OpenSea was, and as far as I can tell still is, a big marketplace for trading nonfungible tokens. These tokens had a vogue in crypto a few years ago, when NFTs from popular series like Bored Ape Yacht Club would sell for millions of dollars. Anyone can create NFTs, and they generally have no intrinsic value; their value is strictly memetic. If a lot of people have heard of Bored Apes and think they are cool, then they will be valuable; if not, not.

Given this, OpenSea played an important role in the NFT boom. Specifically, OpenSea has a website, and it would feature some NFTs prominently on its homepage. People would go to OpenSea’s website to buy NFTs, and they would see the NFTs featured on the homepage, which would make them more likely to buy those NFTs, so those NFTs’ prices would go up. So if you could buy an NFT before it was featured on the homepage, and then sell it after it was featured, you could pretty reliably make money. It is hard to convey how weird 2021 was.

Anyway the guy who picked the NFTs for the homepage in 2021, an OpenSea employee named Nathaniel Chastain, naturally insider traded on his picks:

On August 9, 2021, CHASTAIN purchased ten of the NFT “Flipping and spinning” before it was featured on OpenSea’s homepage, and then sold them at prices 250% to 300% higher than where they were before the NFT was featured.

On September 14, 2021, CHASTAIN purchased the NFT “Spectrum of a Ramenfication Theory,” as well as three other NFTs by the same creator, shortly before the NFT was featured on OpenSea’s homepage. CHASTAIN made over four times the amount of money when he sold “Spectrum of a Ramenfication Theory” early the next morning.

I cannot emphasize enough that his profits on “Spectrum of a Ramenfication Theory” did not come from art connoisseurship or insight into the fundamental value of these NFTs. You put it on the website, it goes up. Chastain made about $57,000 from about 15 trades. 

Is that legal? Well, nothing here is legal advice, but in 2022, Chastain was arrested and charged with wire fraud. We talked about it at the time. It was an interesting case in that “insider trading” is normally a variety of securities fraud, and these NFTs are (almost certainly) not securities. But wire fraud is arguably a generalization of securities fraud. I wrote:

Securities fraud is doing fraud about securities. Wire fraud is doing fraud about absolutely anything, as long as you do the fraud using email or the phone or text messages or a chat app, and in 2022 you certainly do. So every fraud is wire fraud.

You might put these things together and conclude: If insider trading in securities is securities fraud, then insider trading in anything is wire fraud. If insider trading — using information from someone else, to whom you have a duty to keep it confidential, to buy for your own account — is a “device, scheme, or artifice to defraud” under securities law, then surely it is also a “scheme or artifice to defraud” under wire-fraud law. 

That was how I characterized the government’s position when Chastain was arrested. It has an obvious intuitive appeal. This seems bad, you know? He was convicted and sentenced to three months in prison. 

But today the US Court of Appeals for the Second Circuit overturned his conviction. Here is the opinion. “Insider trading,” I like to say, “is not about fairness, it’s about theft.” Lots of people trade securities with an information advantage; you are very much allowed to trade while knowing things that other people don’t know. What makes insider trading illegal in the US is mostly that you are misappropriating that information from someone else. [1] As I put it last year:

In the US, most of the time, it is illegal to trade on inside information about a company not because that’s unfair to everyone else who doesn’t have the information, but because you have some duty to somebody else not to misuse their information. So if you work for a public company, you have a duty to the shareholders not to trade on inside information before disclosing it to them. Or if you are the spouse or golf buddy or therapist of an executive at a public company, and she tells you about an upcoming deal, you have a “duty of trust or confidence” to her not to go trade on it, and if you do that’s a crime.

The same logic applies to wire-fraud insider trading in NFTs. Chastain’s alleged crime is not quite “he ripped off NFT holders by trading on inside information.” It’s “he ripped off OpenSea by misusing its confidential information.” The wire fraud statute forbids “any scheme or artifice to defraud, or for obtaining money or property by means of false or fraudulent pretenses, representations, or promises.” You might think “yes, insider trading on NFTs is a scheme to defraud the people on the other side of your NFT trade,” but that is not how lawyers apparently think about it. Insider trading on NFTs is a scheme to steal OpenSea’s property, which is its confidential information about which NFTs will be featured on the website.

And the appeals court concluded that this is not real property because there was no evidence that it was valuable:

“To be guilty of wire fraud, a defendant must (1) ‘devise’ or ‘intend to devise’ a scheme (2) to ‘obtain money or property’ (3) ‘by means of false or fraudulent pretenses, representations, or promises.’” …

Under these standards, not all information kept confidential qualifies as property. Neither the Supreme Court nor our court has held that confidential information that lacks commercial value will qualify as property under the wire fraud statute. …

Because “the wire fraud statute reaches only traditional property interests,” we must decide whether confidential business information qualifies as a traditional property interest even if it lacks commercial value to the business. … We conclude that it does not. … Confidential information does not qualify as a traditional property interest unless it has commercial value to the company that holds it. …

The [trial court’s] jury instructions would allow a conviction under the wire fraud statute even if OpenSea thought it was merely unseemly to reveal the planned featured NFT before it appeared on the website — and even if the evidence showed that treating the featured NFT as confidential had no commercial value. … In other words, the instructions allowed the jury to convict based the government’s “view of[] integrity” in business conduct rather than the misappropriation of “property rights only.” ...

If the wire fraud statute criminalized conduct that merely departed from traditional notions of fundamental honesty and fair play, “almost any deceptive act could be criminal.”

Because there was no proof that Chastain stole anything valuable from OpenSea, he wasn’t properly convicted of wire fraud.

Two points here. One, this is a strange result, because obviously the information about which NFTs would be featured on the website was valuable. You can tell because Chastain (1) traded on it and (2) made money. OpenSea itself did not regularly exploit this value — it didn’t itself insider trade on its homepage picks — presumably because it thought it was better business not to. [2]  This is a weird feature of “insider trading is not about fairness, it’s about theft”: Chastain was charged with stealing something from OpenSea (information about its homepage) that obviously had value (he made money from it), but the theft didn’t obviously cost OpenSea anything (it wasn’t making money from the information). Chastain didn’t exactly make his $57,000 at the expense of OpenSea; he made it at the expense of other NFT traders. And insider trading laws do not protect other traders.

Two: What does this mean for actual insider trading law? Maybe nothing; the wire fraud statute is different from the securities fraud statute. But there are connections. [3] And I wonder a little whether, in this more fraud-friendly era, courts will be more hesitant to convict people of crimes for misusing corporate inside information. In particular, I wonder if this decision will mean anything for “shadow trading” cases, where an insider at a company gets information about that company (a merger, earnings, etc.) and uses that information to make trades in another, correlated company’s stock. Does that deprive the insider’s company of a “traditional property interest”? Is information about a pending merger a traditional property interest? If insider trading is about theft, not fairness, do you have to prove that the insider stole something that the company was using? How do you prove that? Or is insider trading a little bit more legal now?

There is just so much money being thrown at artificial intelligence these days. Meta Platforms Inc. keeps hiring AI researchers with nine-digit pay packages. Presumably these are people with excellent reputations in their field and a long track record of publications, all of which is hard to fake. But for $100 million, surely someone should try to fake it? Like, go to the right parties, have ChatGPT write some fake papers for you, get really good at interviewing, try to catch Mark Zuckerberg on an off day, see what happens? From Mark Zuckerberg’s perspective, that’s fine. If he hires 100 people and gets 98 world-class AI researchers and two complete scammers, that’s a great outcome; spending $200 million on the scammers is pocket change.

And getting hired at Meta is a relatively hard target. The easy version is, you start a startup, the startup is “we will do AI for ______,” you spin up a product that is some combination of (1) just asking ChatGPT for the answers and (2) having offshore programmers pretending to be AI, and you go out to venture capitalists to raise money in one of the greatest gold rushes in the history of capital markets. “Hi I am doing AI for—,” you begin, and the VCs say “shut up and take my money.” And from their perspective that’s fine! If they invest $50 million in each of 10 AI startups, and they get two world-class AI companies and eight complete scammers, that is a tremendous outcome; they’d be thrilled with that. Flinging money at scammers is a small price to pay to get as many lottery tickets as possible.

The point here, which is not any sort of advice, is that if you are a scammer and you’re not doing an AI startup, you’re wasting your time. And the other point is that in the next couple of years we are going to see a bunch of scam AI companies collapse, and each time there will be stories like “this is very embarrassing for the VCs who threw money at them,” and the VCs will be like “no it isn’t, this is fine, cost of doing business.”

Anyway Bloomberg’s Newley Purnell, Yazhou Sun, and Mark Bergen have a story about Builder.ai and its charismatic founder, Sachin Dev Duggal:

Builder.ai’s audit committee uncovered a web of dubious transactions. The London-based company booked $142 million in sales from resellers that never paid any money and claimed an additional $107 million from customers who made deposits of as little as $1. Such methods, the audit committee found, were used to overstate revenue by 300% as Duggal secured an emergency loan last year. A law firm hired by the company also determined that Duggal “orchestrated a scheme” with a high-profile Indian startup to exaggerate sales through what’s known as “round tripping,” according to documents viewed by Bloomberg News.

Duggal’s alleged financial transgressions fueled the most spectacular collapse yet of the generative AI era. Builder.ai had seized on the fervor around ChatGPT to earn a $1.5 billion valuation on the promise of AI systems that made building apps, as Duggal often put it, as easy as ordering pizza. He described wanting to build a trillion-dollar company. EY named him Entrepreneur of the Year in the UK. …

Duggal had no trouble finding investors, including Microsoft Corp. and the Qatar Investment Authority, and created one of Europe’s largest AI startups — despite a past full of legal troubles and spurned associates. The startup’s unraveling raises critical questions about how closely venture capitalists, law firms and consultants vet startups in hot industries like AI, where the fear of missing out in a boom can trump more careful deliberations.

He raised $450 million, which is like, what, three or four Meta hires? Eventually he was pushed out and the company went bankrupt; we have previously discussed the fake-sales allegations. Here’s a LinkedIn post from Duggal’s successor as CEO, whose first bullet point is “the AI was real,” which is not how you want to lead off your LinkedIn post about your AI startup. 

Some of Builder.ai’s money came from Insight Partners, which is doing just fine:

Insight’s investment in Builder.ai has been an embarrassment internally, but its failure isn’t a big concern because the firm has been able to make record distributions to investors, according to a person familiar with the matter.

Yes! Right! Nobody wins any awards for passing on a bad AI deal, but missing a good AI deal is disastrous. There should be a lot of AI scams!

The simple model of Elon Musk is that he runs a bunch of companies with distinct but overlapping investors, employees, missions, etc., all of which have a single broad purpose, which is to facilitate human life on Mars. Humans do not currently live on Mars, though, and getting them there will require a lot of capital, so for now most of Musk’s entities do some Earth-relevant business to pay the bills. Tesla’s telos is to build (1) Cybertrucks that look ridiculous driving around the suburbs but would look cool on a Martian landscape and (2) droids to work on the Martian moisture farms or whatever, but for now it sells electric sedans to keep the lights on. SpaceX’s ultimate purpose is obvious (you need rockets to go to Mars), but it’s lucrative right now to launch communications satellites. Artificial general intelligence will obviously be useful in getting to Mars, but for now an AI chatbot will lure consumers to X; X is somehow a laboratory for “how a citizen-led government that rules by consensus might work on Mars,” but for now it’s a way to get in fights online and serve ads.

The Boring Co., a Musk entity that makes tunnels, is very important for life on Mars, because Mars doesn’t have much of an atmosphere, and if you lived there you might prefer to be underground. It is much, much, much less important for life on Earth, because Earth does have an atmosphere, and you can just, like, build a house or walk around or drive a car or whatever on the surface of the Earth. There are uses for tunnels on Earth — subways, crossing rivers and mountains, secret lairs — but they are relatively small niches. You could imagine a fairly nice and not-too-distant future where everyone has an electric car, but a future where everyone lives in tunnels is obviously dystopian and hopefully quite distant.

But Musk (1) wants to go to Mars and (2) can’t resist a good name, so the Boring Co. exists now. It just doesn’t have much to do. It sits around making jokes about flamethrowers. Bloomberg’s Kiel Porter checks in:

Musk first hyped Boring nearly a decade ago. He vowed to build a futuristic, ultra-fast underground hyperloop as a revolutionary solution to the world’s “soul-destroying” traffic problems, raising more than $900 million from the likes of Sequoia Capital and Peter Thiel’s Founders Fund.

Since then, the company has started digging just one public project: An underground loop in Las Vegas that ferries conference-goers to and from a convention center and to several hotels. While the city has approved 68 miles of tunnels, Boring has dug about eight miles, and less than four miles are operational. Arguably, its most notable commercial endeavors, in keeping with the company’s tongue-in-cheek name, have been, quite literally, jokes: flamethrowers and a burnt hair-scented perfume.

The tunneling venture has either pitched or been selected to carry out more than a half dozen other US municipal projects since 2017, but there’s no evidence that the company has broken any ground outside of Vegas. Most recently, it has pledged to dig 10 miles of tunnel in Nashville and 10 miles in Dubai. Some initiatives have been stalled or mothballed, while the company has walked away from others. These days, the company is building tunnels to connect Musk’s own properties, according to people familiar with the plans. They say the aim is to allow Boring, SpaceX, Tesla and X employees to shuttle back and forth underneath a road in the rural Texas town of Bastrop, where the billionaire has been expanding his business empire.

It is just not at all a terrestrial company. But wait until you need a place to live on Mars, then you’ll be glad that Musk has figured out tunneling.

Harvard University has a pretty good brand, with that brand symbolizing, roughly, “smart.” People want to be associated with that brand. The trick for Harvard, as a business, is to ruthlessly monetize that brand while also preserving it. Harvard has about 1,600 undergraduates a year. If it sold all 1,600 of those spots to the highest bidders, it would get a lot of money, once, but word would get out, the brand would quickly deteriorate, and it would have a tough time selling spots in the future. If it admitted 1,590 students based on meritocratic criteria [4] and sold 10 spots to the highest bidders, it would (1) get a lot of money for those spots and (2) more or less preserve the brand. For various legal and reputational reasons Harvard does not explicitly do that, but, you know. 

There are cleaner, more ruthless, more scalable approaches. One possibility is:

1. Keep admitting those undergraduates based mostly on meritocratic criteria, thus preserving the brand for the core product; and
2. Sell a lot of noncore products to anyone who wants to pay to be associated with the brand.

That is, you want “going to Harvard” to be an impressive thing that money can’t buy, but you want to offer some sort of “going-to-Harvard-with-an-asterisk” experience that money definitely can buy. You want that experience to be somewhat easy to confuse with “going to Harvard,” but not too easy. You want people to pay for the asterisked experience because they think it is a good substitute for going to Harvard, but you also want some deniability, so that people with the un-asterisked going-to-Harvard experience can say “well that person didn’t really go to Harvard, they just paid a lot for a fancy souvenir” and the core brand is maintained.

Bloomberg’s Janet Lorin reports on that business model:

Harvard University is drawing an increasing share of its revenue from helping businesspeople beef up their résumés, giving America’s oldest institution of higher learning a lifeline as it confronts its largest-ever crisis. …

Executives looking to step up their management skills or get up to speed on new technology have eagerly shelled out for expensive in-person and online courses at the storied university out of their own pockets, or with help from their employers. Those programs are now providing Harvard with revenue that’s relatively insulated from US President Donald Trump’s efforts to deprive Harvard of federal funds — and helping soften the blow as administrators watch expenses rise, endowment returns sag and donations drop.

Overall, executive education and continuing education for nontraditional students delivered nearly $600 million last year, up from $155 million two decades ago, publicly available data from the university shows.

Alumni of Harvard’s undergraduate and graduate programs have sometimes scoffed at such programs, posting memes to social media implying that they lack Ivy League gravitas. … But putting the Harvard imprimatur on classes that let corporate managers burnish their bona fides and expand their networks is providing the embattled university with an important safety net at an uncertain time for US colleges and universities.

Arguably the story here is that Harvard has spent almost 400 years building up the brand and now is as good a time as any to monetize it. And arguably part of the goal of the Trump administration is to reduce the value of the Harvard brand, so they might as well sell high now.

Here is a post from the UK AI Security Institute looking for economists to “find incentives and mechanisms to direct strategic AI agents to desirable equilibria.” One model that you can have is that superhuman AI will be terrifying in various ways, but extremely rational. Scary AI will not be an unpredictable lunatic; it will be a sort of psychotic pursuing its own aims with crushing instrumental rationality. And arguably that’s where you need economists! The complaint people have about economics is that it tries to model human behavior based on oversimplified assumptions of rationality. But if super AI is super-rational, economists will be perfectly suited to model it. Anyway if you want to design incentives for AI here’s your chance.

Treasury Secretary Says Trump Accounts Could Pave Way to Privatizing Social Security. Trump Just Crashed the Copper Market. Trump and Dimon Are Talking Again After Yearslong Rift. KKR Raises $6.5 Billion for Private Asset-Backed Finance Deals. Blue Owl chief warns of manic market for second-hand private equity stakes. Private Credit Has a ‘Real Place’ in 401(k)s, Centerbridge Says. How Podcast-Obsessed Tech Investors Made a New Media Industry. “People feel emotionally tied to Excel.”

If you'd like to get Money Stuff in handy email form, right in your inbox, please subscribe at this link. Or you can