about the Apple versus Epic Games saga.

00:50:04.520 | We've talked about it here.

00:50:06.440 | Epic Games is planning to create a custom app store on iOS

00:50:10.360 | because Europe's DMA, the Digital Markets Act,

00:50:14.340 | has said that Apple now has to allow

00:50:16.560 | third-party app stores in the EU.

00:50:19.120 | So Epic created a developer account based in Sweden

00:50:22.080 | and Apple actually approved the account two weeks ago.

00:50:24.400 | Then on Wednesday, Apple flipped,

00:50:26.080 | terminated Epic's EU developer account,

00:50:28.800 | and Apple said one of the reasons

00:50:30.160 | they terminated the account was because Epic's CEO

00:50:32.100 | publicly criticized their DMA compliance plan.

00:50:35.520 | Additionally, on Monday, Apple was fined 2 billion

00:50:38.600 | by the EU's antitrust regulators

00:50:40.800 | and was forced to remove its anti-steering rules

00:50:43.960 | from music apps like Spotify.

00:50:45.920 | Basically, Apple has been restricting music apps

00:50:47.640 | from informing users about pricing and discounts

00:50:50.600 | and the European Commission considered this anti-competitive

00:50:53.580 | since Apple runs Apple Music.

00:50:55.900 | And they want Spotify to pay 30%.

00:50:58.660 | Daniel Ek, a friend of the pod,

00:51:00.520 | basically did a whole video on this

00:51:03.600 | about how they can't charge 30% more, yada yada.

00:51:06.520 | Sax, you've spoken about Apple's monopoly before.

00:51:10.540 | Your thoughts on what's happening in the EU

00:51:13.040 | and then we'll get into Apple's wider problems.

00:51:16.640 | - I mean, did you just say that Apple booted Epic

00:51:20.740 | from their app store because they didn't like

00:51:22.320 | what Epic said about them?

00:51:24.240 | I mean, talk about-- - Their feelings were hurt.

00:51:26.420 | - Well, like, they're violating Epic's free speech

00:51:29.560 | because they don't like what Epic is saying.

00:51:31.220 | I mean, this is-- - According to Epic, yes.

00:51:32.960 | - This is crazily heavy-handed by Apple.

00:51:36.060 | - Apple heavy-handed with developers, yeah.

00:51:38.040 | - Have they lost their minds?

00:51:39.180 | I mean, this is right out of power corrupts

00:51:41.600 | and absolute power corrupts, absolutely.

00:51:43.940 | I mean, whatever dispute you have with Epic,

00:51:46.320 | you don't boot them out of your app store

00:51:47.840 | 'cause you don't like their criticism of you.

00:51:50.740 | I mean, this is basically proving

00:51:53.280 | exactly what everyone's been saying about Apple,

00:51:55.660 | which is they're too powerful and heavy-handed.

00:51:58.440 | And Apple's coming along and saying,

00:51:59.800 | "Let me confirm it for you guys

00:52:01.380 | "by acting tyrannically against Epic."

00:52:04.900 | I mean, talk about a backfire.

00:52:06.480 | This seems insane to me.

00:52:08.280 | - Yeah, it's super nuts.

00:52:10.280 | Chamat, any thoughts on this

00:52:11.560 | before we get into Apple's other problems?

00:52:13.880 | - I think it's the beginning of the decay of Apple.

00:52:18.880 | Peak Apple, well, let's get into that.

00:52:21.080 | There's tons of headwinds facing Apple.

00:52:22.700 | - You can add a bunch of other things

00:52:23.920 | to the list as well, Jason.

00:52:25.460 | - Yeah.

00:52:26.520 | - The thing is Apple for the last couple of years

00:52:28.220 | has been what is effectively

00:52:30.160 | what we call a GDP plus growth company,

00:52:32.960 | which means that take GDP, two, three, 4%.

00:52:36.440 | Maybe they can grow by a couple of percentage points

00:52:38.440 | more than that, but they are effectively levered to GDP.

00:52:42.160 | Meaning when you look at a Facebook or an Nvidia,

00:52:45.260 | they're growing at 50%, 200%, 2,000%,

00:52:48.280 | whatever it is, that's not tied to GDP.

00:52:50.920 | They're just taking share.

00:52:52.700 | But Apple is a company now that grows

00:52:54.600 | as the economy of the world grows.

00:52:56.400 | So that's not super great for its future prospects

00:52:58.760 | unless it can expand the surface area of where they operate.

00:53:03.800 | And then on that dimension,

00:53:05.680 | there are a few trillion dollar markets

00:53:07.840 | that they can really penetrate.

00:53:09.880 | And they just announced that they've killed a project

00:53:11.960 | in one of those areas, which is autos, right?

00:53:13.820 | Project Titan, which was $10 billion

00:53:16.280 | turned out to be a failure.

00:53:18.580 | So all of these things I think mean to me

00:53:20.260 | that it is effectively becoming a cyclical,

00:53:24.140 | rate sensitive stock.

00:53:25.540 | And then the coup de gras is Warren Buffett.

00:53:29.880 | And Nick and I were talking about it this week.

00:53:32.800 | And what was interesting about Buffett's letter is that

00:53:35.500 | you can tell when Buffett has gotten disengaged

00:53:41.140 | with a company based on the number of times

00:53:43.860 | he mentions it in his annual letter.

00:53:46.500 | So in this example,

00:53:48.060 | this is the number of times Apple was mentioned.

00:53:50.060 | And just to be clear, what I mean by mentions

00:53:52.260 | is not when it's included in a chart

00:53:53.780 | or part of a disclosure.

00:53:54.780 | What I mean is when Warren actually explicitly mentions it

00:53:58.340 | in a positive or even negative way,

00:54:01.300 | or he doesn't mention it at all,

00:54:02.560 | which I think rings very loudly.

00:54:04.820 | He went from basically saying,

00:54:07.180 | Apple was the absolute end all and be all.

00:54:11.500 | And now what you can start to see is this shrinking.

00:54:14.660 | And it's gone from basically a bunch of times

00:54:16.940 | to almost none.

00:54:17.860 | He did mention it once,

00:54:19.140 | but he mentioned it in the context of talking positively

00:54:21.700 | about Coca-Cola and Amex.

00:54:23.560 | And he was lauding these two positions,

00:54:26.260 | but just mentioning that they were not as large

00:54:28.860 | in comparison to Apple.

00:54:29.860 | That's the only mention in this year's annual letter.

00:54:33.440 | What's interesting about that is the last time

00:54:35.060 | that that happened was with, drum roll,

00:54:37.000 | Wells Fargo.

00:54:38.840 | Over 15 and 20 years,

00:54:40.860 | Buffett built up a huge position.

00:54:44.140 | I think he was able to weather the vicissitudes

00:54:47.280 | of the market.

00:54:48.120 | So even when the markets would contract,

00:54:49.780 | he knew when to hold onto that company

00:54:52.120 | until he realized that that company

00:54:53.820 | was not really one of his forever stocks.

00:54:56.060 | And he got out of it.

00:54:58.200 | And the number of times it basically was mentioned

00:55:01.960 | in his letter went to zero.

00:55:03.040 | So interestingly, I think this Buffett index

00:55:07.300 | is a really important one for Apple,

00:55:08.780 | which is it went from a forever holding

00:55:11.620 | that he said he would own forever

00:55:14.380 | to barely getting mentioned.

00:55:15.800 | And above Apple on that list

00:55:18.180 | were all the Japanese trading companies that Buffett owns,

00:55:21.740 | American Express, Coca-Cola.

00:55:23.300 | So that is a person that understands the economy,

00:55:27.440 | I think, better than anybody else in the world.

00:55:29.940 | And so if you're basically taking a lever bet

00:55:32.640 | to the economy as a reason to own Apple

00:55:34.620 | and the person that understands the economy the most

00:55:37.500 | has now started to pivot away,

00:55:39.260 | he started to sell in quarter four.

00:55:41.420 | And then you see all of these things,

00:55:42.700 | antitrust rules, killing projects in trillion dollar terms.

00:55:47.340 | Unfortunately, it speaks for a very bad

00:55:50.360 | next five to 10 years for this company

00:55:51.980 | unless they figure something out.

00:55:53.340 | - I think it's just such a great insight

00:55:55.180 | with the Warren Buffett mentions.

00:55:57.100 | And we should do that forward looking,

00:55:58.380 | like who is he talking about now?

00:56:00.100 | And let's the monitor when he stops talking about them.

00:56:02.540 | I have a theory about this.

00:56:03.900 | I think this is really just about peak iPhone.

00:56:06.900 | If you look at the majority of their revenue,

00:56:08.700 | it's obviously from the iPhone,

00:56:10.660 | which has become massively profitable over time.

00:56:13.380 | But the iPhone revenue has been flat

00:56:15.820 | for a couple of years now.

00:56:17.540 | And they're starting to make their money from services.

00:56:19.820 | Everybody knows Apple One, iCloud storage,

00:56:22.780 | Apple Arcade, Apple Music, Apple Plus TV.

00:56:26.080 | They've got a great collection of services

00:56:28.060 | and services revenues growing.

00:56:29.860 | But I don't think that they are filling in

00:56:33.060 | this growth problem with the iPhone.

00:56:34.740 | And this happened to me for the first time.

00:56:36.380 | And I don't know if you gentlemen

00:56:37.740 | have had the same experience,

00:56:38.660 | but I was buying every Apple phone

00:56:40.900 | and then I would buy every like medium upgrade,

00:56:44.380 | like when they would do like a 12 and then a 12S

00:56:46.740 | or an 11, 11S.

00:56:48.440 | And when I got to iPhone 13,

00:56:49.700 | I was like, this thing's kind of peaked.

00:56:51.560 | And I just forgot to buy the 14, didn't need it.

00:56:54.320 | And then I bought the 15

00:56:55.400 | and I'm sitting here on my 15 and my 13,

00:56:57.020 | I really couldn't tell the difference between them.

00:56:58.500 | And so now I'm a technologist

00:57:00.380 | who would buy it every year, the latest one,

00:57:02.820 | and I don't feel the need to upgrade it.

00:57:04.980 | And I think that a lot of people,

00:57:06.500 | and I have family members who would take two generations off,

00:57:09.780 | now they take three or four generations off.

00:57:11.620 | - I don't even know what number I'm on.

00:57:14.100 | - I don't even know either.

00:57:15.140 | It all looks the same to me.

00:57:16.660 | Okay, so I mean, that's the problem here.

00:57:18.940 | And then if you look at their roadmap,

00:57:20.860 | what is the device post the Steve Jobs era

00:57:24.900 | that they have launched?

00:57:26.220 | How do you know what device you have?

00:57:28.820 | - You go to general in your settings and about,

00:57:31.060 | I just checked it, iPhone 13 Pro.

00:57:33.780 | What version are they on?

00:57:35.380 | - 15 general.

00:57:36.220 | - I'm on 15 and I'm on 13 Pro.

00:57:37.660 | I don't even know.

00:57:38.500 | - Yeah, and you're not price sensitive.

00:57:41.420 | And these things have gotten absurdly expensive.

00:57:43.300 | So they're trying to really extract--

00:57:44.860 | - What is the latest?

00:57:46.660 | - 15.

00:57:47.860 | - 15, I'm on a 14.

00:57:49.060 | - I don't like the hassle of having to reset everything

00:57:52.180 | when I get a new phone.

00:57:53.020 | - Yes, and it's easy to do now.

00:57:55.020 | It's all in the cloud, but even that,

00:57:57.060 | going to the store, buying it, who cares?

00:57:58.980 | It's not differential.

00:57:59.820 | - But then on my apps, I have to re-log in and--

00:58:01.980 | - Frebert, what Android phone are you on?

00:58:05.340 | - Oh, you're a Googler.

00:58:06.260 | Didn't the Googlers all have Android phones?

00:58:08.100 | Or are you your team iPhone?

00:58:11.220 | - I got the Macintosh original in 1984.

00:58:13.940 | I've only been on Macintosh and Apple products since then.

00:58:18.820 | - What iPhone are you on, be honest?

00:58:20.660 | - 15 Pro.

00:58:22.020 | - Okay, so--

00:58:22.860 | - Have you ever tried a Pixel, Frebert?

00:58:26.300 | - I don't like it.

00:58:28.460 | But I just, I'm not used to the OS, I just--

00:58:30.540 | - What was the rule at Google?

00:58:31.660 | If you worked at Google, were you kind of forced

00:58:34.300 | to have an Android phone?

00:58:35.500 | Or did you just keep your iPhone in your pocket

00:58:37.180 | and not take out?

00:58:38.260 | - There weren't a lot of Androids on the market

00:58:40.620 | when I was at Google, because we acquired the company

00:58:43.900 | when I was there, and oh, we acquired Andy Rubin's company

00:58:47.820 | in '04, and I think the first major devices

00:58:52.820 | started to roll out in '07.

00:58:56.340 | And so this was all post my era.

00:58:58.180 | I left at the end of '06.

00:59:00.340 | - Wow, Frebert, that's an amazing chart.

00:59:02.460 | Holy (beep)

00:59:03.300 | - Look at this chart.

00:59:04.300 | Now, here's the global statistics.

00:59:07.140 | The U.S. market share is 57% for iOS,

00:59:10.620 | but on a global basis, iOS only has a 27% market share

00:59:15.620 | in mobile operating systems.

00:59:18.100 | Android is 72%, and all other is less than 1%.

00:59:22.180 | - But look at this, it's all the future GDP is on Android.

00:59:25.960 | Brazil, Nigeria, India, holy mackerel.

00:59:29.660 | - But I think this is one counter to your point, Chamath,

00:59:32.300 | that the macro driver is as the economic position,

00:59:36.740 | the GDP per capita scales in these BRICS nations,

00:59:41.000 | they can start to afford to buy iPhones,

00:59:44.380 | which are generally some multiple of the Android devices

00:59:47.300 | in these markets that you see here.

00:59:49.020 | So while today Android is 72% of the market,

00:59:52.260 | if the emerging markets continue to grow GDP per capita

00:59:55.180 | and iOS continues to be the superior product,

00:59:58.080 | you'll see Apple able to steal into more share over time.

01:00:01.740 | But how do you do that when you have messaging groups

01:00:04.620 | on Android, when you have photos in Google Photos?

01:00:07.860 | Does the switching cost stop that, do you think?

01:00:11.380 | - I don't know the answer to that.

01:00:12.340 | - It's a good question.

01:00:13.680 | They're down to $15 to $25 for these Android phones.

01:00:17.420 | In India, there's a famous-

01:00:18.860 | - That's the difference, right, to these Android devices.

01:00:21.880 | - I think Apple is levered to global GDP,

01:00:24.060 | so they need to figure out some way to grow superior to that.

01:00:27.700 | So what is the market?

01:00:29.020 | What could they do?

01:00:30.300 | Services?

01:00:31.740 | - A company that generally has a bunch of cash flow

01:00:36.500 | being generated by some set of products today,

01:00:39.300 | and you assume stasis over time,

01:00:42.040 | the market share for those products could be eaten away,

01:00:44.260 | you'll see the profits per year go down,

01:00:46.580 | and a company like that will trade anywhere

01:00:48.140 | from seven to 20 times.

01:00:50.280 | The technology value arises from the value of the brand

01:00:52.960 | that you can launch new products, leverage your brand,

01:00:55.080 | leverage your distribution,

01:00:56.100 | leverage the sale of new services, and new products.

01:01:00.340 | I think, to your point, Shamav,

01:01:01.860 | the challenge Apple is facing is that the pool of options,

01:01:05.580 | the portfolio of call options

01:01:07.700 | that you would get a new product coming out of Apple

01:01:09.860 | is shrinking with the car being taken out of that pool now.

01:01:14.860 | Apple Vision Pro, a lot of question marks

01:01:17.340 | on how much it can scale.

01:01:18.700 | As you guys know,

01:01:19.540 | I feel like there's gonna be a market for that device,

01:01:22.020 | but it's a high-end device today.

01:01:24.100 | It has to become more cheap to be more ubiquitous.

01:01:26.980 | - The thing that a lot of these companies confront, though,

01:01:29.380 | is that you can also grow inorganically, right?

01:01:31.720 | You don't necessarily have to incubate these projects.

01:01:34.580 | We can remember the moment Apple had a chance to buy Tesla,

01:01:38.200 | right? - Didn't take it.

01:01:40.700 | - Then they didn't take it.

01:01:42.240 | Apple still has a chance to buy,

01:01:44.300 | pick your company that would take a sweet acquisition offer,

01:01:48.900 | Rivian, Lucid, Polestar, whatever it is.

01:01:51.460 | I'm not saying that these companies are good or not good.

01:01:53.240 | I'm just saying that Apple has the one tool

01:01:56.480 | that they've never used in their toolbox

01:01:58.340 | is the large inorganic acquisition.

01:02:00.880 | And at this point,

01:02:02.920 | if they are proving that they can't execute internally,

01:02:06.420 | the market is going to demand that they prove

01:02:08.340 | they can use that balance sheet as a cudgel

01:02:10.380 | to go and execute it externally.

01:02:12.460 | - Yeah, I think the challenge- - And if they don't do that,

01:02:14.420 | they're gonna discount that cash to zero.

01:02:16.520 | - The challenge is that there's a certain discipline

01:02:21.000 | and quality to the products and the businesses

01:02:23.420 | that Apple produces and runs.

01:02:26.060 | And it's very hard to see that in other markets.

01:02:28.260 | I mean, why would they go buy a money-losing,

01:02:30.340 | low-margin car hardware company

01:02:33.780 | when they are proven to make high quality-

01:02:35.840 | - It depends on the reasons Project Titan failed.

01:02:38.400 | - Yeah, I think-

01:02:40.980 | - The article in Bloomberg basically said

01:02:42.940 | that the reason that they abandoned the project

01:02:45.000 | was in part that they did not know

01:02:48.580 | how to go from where they were to a full production vehicle

01:02:52.920 | that could be level five.

01:02:54.040 | And then when it was proposed that they step down

01:02:56.140 | and just launch a level three autonomous vehicle,

01:02:58.700 | everybody said no, that it wasn't disruptive enough.

01:03:01.400 | Well, if that's the case,

01:03:02.640 | the thing that they made a decision about there

01:03:05.440 | was not going into a market because of regulation,

01:03:07.680 | not because of technical capability.

01:03:10.240 | And I don't think that that's necessarily a smart decision.

01:03:13.160 | They would have been better off going into the car market,

01:03:16.040 | launching a level three vehicle

01:03:17.360 | and just letting the market play out.

01:03:18.560 | They probably, just like they were able to do in music,

01:03:21.500 | have the influence to change the laws,

01:03:24.240 | especially if they stepped in there with their rigor.

01:03:27.100 | And they're not doing that.

01:03:28.140 | So it's a little bit of a head scratcher

01:03:29.920 | what's going on over there.

01:03:30.760 | - Yeah, they need some new products.

01:03:32.120 | I mean, maybe Apple Vision Pro someday

01:03:34.420 | becomes the platform of the future,

01:03:36.420 | but they've squeezed as much revenue

01:03:38.020 | as they can out of the iPhone.

01:03:39.400 | Services seems like the best place for them to make money,

01:03:41.700 | but as we see-

01:03:43.100 | - Well, my idea for them was-

01:03:43.940 | - The app source under assault, yeah.

01:03:46.360 | - My idea was just that they should launch a huge competitor

01:03:48.660 | to AWS, Azure and GCP.

01:03:50.600 | So if you look, every developer that Apple has,

01:03:54.360 | theoretically could have been running on an Apple,

01:03:56.960 | not just on Apple's SDK and APIs,

01:03:59.160 | but they should be running in an Apple cloud.

01:04:00.780 | You could have made that claim

01:04:02.520 | and it would have made a ton of sense for app developers

01:04:04.680 | to have turnkey access to that, right?

01:04:06.760 | And they could have subsidized it.

01:04:08.060 | And by the way, to Sax's point,

01:04:09.960 | that would have been an incredible way

01:04:11.460 | to defend a 30% rev share.

01:04:13.040 | Okay, listen, I'm taking 30%,

01:04:14.920 | but here's a bunch of subsidized access to hardware

01:04:17.640 | and the market would have loved it.

01:04:19.660 | And in this AI shift, they can still do that,

01:04:22.240 | where now people are chipping away at the 30%, right?

01:04:25.880 | People are saying, "Well, I can just build around you."

01:04:29.120 | They need to do something and services, Jason,

01:04:31.280 | the most valuable thing they could do

01:04:34.360 | would be to launch a big cloud, I think.

01:04:36.640 | - Yeah, I mean, they have all the app developers

01:04:38.640 | on TestFlight ready to go.

01:04:40.200 | They just email them and say, "Hey, here's your free storage."

01:04:42.400 | I mean, they could just slowly add features, right?

01:04:45.080 | And they're doing this open source,

01:04:48.040 | I think it's called Maggi or Magpie or something,

01:04:50.800 | but they're doing an image one.

01:04:52.400 | - Well, look, if they could just make Siri work

01:04:56.560 | like the way it was supposed to using an LLM

01:04:59.880 | and have it be like snappy,

01:05:01.560 | that'd be a major upgrade to the iPhone.

01:05:04.460 | I don't think you need to buy a new iPhone for that

01:05:05.920 | 'cause it'd just be a software upgrade,

01:05:07.240 | but just getting Siri to work would be a big win.

01:05:12.160 | - I think that's where they're going with their silicon,

01:05:13.960 | and they just announced the M3 on the MacBook Air.

01:05:16.720 | It's like they're making their own silicon,

01:05:18.000 | and I think it's going to power LLMs locally on the devices,

01:05:21.480 | and that's so powerful.

01:05:22.880 | I was using iPhoto this week,

01:05:25.640 | and in Apple Photos, there's now on certain images,

01:05:29.360 | if you swipe through them, you'll see the little AI icon.

01:05:32.400 | When you click it, it tells you things in the photo,

01:05:35.640 | like that's a bulldog, that's pasta.

01:05:38.120 | And so they're already subtly adding these features.

01:05:40.320 | Now, that's a feature Google's had

01:05:42.120 | in Google Photos for five years,

01:05:44.640 | but go ahead and open your phone right now

01:05:46.280 | and search for a dog or a bulldog

01:05:48.140 | or whatever type of dog you got, you know, Labrador,

01:05:51.400 | and then watch Golden Retriever,

01:05:53.160 | that it actually knows how to do it.

01:05:54.360 | They never announced it.

01:05:55.360 | It's just subtly being put in there,

01:05:57.700 | so there's so many opportunities.

01:05:59.200 | - If they required a hardware upgrade

01:06:01.120 | to get actually good AI built into the software,

01:06:05.200 | then everyone's going to have to upgrade for that.

01:06:07.020 | - You'd be on 16.

01:06:07.860 | - That's a pretty big motivation to upgrade.

01:06:10.280 | - There you have it, Tim Cook.

01:06:11.160 | Somebody send this clip to Tim Cook.

01:06:12.840 | Put AI chips on your phone.

01:06:15.280 | - But this isn't hard, right?

01:06:16.380 | I mean, all they gotta do is just

01:06:18.760 | take the latest open source models

01:06:20.240 | and figure out how to customize them for their own products.

01:06:23.240 | - Yeah, and they're doing it.

01:06:24.080 | They're doing it right now.

01:06:25.120 | All right, let's go to issue four, TikTok bipartisan ban.

01:06:28.640 | Will the CCP agree?

01:06:30.920 | A bipartisan group of a dozen plus lawmakers

01:06:34.280 | introduced a bill that would effectively ban TikTok.

01:06:36.760 | In the House this week, the bill is officially called

01:06:38.800 | "Protecting Americans from Foreign Adversary

01:06:41.360 | "Controlled Applications Act."

01:06:43.120 | Gives ByteDance 165 days to divest from TikTok.

01:06:46.480 | That's the parent company,

01:06:47.600 | the Chinese company that owns TikTok,

01:06:49.480 | the app that's very popular in the West,

01:06:51.600 | especially here in America.

01:06:53.320 | It would make it illegal for companies like Apple and Google

01:06:55.720 | to show TikTok in their app stores.

01:06:58.140 | As you know, TikTok, 170 million US users,

01:07:01.960 | and they claim that they're headquartered in Singapore.

01:07:05.120 | I know people who have worked at TikTok

01:07:06.960 | or do work at TikTok, and they said that's nonsense to me.

01:07:11.160 | The company said it has not and will not share user data

01:07:14.800 | with the CCP.

01:07:16.280 | In my mind, it's an obvious lie.

01:07:17.720 | In 2021, the CCP took a board seat

01:07:19.800 | on ByteDance Beijing-based subsidiary.

01:07:22.260 | That's according to Reuters.

01:07:23.280 | And then in 2022, ByteDance admitted that it accessed

01:07:26.040 | IP addresses and data by journalists covering TikTok

01:07:29.600 | to see if they'd been in the same places

01:07:31.040 | as ByteDance employees, obviously to find leakers.

01:07:33.940 | And ByteDance claims they fired the people involved,

01:07:35.880 | yada yada.

01:07:36.720 | Last year, a former head of engineering of ByteDance US

01:07:39.280 | said CCP members had god mode access to user data in 2018.

01:07:43.420 | Sax, I can go explain more of this,

01:07:46.000 | but I just got to go to you here.

01:07:48.640 | If ByteDance is not spying on Americans

01:07:52.980 | and the CCP is on the board, that

01:07:54.800 | would make no logical sense to me.

01:07:57.120 | And why are they fighting the divestiture

01:07:59.120 | if it's just the financial reason?

01:08:00.920 | Why would they take the CCP off their board?

01:08:02.760 | Do you think it's spyware?

01:08:03.800 | Do you think the US is crazy for allowing this product?

01:08:06.600 | Here in the United States, when we're not allowed to put

01:08:09.960 | Twitter, Facebook, Instagram, pick your social network,

01:08:13.000 | in China, we have zero reciprocity here.

01:08:16.880 | Look, if it's true that TikTok is sharing data with the CCP,

01:08:21.960 | then I think the United States is well

01:08:23.620 | within its rights to either ban it or cause it to be divested.

01:08:28.360 | And I personally like the divestiture option.

01:08:30.720 | I mean, that's what Trump was suggesting during his term.

01:08:33.520 | Because we don't, I think, in the United States like to just

01:08:36.560 | essentially confiscate or destroy people's property.

01:08:40.960 | But I think we are within our rights

01:08:42.520 | to require it to be divested to an entity

01:08:45.560 | that we know is completely separate from and won't

01:08:48.240 | cooperate with the CCP.

01:08:51.120 | This bill is better than the last bill

01:08:54.320 | we talked about that was targeted at TikTok.

01:08:56.680 | I don't know if you remember that one.

01:08:58.400 | But it was weirdly prohibiting Americans from using VPNs

01:09:03.240 | and gave the government the right to go after Americans

01:09:06.040 | who were using VPNs.

01:09:06.920 | So this one seems cleaner and better and more narrowly

01:09:10.200 | targeted at divestiture.

01:09:12.040 | Now, at the beginning of my response to you, I did say if.

01:09:17.460 | I know that everyone's just assuming

01:09:19.680 | that it's true that TikTok is sharing data with the CCP.

01:09:25.160 | But I just want to confirm that that is the case.

01:09:27.920 | Because they are denying it.

01:09:29.160 | And I can understand why people think it

01:09:31.280 | and why it might even be likely.

01:09:33.200 | But since we do have a concept of due process in America,

01:09:37.100 | I do think some evidence should be provided that that actually

01:09:40.640 | is taking place.

01:09:42.120 | Well, we've had whistleblowers inside there.

01:09:45.000 | But Chamath, if the CCP is spying on their own citizens,

01:09:48.480 | what are the chances that they wouldn't take the opportunity

01:09:51.020 | to spy on government officials who

01:09:53.280 | have TikTok on their phone or their kids

01:09:55.760 | and get compromise on them possibly

01:09:57.960 | or no locations of people?

01:10:00.520 | What does your gut tell you?

01:10:02.480 | Is this too dangerous for us to have here

01:10:04.840 | in America under CCP control or this kind of influence

01:10:07.900 | being on their board?

01:10:09.520 | I think you're asking the exact right question.

01:10:13.280 | I have two comments to make.

01:10:15.280 | One is, Nick, if you can bring up

01:10:18.680 | this article about the Google AI IP case.

01:10:22.520 | Basically what happened was that the DOJ filed an indictment.

01:10:29.280 | Actually, I think I sent you guys the actual indictment.

01:10:34.480 | But essentially what happened is there was an engineer at Google

01:10:37.800 | that has been charged with stealing AI secrets for China.

01:10:43.160 | And I don't know whether he's back in China now or not,

01:10:47.080 | but the whole point is that if it's happened at Google, where

01:10:51.680 | there is a motivation for the Chinese intelligence apparatus

01:10:54.960 | and, frankly, every intelligence apparatus

01:10:57.480 | to infiltrate that organization and get access

01:10:59.960 | to all kinds of data, I think we should

01:11:04.000 | presume by default that all of these organizations

01:11:07.800 | are infiltrated.

01:11:09.120 | And I think that that's probably a more conservative and

01:11:11.480 | reasonable posture.

01:11:12.360 | So Facebook is infiltrated, Google is infiltrated,

01:11:15.640 | Apple is infiltrated, TikTok is infiltrated.

01:11:18.840 | So on that dimension, I think that it

01:11:21.640 | should be considered 100% certainty that this data is

01:11:26.040 | getting back to not just the Chinese,

01:11:28.120 | but multiple state-sponsored actors.

01:11:30.960 | So the question about TikTok then, I think,

01:11:34.520 | should be one of business.

01:11:35.680 | And I think Palmer Luckey did a very good job of simplifying

01:11:39.840 | this down to its essence, which is essentially

01:11:42.040 | what he called the law of equivalent exchange.

01:11:43.960 | If you want to just play this, it's like just a few seconds.

01:11:46.000 | I was kind of frustrated that people made TikTok

01:11:48.120 | into a cultural issue.

01:11:49.920 | By the way, I'm totally on the culture war side of it.

01:11:53.960 | But I was saying, practically speaking, you should not

01:11:56.400 | make this a culture war issue.

01:11:58.080 | Don't talk about how it's ruining our used ideals.

01:12:01.040 | Just say strictly on a trade basis.

01:12:04.040 | We cannot allow them to sell this thing to us if we can't

01:12:09.000 | sell the same thing to them.

01:12:10.520 | That should be totally fair.

01:12:11.680 | That's reciprocity.

01:12:13.240 | That's reciprocity.

01:12:14.240 | Me and Palmer Luckey and Singh.

01:12:15.720 | Right, so this is what he calls the law of equivalent exchange,

01:12:18.200 | and I think it just makes a lot of sense.

01:12:19.880 | So on the face, what I would say is, Jake, out of my responses,

01:12:22.760 | I think that the CCP, but also other intelligence

01:12:26.080 | organizations, have infiltrated all of these big companies,

01:12:29.440 | and all of our data is accessible by them.

01:12:33.520 | I'm not going to say on a whim, but I think it's accessible.

01:12:37.080 | I think you have to deal with TikTok as a business issue,

01:12:39.480 | and I agree with Palmer Luckey, which

01:12:41.120 | is they should not be able to sell to us,

01:12:43.520 | but we cannot sell to them.

01:12:44.640 | And I think that that's a fair principle that we can live on.

01:12:47.120 | Reciprocity is a very simple position.

01:12:49.840 | Friedberg, let me use your creativity, your love of cinema.

01:12:55.600 | If you were to use this tool, let's take the most cynical

01:12:58.800 | approach here, or interpretation.

01:13:01.000 | CCP has complete access to the algorithms,

01:13:03.560 | and they want to do maximum damage,

01:13:05.600 | let's say, during the election.

01:13:07.000 | Let's say in a conflict like the one going on in Ukraine

01:13:11.380 | with Russia or in Gaza.

01:13:15.320 | What could they do using the algorithm, using videos?

01:13:19.800 | What would be the doomsday scenario for America?

01:13:22.560 | As in the CCP comes in and influences

01:13:24.840 | the management of this company and tells them

01:13:27.280 | what to tweak and how and why?

01:13:28.920 | Yes, what would they do?

01:13:31.280 | I think we saw this after October 7

01:13:36.800 | that there was a significant surge in pro-Hamas videos

01:13:43.400 | relative to Israel support video.

01:13:46.080 | That's the sort of thing where you could kind of see something

01:13:49.540 | that sets an opinion that may be disruptive to the social fabric,

01:13:55.420 | to the election cycle, that starts

01:13:57.360 | to get shared more frequently and shows up

01:13:59.640 | in feeds more frequently.

01:14:01.160 | Unlike Facebook and other places where there's a linear feed

01:14:03.840 | where you can scroll up and select what you want to watch,

01:14:06.160 | as you know, TikTok has already lined the videos up.

01:14:09.000 | So when you scroll up, they automatically

01:14:11.680 | play the next thing for you.

01:14:13.280 | So the ranking really matters in terms of viewership on TikTok,

01:14:17.240 | unlike a lot of other kind of select-to-play social media

01:14:20.960 | type networks.

01:14:21.500 | Could it shift an election?

01:14:23.960 | Hearts and minds in a war?

01:14:25.740 | Well, I've always said this.

01:14:26.900 | I think it's the craziest thing in the world

01:14:29.700 | that someone can spend advertising dollars

01:14:32.000 | and change someone's vote.

01:14:34.760 | Just think about that fact for a second.

01:14:36.420 | I mean, I've said this before, and people

01:14:38.080 | have told me I'm an idiot for saying it, but--

01:14:40.020 | Meaning that you can, or that it does?

01:14:42.240 | That both.

01:14:43.320 | [INTERPOSING VOICES]

01:14:45.860 | But think about it, like people don't individually

01:14:48.340 | go and gather data and then make an informed opinion about who

01:14:51.800 | they're going to vote for.

01:14:52.880 | Their opinion changes based on seeing an ad.

01:14:55.720 | It is so crazy to me that that's the truth.

01:14:58.580 | I think the country knew that it does.

01:15:00.180 | That's why they prevented it from happening until--

01:15:02.680 | For so long.

01:15:03.620 | Yeah, exactly.

01:15:05.120 | But I mean, I think that's what's so nuts is that there's

01:15:08.000 | no longer a forced discourse that kind of makes people go

01:15:13.000 | out and choose what content they want to consume,

01:15:15.880 | what they want to hear, debate stages, et cetera.

01:15:18.520 | That now it's about who spends the most money

01:15:20.360 | to get the most views in front of someone,

01:15:22.200 | and that that actually influences someone's decision

01:15:24.600 | on who to vote for is what's so compelling to me about why

01:15:29.120 | all of these systems have such extraordinary power.

01:15:31.360 | It's just so amazing to me that the more frequently someone

01:15:33.880 | sees an ad, the more likely they are to buy something

01:15:36.080 | or do something.

01:15:38.000 | Here's the bottom line.

01:15:39.000 | This thing is incredibly--

01:15:40.080 | So the more frequently you show someone a memetic on TikTok,

01:15:42.040 | the more likely they are to vote something differently.

01:15:44.480 | This thing is far too powerful for the CCP

01:15:47.080 | to have any kind of access to it,

01:15:48.680 | for the Chinese government to have any kind of access to it.

01:15:51.100 | It has to be divested.

01:15:52.320 | If you look at what we went through

01:15:53.840 | in the last couple of election cycles,

01:15:55.520 | not being partisan in here at all, but we've said

01:15:58.880 | and we've talked about Hunter Biden's laptop

01:16:00.720 | is but one example.

01:16:02.480 | That was suppressed, obviously, on social networks.

01:16:05.160 | That could have been amplified in social networks,

01:16:07.240 | and it could have had the opposite effect.

01:16:09.160 | Hillary Clinton's emails, Trump this, Hillary this,

01:16:12.760 | Biden that.

01:16:13.880 | You could really sway an election

01:16:16.800 | by putting in specific, subtle information, let alone

01:16:21.320 | taking a hack and releasing it like they did

01:16:24.240 | with Hillary Clinton's email.

01:16:25.400 | So those things might not have swayed an election.

01:16:27.720 | But, Sax, if Putin had access to this,

01:16:30.240 | like somebody who's super capable,

01:16:32.000 | and he had access to 170--

01:16:33.000 | Of course, that's always where it goes.

01:16:34.680 | No, I just--

01:16:35.280 | Is Putin is suddenly pulling the strings of our election?

01:16:38.720 | Iran, okay, Iran, North Korea.

01:16:40.040 | This is like the biggest threat inflation ever.

01:16:42.840 | Okay, look, if you ask people out there,

01:16:47.480 | do you think other people's votes

01:16:49.040 | are influenced by social media?

01:16:50.080 | They'll say, yeah, of course, people are brainwashed.

01:16:51.800 | If you say to them, is your own vote

01:16:54.920 | influenced by social media?

01:16:56.160 | Do you make up your own mind

01:16:57.920 | based on all the information you have?

01:16:59.800 | They'll say, yeah, of course, I'm not brainwashed.

01:17:01.760 | Everybody else is.

01:17:02.840 | And I believe that people are closer to telling the truth

01:17:06.520 | when they're talking about themselves

01:17:07.920 | because they understand their own situation

01:17:10.160 | better than they understand everyone else's situation.

01:17:12.680 | The fact of the matter is that all of us

01:17:14.640 | are constantly bombarded with information,

01:17:17.960 | 24 hours a day, seven days a week,

01:17:19.640 | through all of the channels, both online and offline,

01:17:22.400 | where we get information.

01:17:23.680 | Some of those data points come from advertisements,

01:17:27.080 | but I don't think we take ads very seriously.

01:17:29.560 | We're trained to kind of even just block out the ads.

01:17:33.000 | When I see banner ads or even ads in my stream,

01:17:35.800 | I just scroll past them.

01:17:38.680 | They don't even factor in my consciousness.

01:17:41.280 | I am influenced by accounts that I follow,

01:17:44.960 | but there are accounts that I've chosen to follow

01:17:46.840 | because I think they have signal over noise.

01:17:49.200 | And the more noisy those accounts are,

01:17:51.280 | the more I disregard them and take them less seriously

01:17:54.080 | with advertisements being the most noisy

01:17:56.960 | and least useful channels.

01:17:58.120 | So look, I think at the end of the day,

01:17:59.880 | this idea that we're all being brainwashed

01:18:01.960 | and secretly influenced by maligned foreign actors,

01:18:04.720 | I think, is at a minimum threat inflation,

01:18:07.840 | and that entire narrative might just be completely bogus.

01:18:11.120 | Nonetheless, I do agree that for data collection reasons

01:18:17.440 | and reciprocity reasons, I think it's, like I said,

01:18:20.760 | we're within our rights to require

01:18:22.800 | the divestment of TikTok.

01:18:25.320 | - Yeah, I don't, you know, it's actually, you said on this--

01:18:27.080 | - But don't make this into more than it is.

01:18:29.120 | Again, I think this whole disinformation narrative,

01:18:31.520 | by the way, you wanna know why they push it so hard?

01:18:34.080 | It's because our own intelligence community

01:18:37.520 | wants to be involved, and there are political actors

01:18:41.640 | in the US who wanna regulate, quote, disinformation

01:18:45.440 | on our own social networks.

01:18:46.760 | I'm not talking about TikTok, I'm talking about X,

01:18:48.960 | I'm talking about Facebook, Insta, and so on.

01:18:51.520 | And we saw this in the Twitter files.

01:18:53.360 | We saw what a cozy relationship

01:18:55.440 | the intelligence community had with Twitter.

01:18:57.760 | They were all up in there trying to control

01:19:00.920 | what legacy Twitter was allowing

01:19:04.480 | and what they were censoring. - They blocked the

01:19:05.320 | New York Post URL.

01:19:06.160 | They blocked the New York Post URL.

01:19:07.400 | And you've said on this program that you believe

01:19:09.240 | that could have swung the election.

01:19:10.840 | - Well, I think it was, I think that actually

01:19:13.680 | was genuine election interference.

01:19:15.240 | I actually said that I didn't know

01:19:16.400 | whether it could swing the election.

01:19:17.640 | I think that, but I think that--

01:19:18.720 | - Well, do you think it could have?

01:19:21.240 | - I think it was a story that deserved to be published

01:19:24.600 | so that, and distributed online,

01:19:26.800 | so that the public could take that into account

01:19:28.960 | when they voted, I mean--

01:19:30.480 | - And so if it had been widely received,

01:19:32.760 | do you think, I'm asking a question,

01:19:34.000 | you're not answering it. - How would I know?

01:19:36.520 | Okay, you're asking me a question

01:19:38.440 | that I personally think is irrelevant.

01:19:39.480 | How would I know whether the suppression

01:19:42.520 | of that story swung the election?

01:19:45.300 | How could I know that?

01:19:46.320 | The point is that the American people

01:19:48.720 | were deprived of information that they had every right to--

01:19:52.240 | - Most Republicans believe if that had come out,

01:19:55.000 | it would have swung the election.

01:19:56.520 | - Well, I can't prove.

01:19:57.840 | I can't prove that it swung the election.

01:19:59.720 | What I'm saying is that it was

01:20:01.200 | a type of election interference.

01:20:02.640 | Why was that story suppressed?

01:20:03.920 | Because 51 former intelligence officials

01:20:06.720 | who still maintain close relations

01:20:08.320 | with the intelligence community published a bogus letter

01:20:10.920 | saying that it was Russian disinformation, which it wasn't,

01:20:14.080 | and then that caused our social media sites to suppress it.

01:20:18.080 | So that, to me, is as concerning, if not more concerning,

01:20:22.960 | than whatever it is that TikTok's accused of.

01:20:25.740 | So I don't want social media companies being used

01:20:29.800 | by the intelligence communities of either China

01:20:33.140 | or the United States to swing or to influence our elections.

01:20:37.180 | And we need to be equally concerned about that

01:20:40.140 | as we are about supposed Chinese influence.

01:20:42.900 | - Just to go back in time,

01:20:43.740 | you probably remember the Willie Horton ad.

01:20:45.420 | I mean, that kind of sunk Dukakis, if you remember that.

01:20:47.620 | I mean, media and these ads can really have a big impact.

01:20:50.620 | So people should just look historically.

01:20:52.600 | There have been many moments where,

01:20:54.460 | whether it's Nixon sweating on TV or the Willie Horton ad,

01:20:57.840 | there have been many moments where video can do this.

01:20:59.620 | And I think they could be done even more subtly

01:21:03.120 | by the Chinese by just promoting certain videos.

01:21:06.860 | All right, let's move on to Bitcoin.

01:21:08.540 | Issue five, Bitcoin's back, baby.

01:21:10.700 | It hits an all-time high.

01:21:12.140 | 69,000, interesting number on Tuesday.

01:21:15.740 | 69,000 before dropping, sitting around 68K.

01:21:19.500 | As we're taping, there'd be 75 by the time you hear this,

01:21:23.060 | or 50, it's been ripping in 2024,

01:21:26.700 | up 70% since January 25th.

01:21:30.940 | There's two things going on here

01:21:32.540 | that you've probably heard about.

01:21:33.620 | The Bitcoin ETFs finally arrived in the US.

01:21:36.260 | They were approved by the SEC on January 10th.

01:21:38.900 | We'll get into them in a minute, Chamath.

01:21:40.820 | Obviously, if it's an ETF,

01:21:42.940 | it's super easy to buy them and sell them.

01:21:44.780 | They've been a total hit.

01:21:45.680 | BlackRock's Bitcoin ETF became the fastest ETF

01:21:49.060 | to ever reach 10 billion in assets.

01:21:51.340 | Also, as the technical crypto heads in the audience know,

01:21:55.180 | there is a halving happening in April.

01:21:58.660 | This happens about every four years at the current pace.

01:22:00.900 | Last one was May of 2020.

01:22:02.300 | When these halfings happen,

01:22:04.900 | the mining rewards are cut in half.

01:22:07.180 | This reduces the supply of new Bitcoins

01:22:08.900 | entering circulation and can cause some swings.

01:22:12.940 | Chamath, we released a clip, I think,

01:22:15.460 | just about your prediction, and you nailed it again.

01:22:20.340 | You said this would be a big year for Bitcoin.

01:22:21.860 | So your thoughts on being right.

01:22:24.920 | - I don't have much thoughts on that,

01:22:27.860 | but my two comments are that

01:22:29.360 | I talk to a lot of Bitcoin traders and folks

01:22:33.100 | that seem to have a very good pulse and touch

01:22:36.860 | on this market.

01:22:37.700 | I don't say that I do

01:22:39.020 | because I don't really look at it every day,

01:22:42.620 | but they seem to think that this thing

01:22:44.100 | is on a death march to 100K.

01:22:45.980 | I'm not sure whether that price is realistic or not

01:22:48.880 | in a year, but I will say that

01:22:50.420 | we're gonna get to a tipping point

01:22:53.980 | where everybody really talks about this.

01:22:56.660 | I still don't think we're there yet.

01:22:57.940 | I think we're just at the beginning.

01:22:59.740 | But when you see the inflows into these ETFs,

01:23:02.560 | Jay Cal, it's like a very big deal

01:23:04.400 | because it just allows every mom and pop individual

01:23:08.780 | to buy some to the extent that they want to own it

01:23:11.220 | or they want to speculate on it, whatever it is.

01:23:13.620 | So I think it's been a very big year,

01:23:17.220 | and I think that psychologically

01:23:19.160 | it's proven a lot of folks wrong

01:23:21.740 | and it's a setup for something really constructive.

01:23:26.280 | The other thing I'll say is that it's not just Bitcoin,

01:23:30.540 | but as goes Bitcoin, there are a handful of other things.

01:23:34.860 | People are now speculating that there's gonna be

01:23:37.320 | an Ethereum ETF that gets approved as well,

01:23:40.120 | because if you approve one,

01:23:41.920 | there's probably legitimate cause to approve a few others.

01:23:44.500 | So these things are becoming part of the financial fabric.

01:23:48.100 | And I think that that should not be underestimated.

01:23:50.460 | - And the SEC is still taking action

01:23:52.220 | against certain bad actors in crypto.

01:23:54.660 | There were a couple of those this week,

01:23:55.940 | but Bitcoin, Friedberg, is incredibly resilient,

01:23:59.120 | just on a technological basis.

01:24:01.140 | The fact that it hasn't broken down under stress,

01:24:04.860 | it hasn't had a denial of service type of attack,

01:24:08.680 | or a government hasn't been able to capture

01:24:13.220 | 51% of the mining, or some great amount of it,

01:24:16.680 | or just even be hacked in any way,

01:24:19.380 | you have to be impressed by the fundamental technology here,

01:24:22.180 | Friedberg, maybe you could speak to that level of success,

01:24:25.380 | that this thing is so stable

01:24:26.860 | and trustworthy and reliable to date.

01:24:29.900 | - It's got a great incentive model.

01:24:32.180 | As long as Bitcoin price remains high,

01:24:34.120 | the miners will still be there

01:24:35.220 | and the system will keep running.

01:24:37.540 | If Bitcoin price drops, transaction fees will decline,

01:24:41.960 | the value of mining will decline,

01:24:44.520 | and it kind of goes the other way as well.

01:24:48.440 | I think the real question is, in the last couple of years,

01:24:51.100 | have we really seen a change in Bitcoin

01:24:53.340 | being used for transactions or for commerce

01:24:57.600 | in any meaningful way?

01:24:59.060 | I think the answer is still likely not.

01:25:01.340 | - No, definitely not.

01:25:02.180 | - And it's really a stored value system,

01:25:04.100 | and it's become this kind of stored value asset.

01:25:06.300 | - Friedberg, this microplastics thing,

01:25:08.860 | we talked about it on the show,

01:25:10.860 | and since that time, I refuse to open plastic bottles.

01:25:15.020 | I'm doing all glass.

01:25:16.300 | I'm getting rid of all this goddamn plastic.

01:25:18.540 | I already did glass bottles in my house 'cause I'm cheap,

01:25:21.400 | and I like to fill it from my water filter,

01:25:22.860 | but we're uncovering more information.

01:25:26.360 | And then I saw this headline this week

01:25:28.080 | that microplastics are in our blood streams in some cases,

01:25:32.460 | and what the heck does that mean?

01:25:35.400 | - It's worse than that.

01:25:36.320 | - Team of scientists in Italy collected samples

01:25:40.380 | from patients that had plaque removed

01:25:43.340 | from their carotid artery.

01:25:44.580 | It's a kind of common cardiac procedure

01:25:48.100 | where you get plaque that blocks up in your carotid,

01:25:51.820 | they go in, they remove the plaque.

01:25:53.700 | So a total of 304 patients agreed to have the plaque

01:25:57.360 | that was removed from their artery submitted for analysis.

01:26:01.300 | And then what this team did is they took that plaque

01:26:03.500 | and they studied it to see

01:26:06.160 | how much plastic was found in that plaque.

01:26:10.880 | And they used a bunch of measurement techniques to do this,

01:26:14.960 | including electron microscopy and mass spec,

01:26:17.960 | 'cause it's really hard to find these molecules.

01:26:20.500 | And microplastics or nanoplastics, remember,

01:26:22.820 | are less than five millimeters in size

01:26:25.280 | with a mean level of 21 micrograms per milligram of plaque.

01:26:30.280 | Roughly one per 50 is the ratio of plastic to plaque

01:26:35.120 | that they found, which is really incredible

01:26:37.280 | because it shows that plastics,

01:26:39.020 | these little nano and microplastics are accumulating-

01:26:40.860 | - Incredible good or incredible bad?

01:26:43.140 | - Incredible bad, that these microplastics,

01:26:45.660 | these nanoplastics are accumulating in the human body.

01:26:48.640 | Now here's the scary part.

01:26:50.520 | They then did a follow-up 34 months later.

01:26:53.200 | The patients that had plastic in their blood

01:26:56.100 | had a four and a half times higher ratio

01:26:58.940 | or likelihood of having a heart attack,

01:27:01.540 | stroke or death from any cause.

01:27:04.100 | So all of these major health effects

01:27:07.660 | were four and a half times elevated

01:27:09.480 | in patients that had plastic in their blood.

01:27:11.580 | This was published in the New England Journal of Medicine,

01:27:13.580 | if I didn't say it.

01:27:14.420 | It really indicates that there is

01:27:15.700 | this kind of cumulative problem

01:27:18.060 | and that the cumulative problem is likely leading

01:27:20.940 | to really adverse health outcomes.

01:27:23.260 | And I'll just highlight one other paper

01:27:24.860 | from a team in Germany and Norway back in May of 2022.

01:27:29.140 | And this team tried to figure out how plastics

01:27:31.540 | are causing adverse health effects in the body.

01:27:34.260 | And they had a theory like let's put little microplastics

01:27:36.720 | or nanoplastics together with all the human cells

01:27:38.940 | that we know, shake it up and see what happens.

01:27:41.860 | And what they found were that these little plastic fragments

01:27:44.440 | were binding to dendritic cells and monocytes,

01:27:47.300 | key cells in the immune system.

01:27:49.500 | And when those cells were bound by plastic,

01:27:53.020 | they release these cytokines

01:27:54.540 | and the pro-inflammatory signals go through the roof.

01:27:57.340 | It causes the immune system to go high wire,

01:27:59.500 | increases inflammation and the cascading effects

01:28:02.220 | of that obviously can ultimately lead

01:28:04.300 | to many of the events that we're mentioning

01:28:06.180 | were measured in this set of patients in Italy.

01:28:08.620 | So again, we're just starting to uncover these effects,

01:28:12.380 | this concept that microplastics and nanoplastics

01:28:15.420 | that are accumulating.

01:28:16.240 | And let me just say these plastics are mostly PET,

01:28:19.120 | which is what we use to make plastic bottles

01:28:20.940 | that we drink water and drinks out of,

01:28:22.540 | and PVC or polyvinyl chloride,

01:28:24.820 | which is what a lot of our plastic plumbing

01:28:27.100 | and piping is made from.

01:28:28.700 | And so as little tiny bits of these plastic materials

01:28:31.220 | either are exposed to sunlight and break off

01:28:33.140 | and end up in our water and food supply and we consume them,

01:28:36.420 | they are slowly accumulating in bodies

01:28:39.060 | and they may be driving inflammatory response,

01:28:41.200 | they may be driving adverse health outcomes.

01:28:43.540 | We're really kind of tip of the iceberg

01:28:45.160 | and really studying this, understanding and analyzing it.

01:28:47.620 | But here's another really interesting empirical data set

01:28:50.680 | that highlights that this really is a pretty significant,

01:28:54.300 | half the patients had it and of that half,

01:28:57.280 | they had a four and a half times higher chance of dying

01:28:59.680 | or having a heart attack or a stroke

01:29:02.000 | in the 34 months that followed.

01:29:03.980 | - Yeah, the thing that that study said, which was nuts,

01:29:07.880 | is it looked like the nanoparticles,

01:29:10.640 | the nanoplastics and microplastics

01:29:12.280 | were effectively acting as scaffolding for plaque.

01:29:15.080 | So in almost like it was a shim that allowed it to grow.

01:29:20.080 | The question is, would it have grown faster

01:29:22.740 | than it would have otherwise?

01:29:24.100 | That's even scarier.

01:29:25.260 | So I did not like reading that paper.

01:29:30.260 | That really freaked me out.

01:29:32.140 | Really, really freaked me out.

01:29:34.900 | - I was drinking water from plastic water bottles this week

01:29:37.520 | and every time I drink water out of a plastic bottle now,

01:29:40.260 | I'm like nervous every time I take a sip.

01:29:43.660 | - You're not supposed to double the risk

01:29:45.400 | of all cause mortality by drinking a Fiji water.

01:29:48.780 | You know what I mean?

01:29:49.620 | - Four and a half X, four and a half X.

01:29:51.340 | - It's crazy.

01:29:52.260 | - In 34 months.

01:29:53.420 | - Think about the cumulative effect over time.

01:29:55.420 | - Over time.

01:29:56.820 | - Imagine drinking water out of a plastic bottle,

01:29:59.040 | thinking you're doing the right thing

01:30:00.260 | and then trotting over to the recycling bin.

01:30:02.860 | And you do that for 20 years, you may be killing yourself.

01:30:07.520 | - Well, there it is.

01:30:10.440 | - So wait, what's the, if you, if the water's--

01:30:13.420 | - Glass bottles, you must use, you cannot use plastic.

01:30:18.060 | You just cannot.

01:30:18.900 | - It's over, no plastic, no plastic.

01:30:20.780 | - You can't, you gotta stop.

01:30:21.620 | It's over, it's done.

01:30:22.440 | - We're drinking, it's done.

01:30:23.280 | Glass, cans, good.

01:30:24.840 | Plastic, stainless steel, stainless steel is fine.

01:30:28.580 | - Just like I mentioned when we talked about this

01:30:30.140 | a few weeks ago, the carbon footprint,

01:30:33.020 | the environmental cost, the cash cost is much higher

01:30:35.600 | with all these alternatives to plastic.

01:30:37.380 | So there are big challenges with respect

01:30:39.600 | to having some big, massive response to plastics

01:30:43.020 | used in our supply.

01:30:43.880 | But, you know, living in the luxury world

01:30:46.540 | that we all get to live in, we get to have that choice

01:30:48.480 | and we'll make that choice.

01:30:49.940 | But it's a real problem for humanity

01:30:52.240 | because plastics are so ubiquitous in so many things

01:30:54.940 | and they've enabled affordability of consumer goods.

01:30:58.680 | - This is such bull (beep)

01:30:59.880 | Honestly, like all you have to do is have glass bottles

01:31:02.500 | or carry a water bottle with you.

01:31:04.380 | Like I have a Contigo one I like, I carry it with me,

01:31:07.980 | I empty it, I fill it. - No, but you can't.

01:31:10.140 | - And I have a water filter in my house

01:31:12.660 | and we fill water bottles and put them in the fridge.

01:31:14.820 | - No, but what if you like yogurt?

01:31:15.960 | Yogurt comes in a plastic container.

01:31:17.400 | There's all kinds of stuff.

01:31:18.700 | You can't avoid plastic.

01:31:20.420 | That's what's so scary.

01:31:21.260 | - We try to in our house. - That's what's so scary.

01:31:22.920 | - We do have a French yogurt that comes in glass bottles,

01:31:25.100 | but yes, it is hard.

01:31:25.940 | - Ah, yes, we do the French yogurt in glass bottles.

01:31:27.700 | - It's so good.

01:31:28.540 | - It's so good.

01:31:29.360 | - That's delicious.

01:31:30.200 | - Wait, where do you get it from?

01:31:31.760 | - There's a French yogurt that comes in glass bottles.

01:31:33.760 | - A French yogurt.

01:31:34.600 | - But how do we get water in plastic?

01:31:35.760 | - It's called like Le Benoit or something, Le Benoit.

01:31:38.760 | - The water, you just install a filter system,

01:31:40.860 | which you have at the mausoleum

01:31:42.820 | and just fill huge glass bottles.

01:31:45.860 | - Just have your staff fill glass bottles

01:31:47.860 | and put them in the fridge and don't throw them away.

01:31:49.860 | And give your kids like some of these thermoses or whatever,

01:31:52.260 | but don't have the water bottles.

01:31:54.260 | Like, you know, it's so funny.

01:31:55.460 | It's been a disaster in the poker game in some ways.

01:31:57.940 | We got rid of it and there's been way more broken glass.

01:32:01.740 | People knock over the, you know, the side tables.

01:32:05.260 | I get it.

01:32:06.100 | It's been a huge pain, but I will not go back.

01:32:08.700 | - No, it's the right thing to do.

01:32:10.580 | - Absolutely. - It's the right thing to do.

01:32:11.620 | - Okay, well, if Tamath is endorsing this,

01:32:14.520 | I guess I'm gonna take it seriously.

01:32:16.880 | - I think you gotta do it.

01:32:18.100 | - My opinion doesn't matter, Sax.

01:32:19.620 | I'm not saying you're wrong.

01:32:20.940 | I'm just saying that your threshold

01:32:22.860 | for becoming concerned is lower.

01:32:25.980 | And then if it hits Tamath's threshold, which is higher,

01:32:29.540 | I'm gonna take it more seriously.

01:32:31.940 | - As your bestie, I would like you to stop using plastic

01:32:34.900 | for the rest of your life.

01:32:35.980 | - Okay. - Please.

01:32:36.800 | - If you're in a position to do it,

01:32:37.740 | I would ask you to not do it.

01:32:39.420 | - All right, everybody.

01:32:40.260 | - Can all you (beep)ers come back to the Bay Area

01:32:43.060 | so we can see each other? - Yes, absolutely.

01:32:44.340 | I miss you guys. - I heard this.

01:32:46.180 | - Love you guys.

01:32:47.020 | I miss you.

01:32:47.840 | I miss you guys, too, for the Sultan of Science,

01:32:50.840 | the Chairman Dictator, and the Rain Man, David Sax.

01:32:54.880 | I am the world's greatest moderator.

01:32:57.280 | Okay, love you guys.

01:32:58.160 | I gotta go. - Love you.

01:32:59.000 | - Love you. - See you next time.

01:32:59.820 | Bye-bye.

01:33:00.660 | (upbeat music)

01:33:03.240 | ♪ I'm going all in ♪

01:33:05.240 | ♪ Rain Man, David Sax ♪

01:33:06.560 | ♪ I'm going all in ♪

01:33:09.220 | ♪ And it said ♪

01:33:10.060 | ♪ We open sourced it to the fans ♪

01:33:11.520 | ♪ And they've just gone crazy with it ♪

01:33:13.380 | ♪ Love you, Westie ♪

01:33:14.220 | ♪ I'm the queen of Kinwan ♪

01:33:15.480 | ♪ I'm going all in ♪

01:33:16.640 | ♪ What your winners like ♪

01:33:18.340 | ♪ What your winners like ♪

01:33:20.580 | ♪ What your winners like ♪

01:33:22.240 | ♪ Besties are gone ♪

01:33:23.320 | ♪ That's cold 13 ♪

01:33:24.980 | ♪ That's my dog taking it ♪

01:33:26.200 | ♪ Oh, it's your driveway ♪

01:33:27.040 | ♪ Sex ♪

01:33:27.860 | ♪ Oh, man ♪

01:33:29.980 | ♪ Oh, man ♪

01:33:30.820 | ♪ My appetizer will meet me ♪

01:33:32.100 | ♪ We should all just get a room ♪

01:33:33.580 | ♪ And just have one big huge door ♪

01:33:34.980 | ♪ Because they're all just useless ♪

01:33:36.400 | ♪ It's like this like sexual tension ♪

01:33:37.800 | ♪ But they just need to release them out ♪

01:33:40.480 | ♪ What you're the bee ♪

01:33:42.300 | ♪ What you're the bee ♪

01:33:44.220 | ♪ Bee ♪

01:33:45.780 | ♪ We need to get merch ♪

01:33:46.740 | ♪ Besties are back ♪

01:33:47.580 | ♪ I'm going all in ♪

01:33:52.580 | ♪ I'm going all in ♪

01:33:58.200 | [BLANK_AUDIO]

E169: Elon sues OpenAI, Apple's decline, TikTok ban, Bitcoin $100K?, Science corner: Microplastics

Chapters