In the early days of Bitcoin, it was (1) not worth very much and (2) used mostly to buy drugs online. So a lot of people had the experience of, like, acquiring 500 Bitcoins to buy some molly, getting 100 Bitcoins back in change, taking the drugs and forgetting about the whole transaction for a decade. And then reading a news story about Bitcoin and thinking “oh damn that’s like $7 million, I wonder where my Bitcoins are,” and not being able to find them. Somewhere there’s a Bitcoin wallet that they “own,” but they have long since forgotten the private key or password that would unlock it. Oops! On some estimates this is like 20% of all Bitcoins, an absolutely incredible financial system.

Various technologies offer some hope for recovering those Bitcoins. Last week I mentioned “a guy who used Claude to recover his Bitcoin wallet password that he lost 11 years ago while stoned,” through some combination of fixing a bug in the password configuration and, I hope, psychoanalyzing him to figure out what words he was thinking about while stoned in 2015. Possibly a promising approach. Or of course quantum computing might unlock all the Bitcoin wallets. That wouldn’t help the people who lost them — it would help the people with quantum computers — but, in the crypto ethos of “code is law” and “not your keys not your coins,” maybe that’s fine. If you’re smart enough to crack open a Bitcoin wallet, the Bitcoins are yours, arguably, sort of.

There is another, older, somewhat less effective technology, which is law. If you can get a court to declare that you own a wallet full of Bitcoins, that’s something. It’s not as good as having the private key to that wallet: Without the private key (which the court can’t give you), you can’t transfer the Bitcoins on the blockchain, so you can’t spend them or use them in the ways people normally use Bitcoins. But it’s better than nothing. If you had a final judgment from a New York court saying that you own 100 Bitcoins at some particular blockchain address, then you could tell people “I own $7 million worth of Bitcoin” and that would be true in a legal, though not a cryptographic, sense. What could you do with that $7 million of inaccessible wealth? Things? Could you borrow against it? Would a bank lend you $1 million of real dollars against $7 million of legally owned but inaccessible Bitcoin? I mean, probably not, but worth a shot? [1]

Something like this is sort of what James Howells, the funniest Bitcoin guy, is going for. Howells at one point had 8,000 Bitcoins, but the hard drive containing the private keys is now at the bottom of a garbage dump, and nobody will let him dig up the dump to find it. A minor setback! He has, if not quite a court order saying that he owns the Bitcoins, at least a long and well-documented history of telling everyone he owns them. I wrote last year:

The normal, crypto-y way to prove that you own a stash of 8,000 Bitcoins is to have the private key for that stash, allowing you to send those Bitcoins over the blockchain. But being constantly in the news for a decade for owning that stash of 8,000 Bitcoins is also a kind of proof of ownership. Like, I could go around saying “actually those 8,000 Bitcoins are mine,” and Howells has no more cryptographic proof of ownership than I do, but you wouldn’t believe me, would you? He’s been telling this story forever and he seems to be quite sincere about it.

And in fact Howells said he was “planning to tokenize his legal ownership of the lost 8,000 BTC into a new Bitcoin Layer 2 smart token named Ceiniog Coin,” sure. The legal ownership (or whatever it is that Howells has) is arguably worth something, [2] even if it is not nearly as good as actually having the private keys.

Fine, good, fun, but here is a magnificent logical leap. Why should your legal ownership of Bitcoins be limited to the Bitcoins whose private keys you forgot? Thousands of people have forgotten private keys to thousands of wallets containing thousands of Bitcoins, some of them for years and years. How can they prove that those Bitcoins are theirs? What if you claimed all of them? 

How could you claim all of them? We have talked once or twice around here about the Finders Keepers Law, which is a joke but which also more or less exists. Here is Article 7-B of New York Personal Property Law, “Lost and Found Property,” which spells out the circumstances in which people who find lost property can keep it. Basically you’re supposed to deliver it to the police, and you and the police are supposed to make a certain amount of effort to find the original owner, and if the owner never shows up then the police eventually give it back to you and you can keep it. (Not legal advice; consult a Finders Keepers lawyer.) 

So in theory, you could take all of the abandoned Bitcoin wallets to the police, say “hey somebody lost these,” advertise “hey if you lost this Bitcoin wallet let me know,” and then wait. Probably no one would contact you, because they forgot about their Bitcoin wallets and are not going to recognize the public blockchain address in a long list. Eventually, the police would say “well nobody asked for these wallets back so they’re yours now,” and you could take your little receipt from the police and go to court and ask for a court order saying that you own thousands of Bitcoins. And maybe the court would give it to you. And then you’d have something. Not quite thousands of Bitcoins. But not quite not.

How can you “take all of the abandoned Bitcoin wallets to the police”? Well, a Bitcoin wallet is arguably just a string of numbers, the public key for the wallet; you could just write down a bunch of those strings of numbers and hand the list to the police. Is that quite the same as delivering actual abandoned property to the police? No, but (1) I am not an expert in Finders Keepers Law and (2) I suspect nobody is really an expert in this area of Finders Keepers Law. This is the bleeding edge of Finders Keepers Law.

Here’s a guy with a trade:

A New York man identified in court documents only as Noah Doe has filed a lawsuit in the Supreme Court of the State of New York seeking legal ownership of 39,069 abandoned Bitcoin wallets — assets he discovered using a self-developed algorithm, reported to the NYPD in compliance with lost and found property law, and spent over a year attempting to return to their rightful owners before filing suit. …

He developed what the filing calls “the Algorithm” — criteria and methods for identifying wallets meeting the legal standard for abandonment: dormant or inactive for at least five years, self-custodied rather than held by an exchange, and unresponsive through multiple periods of significant cryptocurrency price appreciation that any reasonable owner would have acted on.

Between December 26, 2024 and April 14, 2025, Noah Doe ran his Algorithm and identified three separate batches of found wallets — 1,544 in December, 546 in March, and 39,911 in April — all using his personal computer in New York City. Each time, within days of discovery, he physically brought a USB drive containing the wallet addresses to the NYPD’s 17th Precinct, where officers issued receipts and property invoices, per the complaint. The NYPD took custody of each drive before returning them months later. …

Of the 42,001 total wallets found, 2,932 were subsequently removed — including 424 that took on-chain action to demonstrate they had not been abandoned. The remaining 39,069 Bitcoin wallets took no action and are the subject of the lawsuit.

That is from NewsBTC this week; here is a story from Crypto Citizens Network, [3] and here is the complaint. (The complaint is “ABC Co. v. John Does 1 - 39,069”: Technically he is suing the anonymous holders of all 39,069 forgotten Bitcoin wallets for a declaration that they are his now.)

Will this work? I have no idea. It is not at all clear to me that he “found” these abandoned Bitcoin wallets; certainly he didn’t find them “in” New York. He delivered something to the police, and no one claimed it, and the police gave it back to him, but it’s a stretch to say that he delivered “the Bitcoin wallets” to the police. Anyone could have written down the same list of public keys and handed it to the police; arguably he handed the police a trivially reproducible reference to the abandoned property, not the property itself. Still, maybe?

If it does work, what will he do with these abandoned Bitcoins? Not spend them; even if he wins a court order saying that the Bitcoins are his, he still won’t have the private keys and won’t be able to transfer the Bitcoins. But he’d be able to say, accurately, that he owns many millions of dollars’ worth of Bitcoins. Which is better than not owning many millions of dollars’ worth of Bitcoins. Even if they’re all someone else’s Bitcoins. Finders keepers!

Debt with strong covenants is safer than debt with weak covenants. If a company borrows money, and its borrowing agreements say “we will not pay out any money to shareholders and will use any free cash to pay down debt as fast as we can,” that is safer than if the agreements say “we can do whatever we want with our money and we’ll see you in seven years when the loan matures.” The first arrangement should get a better credit rating and a lower interest rate. 

What if the company’s documents say “we can do whatever we want,” and the lenders go to the company and say “we’d actually prefer stronger covenants,” and the company’s chief executive officer looks them in the eye and says “for complicated reasons we can’t change the covenants, but you have my word of honor that every penny of profit will go to paying down debt”? The lenders might be reassured, no? It’s not as good as a written covenant. But if the company then goes and blows a bunch of money on dividends to shareholders, the lenders can at least go to the CEO and say “we had your word of honor” and make her feel bad. Maybe they can sue, too, not for a breach of the borrowing agreements but maybe for fraud. If the CEO’s word of honor carries weight with the lenders, then they probably should charge a lower interest rate with that promise.

What about the credit rating? The company can go to the credit ratings agencies for a rating, and the ratings agencies can say “hey you have a lot of debt,” and the CEO can look them in the eye and say “you have my word of honor that every penny of profit will go to paying down debt.” If the CEO’s word of honor carries weight with the ratings agencies, then maybe they should give the company a better credit rating. On the other hand:

• They will get questions, from investors, about their rating. “Why did you give this company such a high rating when it has so much debt?” Responding “the CEO gave us her word of honor that she’d pay down the debt” is problematic, in part because the meeting with the CEO was probably confidential, and in part because you sound kind of naive saying that. Get it in writing!
• If the company then goes and blows a bunch of money on dividends, the ratings agencies can go to the CEO and say “we had your word of honor,” but can they sue? About an oral statement in a confidential meeting? They didn’t even buy the debt; how were they defrauded? It’s all a little secondhand. Get it in writing!

So just get it in writing. Here’s a fun story from Bloomberg’s Davide Barbuscia about a not-quite-covenant in Paramount Skydance Corp.’s acquisition financing:

Paramount is financing its Warner Bros. acquisition with a daunting roughly $50 billion of debt, leaving investors skeptical of creating a heavily leveraged entity in a turbulent media industry.

To ease those concerns, Paramount Chief Executive Officer David Ellison privately promised ratings agencies including S&P Global Ratings that the family — which controls Paramount — would step in to tame leverage at the merged entity. The credit graders then pushed for that commitment to be made public and sure enough, Paramount revealed it in a regulatory filing last week.

“It was a verbal comment from David Ellison,” said Naveen Sarma, sector lead for US media and telecom at S&P. “I think all the agencies, and certainly us, insisted on a public disclosure as well.” …

The public disclosure proved decisive for S&P, which viewed the wealthy family’s backstop as a tacit commitment to inject additional capital if needed. While S&P has said it expects to lower the post-merger company’s rating by one notch to BB, Sarma noted that without the family’s backstop, the expected rating would have gone down by two notches.

Here is the relevant section of the filing:

Commitment to Deleveraging

In discussions with certain ratings agencies relating to the financing for the Acquisition, including the Exchange Offers and the potential Acquisition Financing Transactions, Paramount communicated to such ratings agencies that it is Paramount’s and its controlling stockholder’s plan and commitment, following the consummation of the Acquisition, to delever below a net debt to adjusted EBITDA multiple of 3.75x by fiscal year 2028 and a net debt to adjusted EBITDA multiple of 3.0x by fiscal year 2029, and that they will take steps to deliver the deleveraging targets.

Part of what is going on here is that S&P doesn’t want to look stupid: By getting it in writing, they can publicly support their rating decision. But part of what is going on here is that putting this statement in a public Securities and Exchange Commission filing makes it somewhat binding. It makes it a quasi-covenant. If, in 2028, Paramount Skydance is at 6x debt to EBITDA, its creditors won’t be able to put it into bankruptcy — it won’t be an event of default on its debt — but they will be able to tap that paragraph and say “ahem.”

Or sue. “Everything is securities fraud,” I often say around here. This is often annoying for public companies, but it is sometimes a useful commitment mechanism. If you explicitly say “we have committed to deleveraging” in your SEC filings, and then you don’t deleverage, you will probably get sued for securities fraud, and you’ll deserve it. [4]  And because everyone knows that, you can get a better rating for your debt. 

“There is no insider trading on Polymarket,” Bill Ackman told Eric Adams publicly, on X, less than nine months ago, meaning not “no one does it” but rather “there are no rules against insider trading on Polymarket so go right ahead.” I said at the time that he was wrong (though in a somewhat nuanced way), but he is a sophisticated financial industry professional and he was not the only person who thought that. There had, at that point, never been a prosecution for insider trading on Polymarket, and Polymarket’s general air of crypto anonymity and quasi-illegality encouraged the notion that there never would be. Sure, go nuts, insider trade, it’s just a weird online gambling site.

In a few years everyone will have internalized the idea that, in fact, insider trading on prediction markets is just as illegal as insider trading on stock markets, and with a few embarrassing exceptions, Polymarket traders will (1) not insider trade, (2) be a little clever about covering their tracks or (3) do it from non-extradition countries. But for now we are in a golden age of detecting prediction-market insider trading, in that lots of people thought it was more or less fine, did it carelessly, and are slowly finding out that nope it isn’t. Anyway:

A Google software engineer was charged with insider trading on Polymarket, where he allegedly made more than $1 million betting on one of last year’s most popular Internet searches.

Michele Spagnuolo was charged in a complaint unsealed Wednesday in federal court in New York. … 

According to the complaint, Spagnuolo, an Italian citizen who joined Alphabet Inc.’s Google in 2014, had access to company data that tracked user searches when he bet that Google’s most-searched person in 2025 would be the singer D4vd. …

At the time, Polymarket assigned a “near-zero probability” that D4vd would be the top-ranked search over figures like Pope Leo XIV and Kendrick Lamar, prosecutors said. When D4vd was publicly announced as the top-searched person in December, Spagnuolo allegedly made around $1.2 million.

Here is the Justice Department announcement, which calls him “MICHELE SPAGNUOLO, a/k/a ‘AlphaRaccoon,’” sure. Don’t use your employer’s secret information to trade on prediction markets, is the main takeaway here, which would be extremely obvious if I said “the stock market” but which people are still figuring out about prediction markets.

One standard model is that artificial intelligence can easily do a lot of entry-level white-collar work — building spreadsheets, formatting presentations, etc. — which will allow senior bankers and lawyers and consultants to operate with fewer junior employees. This will be good for the profitability of professional services firms, but bad for recent college graduates, and also bad for the long-term viability of the apprenticeship model on which those firms run. How will senior bankers develop deep connoisseurship for valuation, if they do not spend years as junior bankers building models? Where will the senior bankers come from, if there are no junior bankers?

Another model, though, is that every recent college graduate has spent the last few years using AI to do their homework, while the average senior partner is used to a certain way of doing things, scrawls “pls fix” in pen on printed copies of presentations, and is not about to go learn how to use a new piece of software. So what the banks, consultancies, law firms, etc., should do is hire a lot more entry-level employees, so they can automate everything, including whatever the senior employees are up to. This will be good for the profitability of those firms, and also for the recent college graduates, though arguably bad for the senior partners. What are they getting paid for?

Two objections to this model are:

1. No, the entry-level work can easily be automated by AI, but the senior-level work — the relationships, the connoisseurship, the judgment — can’t be. Nice try, recent college graduates.
2. Even if this theory was correct on the merits, the senior partners run the place; why would they hire a bunch of 22-year-olds to replace them?

Anyway here’s this:

Bank of New York Mellon Corp. is bullish about the promise of artificial intelligence, tripling the size of its intern and analyst classes since 2022 to tap into younger employees’ embrace of AI, Chief Executive Officer Robin Vince said.

“I am an AI optimist,” Vince said Thursday in an interview on Bloomberg Television’s Surveillance. “I think this is a revolutionary technology that for sure has to be well-managed.”

The additional intern-and-analyst investment is a nod to younger workers’ natural inclination toward AI, Vince said, without elaborating on the size of those employment changes. The bank wants to “continue to bring new skills, fresh people into our organization” while also equipping existing teams with AI proficiency, he said.

The good news for the younger workers’ might be the older workers’ natural disinclination toward AI.

US Funding Markets Are Flooded With Cash That’s Here to Stay. Kirkland & Ellis to spend $500mn building its own AI technology. Caesars Agrees to Be Taken Over by Fertitta in $5.7 Billion Deal. The $71.5 Billion Merger of Two Giant Railroads Just Hit a Snag. KKR’s Nuttall Says Trading Private Credit Is ‘Likely to Happen.’ Aircraft owners lease out engines rather than whole planes as prices soar. One Million New-Car Buyers Are Gone and They’re Not Coming Back Soon. ECB Says Leveraged Hedge Fund Bets Risk Bond Market Instability. The Degenerates of the Drone Stock Bubble. ‘I Was Just Sad’: Immelt Pens Substack About Post-GE Struggles. Court rules against Marc Rowan in Hamptons lobster shack dispute. F.B.I. Arrests C.I.A. Official With $40 Million in Gold Bars in His Home. Fired JPMorgan employee who claimed wrongful termination over $642.50 ‘Super Bowl’ deli platter awarded $4M.

If you'd like to get Money Stuff in handy email form, right in your inbox, please subscribe at this link. Or you can subscribe to Money Stuff and other great Bloomberg newsletters here. Thanks!