What's the story with the California wealth tax? Can somebody explain this to me? Okay. So, the SEIU, the Service Employees Union, filed a ballot initiative, which means a direct-to-voter vote to amend the California Constitution to introduce a one-time billionaire's wealth tax where billionaires, anyone who has assets over a billion dollars net of their debt, has to pay a one-time tax of 5% of their net worth, including their private stock, including their real estate. 5% of their net worth, not of their income, of their net worth, the entire net worth, onetime payment to the state of California, and then there's an allocation on how that money will be spent, but it's a onetime billionaire tax. Now, it is very likely that this sort of an amendment to the California Constitution is not constitutional and actually cannot be made and will not actually go into enforcement even if the voters do vote to approve it in both a federal and a state level based on this concept of uniformity which is that you have to tax everyone equally except for the case of an excise tax which is like income or a transaction. You're allowed to tax disproportionately based on the size of the income or the size of the transaction. But if you're going to tax on property, if you're going to tax on an asset, you have to tax everyone uniformally.

So it is likely not going to go into effect if it does pass. However, it is very likely the case that the SEIU is simply using this as a baiting mechanism to get people to stand up and denounce it and then they will be in a position to attack those people and destroy them and use this effectively as a political fodder for this next election cycle. That's what it seems like the true kind of motivation right now.

Well, we don't know who's going to come out against it, but there's an effort to try and get top Democrat officials in the state of California to say, "This is silly. If you do this, people will leave the state." yada yada. So, that's kind of a quiet underway effort. But I don't know why the citizens of California, the majority of citizens of California would not vote for this. Why? Who wouldn't want to tax the billionaires 5%. Come on. Like, the way that it's written, it says, "Hey guys, we're 30 billion in the hole and there are 200 Californians that control two trillion dollars. We're just going to ask them to pay a onetime fee of 5%." And I don't see how anybody would say that doesn't sound unreasonable at the ballot box.

And then the people that step up against it and are vocal against it and point out like hey in France when they did this they lost like 40% of their revenue because all the wealth left the country. The reality is that this sets it up to go through the legislature because if it goes through the will of the people and it gets overturned as you say Freeberg then if you're legislatively smart then you'll actually push it through the state senate. And then it will not get vetoed like because then it's like hey listen it's clear that the people want this. So I think that you'll have some kind of progressive taxation system that conforms to the law.

I mean they're already trying to extend Prop 55, which is the progressive tax for people making over a million dollars. They're going to get that passed. That's going to be this incremental tax on income. But the onetime wealth, I think the million dollars thing, I think that's harder to hunt because there's too many people that that touches. A million dollars today in 2025, not to be glib, is just not what it used to be. But a billion dollars does cut off most people except for a couple of hundred. That is true. And I think that for example, it's very reasonable to then charge a 10% excise tax on selling appreciated stock, right? Why not? There's all kinds of ways that you can get billions and billions and billions of dollars. So, I don't know. I think that this is more of a trial balloon to say, can we draw a clear line between 200 Californians and the rest of California? Yeah. And to the extent that that bright line becomes visible and it's okay, people are going to go ham. They're going to try to get as much as they can.

The reality is, as we all know, I mean, Larry Ellison left the state. Many of the founders CEOs who have built large technology companies in California. Elon left the state will eventually at some point break and say, "Okay, I'm moving my company out of state and I'm leaving the state and I'm bringing the employees with me and I'm bringing all of the economic value of this business with me." And people will never learn that lesson because it's so much easier to sit in front of a voter and say, "Hey, should we tax these 200 people to give you better benefits?" 97% of people will say, "Absolutely." Very few people will sit and think about the consequences of what's going to end up happening.

One of the motivations for this bill, and this is why it's being proposed by the SEIU, is that there are these massively ballooning pension benefits and pretty significant increases to the pension programs for both private and public pension funds in California, which has actually become a very visible liability for the state and for some of these private pension programs. and they're trying to fill the pension hole, which we've talked about in the past, but there is a multi-trillion dollar unaccounted for pension liability in this country that's going to have to come from somewhere. You're either going to have to print the money because the federal government's going to step in and fill the hole in all these pension obligations or they're going to have these massively progressive tax programs to try and fill the hole. And if and when they do, as we all know, there will be an economic cycle that will be pretty nasty, which is all the value will leave that jurisdiction and move elsewhere. But let's see.

The crazy thing with this, the other like I read it because I was like, "Oh my god, what's going on?" The two things that they they obviously got somebody very clever to draft it because any Roth IRA over 10 million counts. And normally in these wealth calculations, you you keep your deferred retirement accounts off the table. They're typically not included. So folks, and I'm not going to say who they are. we all know who have tremendously appreciated Roth IRAs those are included and then the other thing is that if you actually did any tax structuring the real valuable tax structuring is where you set up these trusts in Wyoming and North Dakota and you do these interparty loans where you can lever up 10 20x so you can transfer billions and billions and billions of dollars out of state but then you have these obligations those are negated and they don't count so all that tax structuring goes out the So you can get into a very difficult situation here where they're like, "Hey, you owe us $500 million, a billion dollar, $2 billion, and the only way to pay it is to have an IOU to the state of California." Which is crazy.

A wealth tax has been tried in many places at many times it always backfires because whatever the tax benefit is that you get for the state it's greatly outweighed by the economic depression that you get by the wealthy people the job creators companies leaving. And as soon as you cross that line of going from no wealth tax to any wealth tax, enough people of wealth can see the tea leaves, they can see the writing on the wall that they have to leave. And that's why I think even if they say this is a one-time thing, we all know that it won't be one time. If they get away with it, it'll become a regular thing. If it's to plug a deficit, they're going to run deficits every year. Exactly. And this isn't even to plug an emergency situation or an unfunded liability like some one-time thing. This is just a regular operating exactly. So they will have no incentive to fix their mismanagement of the state and their deficits and all that kind of stuff.

And it's not just going to be billionaires eventually the line will of course get pushed down be gone. The billionaires always like like the income tax in the US. I think it was a 1% income tax originally and it was like just a onetime thing for wealthy people and then it became a smaller thing for you know lower income people and then eventually as we all know every person has to pay a tax every property has tax and so on. I mean these are this is the problem with government.

Here's what I wonder about is, you know, what are guys like Jeffrey Katzenberg or even Ari Emanuel thinking about right now because they're kind of the higherups in the Democratic party behind the machine, sort of the oligopoly that kind of runs the machine. I think there are these guys who are very, very wealthy who think that they can control the machine well enough that they basically are still in control of this thing, right? That in other words, that the tiger won't eat them, right? The tiger is socialism. You know, they think they've got the tiger under control enough that it won't eat them. But I don't think they do. Maybe they don't. Maybe this is the tiger breaking loose. And I think you pointed this out that there was an attempt in the legislature last year to pass a wealth tax and it was quietly killed behind the scenes. And I actually think that Gavin Newsome might have something to do with that because he has presidential ambitions. So he can't let the state go full socialist. But you just kind of wonder, okay, well, if these guys lose control of the strings they have to control the beast of socialism, does the whole thing just spin out of control?

Just to let people know about the France situation back in I don't know 2011 2012 they did get rid of Gerard Depardieu which was kind of a win but Bernard Arnault said he was going to go to Belgium and said it was a clerical error and he unwound it but that was a clear signal the richest man in France. He's basically the entrepreneur LVMH together. I mean, it's their biggest company. It's the one that does all the luxury goods, all the craft goods that they're so famous for. I mean, yes, him threatening to leave France is, you know, accidentally filing paperwork. Oops. What an accident.

Here's your look at uh New York City. New York State tax 10.9% city 3.876% and the 2% tax put you at 16.8% for living in New York is 17%. I mean if you were making $10 million a year is it worth $1.7 million? You could get a plane. You could live in Florida. You could come to New York 150 days a year. There's really five good months in New York. the fall, the spring, and that's about it. You know, you go see the tree at Christmas, but it's cold. Well, that's not realistic for most people. And especially if you have kids and you care about them, you'd like them to be rooted somewhere. You're not going to shlep them around every month to arbitrage taxes. Yeah. Well, I mean, I do think they're they're going to test that at 17%.

This is the free rider problem that we have is no one's going to want to stand up against it and the thing will just kind of pass. By the way, if you're a billionaire CEO of a public company in California, you have everything to lose to stand up and oppose it. Your employees will run. Your shareholders will attack you. You'll look awful in PR. So, everyone's going to sit quietly and start looking at houses on Zillow in Austin or Miami and be like, "Where should we move to next year, honey?" You know, like that's the conversation that's going on.

Didn't you say it was retroactive? It's retroactive to 2026? So, if it passes, you have three months. Two months. Yeah. But again, I don't think it passes muster with the constitutional reads. Even if Friedberg is right that there's a good chance that it'll be found unconstitutional, how many years in the course is that going to take? And who's going to stick around waiting for that? In fact, the rational thing to do is pull up stakes before January 1st and leave right now.

There's a part of the bill that they could cleverly use which is called this ODA which is effectively this IOU mechanism and they could essentially say when these assets transact you owe us 5% on an excise basis and by the way there's an attestation that you have to file you have to file a legal document and this was quite well written in there which said you must attest that you have less than a billion dollars. Okay, now what? Okay, then I have to attest that it's more and then I have to know how do you even mark your whole portfolio to market if you have a lot of privates? They do not allow discounts. They do not allow liquidity discounts. It says if you are a reasonable buyer and a reasonable seller, you have to transact this at market price. So, for example, imagine you owned a sports franchise and the sports franchise, if you sell a minority share, you're typically selling it at a discount off the table. If Forbes says it's worth 10 billion and you own 10%, that's a billion dollars for the purposes of this calculation. I'm going to pay $50 million to keep it even if you paid 50 million to buy it.

Big breaking news this morning. Huge scandal in the NBA. The FBI just arrested 30 people in a sports betting and gambling probe. This hardly seems real. Chauncey Billups, who is the current Blazers coach and was just introduced into the Hall of Fame, got pinched for a poker game he was running allegedly with the mafia that was rigged 17 different ways allegedly to Sunday. Terry Rozier allegedly is a point guard for the Miami Heat. He told his friends, this is crazy, you know, in the over-unders, you know, hey guys, bet the under on me uh in rebounds cuz I'm going to uh take myself out of the game with an injury allegedly. And uh his friends allegedly made 200 grand off this. Okay, just allegedly for the whole goddamn thing. Uh this is going across 11 states and a bunch of crime families.

I think it's crazy. I think you're seeing a lot of these trends converge all at the same time. Meaning, you have the emergence of all of these prediction markets. you have a lot of data science and AI being used that shows that there's a lot of odd behaviors. So, it really was the squares versus the sharps. And if you had the inside edge, you were just printing money. Now that all of that is becoming more transparent, there's a lot less margin. Then what happens is you have these laws passed in the 11th hour. There was a an important gambling law that was inserted into the big beautiful bill that has implications to all of this. And now you're seeing the feds. The crazy thing to me is a press conference where Kash Patel is talking about this. I mean, that's like serious business when the FBI director is front and center talking about all this. So, I don't really know what it means to be honest. I was shocked at the scale of it and I was shocked that it's on the radar of the feds. This took I thought this is like pretty typical tickytacky stuff, but clearly there's something bigger. I don't know exactly what that bigger is, but something is happening where all these markets are smashing together. There's just a big cleanup effort going on.

Are these people too dumb to understand that their $10 million contract to play in the NBA every year or $20 million contract is more important than your friends betting the under or over. And how dumb are they? I mean, to not know that the people running a sports book would look for weird action. like why is one player getting $200,000 on their overunder for rebounds and the other players are getting 20,000? What are your thoughts here?

I think gambling generally as we call it should be decriminalized and I don't like this state-by-state setup with gambling. I think we should have a federal regulatory body to oversee, monitor, and the problem is you have state gambling commissions and we have a state-by-state kind of patchwork of regulatory authority that makes it very hard to standardize, track, and provide also guidance and feedback. I would much rather see this all kind of get handled at the federal level and and better organized. To Chamath's point, this is not going away. People love to bet on stuff. They love to gamble. This is part of sports. This is part of the culture. They're not going to just turn it off. The poly market model is the best model because it creates a market and so as information flows in that market will dynamically adjust and everyone will get a more fair price.

Did you see the regression that they did on the poly market trades and how well they're in the money? Basically what it showed is like the front money is the sharps, the back money are the squares, but you have to fade the trade in the first week. So there's a very scientific method where if you want to make money on poly market, it became pretty clear. Jason, look at this. I sent Nick the tweet, but this is incredibly systematic. This is over many many many markets. But basically, 89% accurate one week out, but in the final four hours, it jumps to 95, which means that if you follow the sharps along this pattern, you're going to make money. Poly Market actually has the news before the news does. This is one of the most like powerful outputs of Poly Market is they're actually getting a read on what's going on in the world before the media recognizes it, before the public recognizes it. Because when you put money up, it actually turns out that when people have incentives, that market will find the truth.

Somebody needs to build the app that makes all of these things fungible. And by all what I mean are cryptocurrencies, betting markets, equities, and options markets turning into. Yeah. And the reason is there's just no reason to go to nine different sites and have nine different accounts. And the most important thing is to do KYC and AML across nine sites to get access to liquidity, credit, and margin. You'll want to do it once. And then you'll want to have a large pool of capital that you can trade across anything. So I can go long Nvidia but I can also go short the nicks and then I can own some Bitcoin all in the same trade.

Now to the earlier question, if we end up there where poly market does become the truly liquid market across all of these kind of predictions all of these assets then a lot of what we are seeing with respect to insider trading insider information becomes much more apparent. So, the problem with the sports betting is that there's a one-sided bet. The casino sets the odds or, you know, whomever is setting the odds, and then you're either taking one side or the other. And so, if you have the insider information, you're taking the side that creates an arbitrage opportunity for you. But if you were to do that in a liquid market where there's someone taking the other side in a dynamic way, then the market very quickly moves because of the inside knowledge you have. And that inside knowledge is now reflected in the underlying asset price in the underlying odds that you get for that bet. And so Poly Market actually brings truth and transparency to what is currently an insider arbitrage opportunity and it may actually solve some of these fundamental problems in in gambling.

Let's talk about this Amazon outage. Tough week for Amazon. They had this huge outage in the beginning of the week and then they had a bunch of leaked documents about their plans for uh jobs. And uh Monday, massive AWS outage, 2,000 companies, 4 million users unable to function on the internet for half a day, 15 hours, 20 hours. Uh and then on Tuesday, internal docs viewed by the New York Times showed Amazon plans to not hire uh 600,000 plan jobs uh because of robots by 2033.

So this isn't they're planning on laying off 600,000 workers, but rather they're just pulling back their hiring plans and ramping up their robotic plans, which you would expect. uh and uh their goal according to these internal leak documents is to automate 75% of warehouse operations. We talked about this the last couple of weeks.

I think the AWS story is interesting in terms of its implications for the clouds. There's effectively three major cloud vendors that compete with one another. AWS, Microsoft, and GCP or Google Cloud. And I'll just give you these numbers. AWS is $124 billion revenue run rate. Microsoft 120 billion. And Google Cloud 54 billion. But AWS, which is slightly larger than Microsoft, is only growing 17% year-over-year. Microsoft 26% year-over-year. And Google Cloud is accelerating at 32% year-over-year. And some say getting closer to 40% growth rate.

The big thing I hear from partners and enterprise customers of these cloud services is that many of them if not all of them as they scale up move to a multicloud model. So none of them want to be dependent on a single cloud. Many folks started on AWS because AWS was the OG. Back in the day when I was running Climate Corp, I was the largest EC2 user on AWS for about a year and a half, which was their elastic compute cloud service. We were running all these models back then. So I knew that service very early on and it was very unique. It was very powerful and so a lot of companies that are old school established themselves on AWS very early on. But the outage that happened this week, I think starts to highlight for folks that they can't and shouldn't have a dependency on a single cloud service provider and will only accelerate the diversification of companies into the other clouds. And so I do think this is actually a very beneficial situation for Microsoft and GCP and to your point perhaps even Oracle in terms of giving those sales teams which are very aggressive a hard story to go and sell for and say guys you don't want to just sit on AWS in case this happens again we've got better infrastructure we're more reliable etc than these other guys so come and move over to us and that might be a little bit of a naive simplistic kind of reductive way to think about what happened this week but I we are seeing the smaller competitors accelerate and I think that this might be another kind of moment of acceleration for those folks.

I think there are two markets there's the AI market then there's the non-AI market. In the nonAI market, everybody has everything. It all looks effectively the same. There's certain products and services that are unique to Azure versus GCP versus AWS, but by and large, the market is big enough and important enough that you'd have to be pretty insane to take a single vendor approach. And so what typically happens in these markets is that they start off really small. One person has all the share. And then as the market becomes very valuable and very big, everybody diversifies because it's a riskmanagement thing. And these things flow into the disclosures you have to make as a public company. And if you didn't have that diversification and something bad happened and it impacted your business, you could get sued. So there's all these reasons why eventually all these three big companies will converge effectively roughly a third, a third, a third. In nonAI, it's a third, a third, a third. It will it'll take circuitous paths, but that's where we'll end up. I think the way it works in AI is that you initially right now we're in this early phase where there's two paths. Path one is you need a specific model and it's relatively well integrated using a specific subsidized form of hardware on one of the hyperscalers but eventually you'll get more of that abstracted away as it gets pushed into the infrastructure so that you have less dependence on one model. There's a lot of work that has to get done and a lot of in-memory infrastructure that is not yet built that has to exist but once that exists it'll be easier for all of us at the application level to view these models a little bit more fungibly. And then at the bleeding edge you'll have the folks that basically give you some form of a hypervisor or virtual machine or the bare metal and that's where the neoscalers are doing really well but I think my point is that in any important market in compute in technology where there really isn't much of a differentiation I think you'll end up with these hyperscalers at a third a third a third now if one model is way way better and it's only on one of the clouds because Google writes a big check or Amazon writes a big check I could see that swaying the AI share but in the absence of that I think cheaper faster better is sort of the end destination for everybody.

When you look at this Amazon stuff that came out, I just wanted to point out a couple of things, it's not just that they're not hiring these 600,000 jobs. It's that they are in fullblown crisis preparation for this. They they have crisis teams writing up how to handle this and be a good corporate citizen. And they're talking about having parades and paying for toys for tots. And they're even trying to get the executives to say things like cobots as opposed to robots. Let's not call them that. Let's call them co-workers and co-bots.

And when you look at this, just to open up the aperture here right now, Walmart and Amazon are the number one and two employers in the US. 2.1 million people work at Walmart, over a million at Amazon, and three million people, as we know, work in taxis, Uber, Door Dashers. All those jobs are at risk. And we talked about this back in June when Andy Jassy telegraphed all this in a blog post where he said the next few years we expect that this will reduce our total corporate workforce as we get efficiency gains from using AI extensively across the company. They believe that they're going to have significant job displacement.

Elon replied to my tweet and said, "AI and robotics replace all jobs. Working will be optional like growing your own vegetables instead of buying them from the store." And Senator Bernie Sanders came out and said, "I don't often agree with Elon Musk, but I fear that he may be right when he says AI and robotics will replace all jobs." So what happens to workers when they have no jobs or income? AI and robotics must benefit all humanity and not just billionaires. And I think this feeds into your story for the last two years on this podcast, Friedberg, which is the rise of socialism. These things and Bernie Sanders being the standard bearer for democratic socialism. These things are starting to come together.

I think that you are characterizing AI automation and technological progress as the core driver of the socialist influence and what I would argue is that the actual core driver of the socialist influence is the fact that we put in place a lot of people into government, passed a lot of laws that caused an increase in spending because we promised people that the government would do more for them over the last 40 years. That is not possible in a true market- based system. And so by telling everyone, hey, we're going to make sure you get better jobs. We're going to make sure you all get housing. We're going to make sure you get education. You cannot actually get a government to effectively do that because what ends up happening is the government inflates the cost of those things and the market doesn't actually work. So the truth is this is now like all other things a scapegoat for the true cause of the socialist movement which is that government has become too big, too unwieldy and its natural inefficiency has distorted markets to the point that there is maybe no point of return anymore. And people will not see that. They do not see it and they're going to look for reasons and they're going to look for scapegoats and they're gonna say, "Oh my god, look over there. There's a robot. That's the reason I'm losing my job. Oh my god, look over there. There's a rich person that works at a pharmaceutical company. That's the reason I can't get healthcare." Or an immigrant took my job, right? Is the one from the last 20 years. And so, fundamentally, I think that people aren't willing to and they're not going to see the true cause because there's no one that runs to go work as a politician that is going to raise their hand and say government is the problem. No one ever has gotten elected in a democracy doing that. So the natural course of things over 250 years is that people raise their hand and they say I'm going to give you more and I'm going to use the government to do it and then they go into the government. They make the government bigger and as a result of making the government bigger, the government is spending more. The dollar goes down. The performance of the services goes down and fundamentally we end up in a socialist spiral.

Tesla reported their earnings on Wednesday. Record revenues, 28 billion, up 12% year-over-year. massive amounts of free cash flow. Four billion I think they're up to 40 billion in cash uh which is always great when you're going into uh some big capital intensive projects like Optimus and like self-driving. Downside operating profit fell 40%. uh stock dropped a bit 4% but bounced back and uh on the earnings call Elon emphasized the importance of his trillion dollar pay package which will give him just but 12% uh additional stake over the next 10 years if he hits absurd targets that would make everybody who holds the share uh shares in the company extremely wealthy uh and they would benefit uh more than Elon himself and here's his quote my fundamental concern with how much voting control I have at Tesla is if I build this enormous robot army. Can I just be ousted in the future? I don't feel comfortable building that robot army if I don't have at least influence over it.

And he called Glass Lewis and ISS corporate terrorists. These are the people who vote on behalf of passive index funds for things like who's on the board of Tesla. Vote for Elon's pay package will be number six. Poly market thinks it's going to pass.

The first is at the foundational technology layer. And Nick, I sent you this tweet, but it's what he said about AI5. I've made these comments before, but he had these multiple efforts with Dojo and other stuff that he merged into one unit. And the quote is pretty incredible. We're going to focus TSMC and Samsung on AI5. The chip design is an amazing design. I have spent almost every weekend the last few months with the chip design on AI5. By some metrics, it will be 40x better than AI4. We have a detailed understanding of the entire stack. With AI5, we deleted the legacy GPU. It basically is a GPU. We also deleted the image signal processor. This is a beautiful chip. I've poured so much life energy into this personally. It will be a real winner. Why is AI5 so important? What AI5 is is the building block of a system that I think you'll start to see not just in the cyber cabs but also in Optimus.

The second thing was what he said about his energy business which I think is the critical adjunct to believe robotics and autonomous cars. If robotics and autonomous cars work, what you really need is an energy business beside it that is humming and on all cylinders. Why? It's how you make LFP battery cam that will be the limiter. Energy will be the limiter. But what he's showing, and Nick, I sent you this tweet, is that business is just on a tear. It's printing $ three and a half billion dollars a quarter, and its operating margins, an energy business, 30%. And so what you're going to see are battery packs of all shapes and sizes, the huge battery systems that's going to go into data centers, but then all the way down, I think, to the small LFP cam that he's going to need to power all these things.

And then the third thing is his comments on cyber cab which is that this thing is just going to be a shock wave. So I read all of those things and I was very bullish. I think that he is humming on all cylinders on the critical layers of the stack that he needs to build this next version of Tesla. My concern I think there's a real concern that I have that this vote is going to go down to the wire. I think that ISS and Glass Lewis, I think that these organizations are pretty broken. I think the way that they make decisions are hard to justify.

Can I just shine a spotlight on one of those points that you made with these proxy advisory services? So I think for years people have wondering why did corporate America go so woke especially in the early 2020s where they created all these DEI departments and you know they didn't have to do that and a big part of the reason is that those initiatives came from Glass Lewis and ISS. I think Elon's jokingly called ISS ISIS but basically what happens is they make recommendations for how shareholders should vote on different resolutions and the index funds basically just defer to them for whatever they should do. So they effectively control or almost control the voting for all these board level resolutions that every public company has to make. And so they've been the ones who've been imposing all these DEI requirements, all these ESG requirements, if you're wondering where those things came from, because just these two companies, which no one's ever heard of, they were captured a long time ago, meaning they were captured by the woke crowd years ago. And so this has really been the root of why corporate America has gone woke for a long time.

Our zar of AI, our civil servant study reveals AI models are showing hidden biases in how they value human lives. Back in February, center for AI safety published a study showing that LLMs have well-defined biases for race, gender, ethnicity. The title of this study, utility engineering, analyzing and controlling emergent value systems in AIs. Paper found that open AIS GPT40 favored people from Nigeria, Pakistan, India, Brazil, and China over those from Germany, the UK, and US relative to Japan as a baseline. Here's another one. Valuing people with Joe Biden as a baseline. Bernie Sanders, Beyonce, Oprah, all better. Paris Hilton, Trump, Elon Putin, all worse.

Twitter users and AI analysts called Artothereum decided to update the papers problems with new LLMs. Consistently ranking white people last, Claude Sonnet, GPT5, uh, and consistently ranking white Western nations last as well. your thoughts here on the biases we're seeing in some of these models and these early studies to track it. Yeah, I think what the paper purports to show is that almost all of these models, except for maybe Grok, view whites as less valuable than non-whites and males as less valuable than females and Americans as less valuable than people of other cultures, especially global south. And if the results are true, it does look like these models are pushing a woke bias that makes that sort of distinction between oppressed and non-opressed peoples and gives more worth or weight to the categories that they consider to be oppressed. This does appear to show significant bias.

The question is how does this bias get into the models and there's a few different possibilities one is that the training data is just biased like if they're training on Wikipedia we know that Wikipedia is massly biased because they literally have censored the leading conservative publications from being citations and sources in Wikipedia. The co-founder recently just revealed that that they don't allow Larry Sanger just said that they don't allow the New York Post for example to be a source in Wikipedia or a trusted source. So if AI models are training on Wikipedia, that's a huge problem because that bias will now cascade through. And same thing if they're training on say mainstream media or leftwing media but not right-wing media and they don't have a way of correcting that. So that's one source of potential bias. Another source of potential bias is just the engineers these companies, the employees and the staff do tend to be I mean if they follow the trend of other tech companies are 90s something% Democrat versus Republican and that does over time trickle into these models. And then finally, I think another source of potential bias is DEI. And we saw that when you remember this is like a couple years ago when Google launched Gemini and that that problem with, you know, Black George Washington. That was because you had DEI advocates in these meetings and that somehow trickled into the the model.

Now, one thing that's very concerning is that the push for DEI to be inserted into AI models, which was explicitly part of the Biden executive order on AI, has now moved to the state level, and they're just doing it in a more clever way. They've rebranded the concept. They call it algorithmic discrimination. We talked about last week how Colorado has now effectively prohibited models from saying something bad about a protected group. And that list of protected groups is very long. It's not just the usual groups. It even includes groups who have less proficiency in English language. That basically requires DEI. I mean, you have to have a DEI layer to prevent that. So I think that we've gone from models being required to promote DEI, which is what the Biden executive order on AI did explicitly, to states now prohibiting algorithmic discrimination, which is effectively a backdoor way of requiring DEI models. So that's a whole other area of potential model bias that I'm very concerned about. And honestly, that's just getting started because I don't think the AI companies have even had time yet to implement the Colorado requirements. I'm not sure they figured out how they're going to. But just one other piece of news since the last time we talked about this is now in California, the civil rights agency that deals with housing has now embraced algorithmic discrimination and Illinois has also embraced it. So this concept of algorithmic discrimination is spreading. Other states are now adopting it. And I do think that where it's going to lead if it's not stopped is right back to DEI, you know, AI.

It's probably that we need to start by rewriting these benchmarks. Remember that all these models, you know, when you do a big training run, you go and you try to run it against some set of benchmarks. The problem is that these benchmarks, I think, are overfit to a legacy way of thinking. And as Sax says, we need to revisit what those are and make them more objective and make it harder to actually get a good score unless you can be shown to be valuable. The second thing is that we may need to ask people in these next generation training runs to do a version that is built entirely on synthetic data where you have these judges determining whether this data is accurate or not from first principles and then you can compare them in a much more apples to apples kind of a way. But in the absence of that, the bigger problem you'll have is legislators trying to clean it up on the back end where there'll be these third parties that will go and take these models and show that these biases exist. They'll exist on both sides and then laws will get passed. Whole market gets mucked up and sullied. Everybody will get slowed down. So I think we need to change the benchmarks. We need to ask these companies to train on synthetic data. We need to have real disclaimers on what the sources and the weights are that you use if you don't do that. And we need federal regulations so that there aren't 50 sets of rules here otherwise we're screwed.

I'm more of a free market guy, so I would not ask where the data comes from or force people to use synthetic data or tell them how to do it. I think that this paper is useful in that it elucidates an important set of biases that the market can now say that is ridiculous and now the models will train and use that as a marketing exercise to say we are not biased. And so my my free market philosophy would dictate that this kind of elucidation will effectively create a vector upon which consumers will make choice in the market on what LLMs they want to use. Like Elon's going to harp on this. He's going to say look my Grok model Grock 4 fast is the only one that doesn't have this bias and that will cause more people to use his model and he will be able to take that benchmarking data and demonstrate. And some people they might want to have a biased model and they might want to say hey this one aligns with my philosophy my values my view and I want to use it.

Look, I think the question here that Friedberg raises is whether the market can just sort this stuff out on its own. And I think that would be great if it were true. But I do think it ignores the fact that in a lot of markets we have monopolies or oligopolies. We have institutions that have a lot of power and are very very hard to correct. So for example, Wikipedia has achieved a dominant position. I hope Rockedia challenges it and is able to fix that. But the easier path might just be for Wikipedia to stop blackballing and censoring conservative publications. In a similar way during the whole censorship era when the major social networks were all shadowbanning and censoring conservatives, it's not really realistic to have to start a whole brand new social network and overcome all of Meta's or in that time Twitter's network effect, right? Just to basically get a few accounts restored. When Gemini put out saying George Washington was black, people stopped using it. They're like, "This thing's a joke." So, I do think that consumers are not dumb and I don't believe in taking away agency from consumers. I think if I wanted to create subtle chaos, what I would do is make very small changes where none of these things are at the obvious stupidity of a black George Washington. But they can start to set the trajectory of a narrative forward and slowly over many many many years change the underlying content and what those models would do would be training kids over years if not decades one way of thinking versus another. I 100% agree with that that's the endgame here. By the way, in my opinion, that was the endgame for the Biden approach of requiring DEI values in these models. Indoctrination.