ValueArb

Jim Carrey won't, obligatory Ace Ventura

Also adjusted for tax rates! Beating my dead horse here;) I've always wanted to put in the time to make interest rate and tax rate adjustments to CAPE (Shiller PE), but the reason I don't is everything you wrote here. Making it "more accurate" isn't going to magically turn it into a market timing tool, the market and economy are both too random and mysterious to be easily predictable. So in the end all you get is a tool that claims the market is overvalued for a decade+, far more than enough time to kill any fund that relies on it for investment decisions.

Adam Neumann is the greatest fabulist of the century (so far). Now he's trying to buy back WeWork out of bankruptcy, and told the company he is ready to make an offer in partnership with Dan Loeb's Third Point LLC. Except that Dan is saying whaaaaaat? The reason Adam is a billionaire and we are not, is that he sees a vision and convinces others to make it happen, even if its not really true at the moment. Because just believing and saying it will make it happen! It’s February! If you are dropping Third Point’s name now, and you last had “preliminary conversations” with them in October, that suggests that your bid is not all that real. If you had any better meetings with anyone else since October, you’d be dropping their name rather than Third Point’s.

What’s a good way to estimate future natural gas prices? I’ve found a potential investment that would lose half its value if natural gas prices goes lower (even to zero), might lose 25% if they stay at current levels, and is a 5x-10x if they go up significantly. It’s an intriguing range of outcomes but what the heck do I know about natural gas prices?

Russian “technology”, training and morale so terrible that one of its most advanced SAM installations can’t shoot down two Storm Shadows flying directly over it on their way to massively damage a key airfield.

Did we lose the Korean war when Truman sacked MacArthur or the Civil War when Lincoln sacked McClellan? Civilian control of the military is a good thing, as long as his replacement is competent.

Why hire an American CS major for $100K when you can hire an Indian, Russian, or Chinese developer for $50k and have them work remote? Why not let them immigrate here so they pay US taxes? Immigrants aren't taking jobs, they are creating jobs. Paying people to clean my house and do my landscaping frees time for me to make 5-10x as much with those hours. American high school dropouts work at McDonalds, not doing landscaping for half the pay. When I ran product development for a PC software company many moons ago I got a resume and sample application demonstrating graphic algorithms from a foreign developer. It was so impressive that we paid to fly him to interview, hired him on the spot, and spent over $20,000 (in 1990s dollars) to get him a green card in order to stay and work for us. The product we built became so important it was about a third of our revenues, which meant it created roughly 40 high paying jobs. If we hadn't hired this genius, he would have almost certainly been hired by a foreign competitor and most of those jobs would have gone to them. The reason the US is so wealthy is because we saved and attracted immense amounts of capital, and that capital creates new higher paying jobs and makes us more productive at existing jobs. But we have a problem in that our declining birth-rate makes it hard to fill all the jobs, and having immigrants come in to take the worst lowest paying work on pushes americans up to better jobs. And over time those immigrants work hard and advance and save and help create even better jobs. We should target the smartest, hardest working foreigners and get them to immigrate here instead of throwing up roadblocks. A simple system of work visas would eliminate almost every problem we have with illegal immigraiton.

I have a backlog of new tender offers to read through and came across this one. Alternative Liquidity Index put out an offer to buy up to 10M shares in SVA (Sinovac Bio) for 3 cents. Now SVA stopped trading nearly 5 years ago due to a legal dispute over a share exchange that has yet to be settled, and last trade was at $6.50. Book value reported at year end 2022 was roughly $100 per share, and they claim $14/share in cash and roughly $80/share in investments as of June of last year, I guess I have to give Alternative Liquidity some props for an audacious attempt, but if they really think they'll be able to trade these shares someday (and that they are anywhere near book value) maybe offering a few bucks per share might actually get them some in volume. Or SVA is just a fraud, so why buy at even 3 cents? https://www.sec.gov/Archives/edgar/data/1084201/000110465923097975/tm2325464d1_ex99-a1.htm

I think you covered most of the issues with it. Another is when you leave a big limit order out at say, $13, hoping to pickup cheap shares, then sleep in the day they announce pre-market open that they are being bought for $9, and you wake up to find your order filled and you are sitting on a big loss.

This was great, thanks!

That's not how pattern matching models like ChatGTP work to my knowledge (but please correct if you think I'm wrong, I'm an AI amateur who has only built models using images for minor projects, not in real work). But my long experience in software development leaves me skeptical about how easy it will be to turn AI models into trustworthy products. The model will likely sometimes return the wrong decisions with the legal question given it. In the case of ChatGTP it will hallucinate imaginary decisions that never existed, but even if we constrain our model input set to proper legal databases instead of the World Wide Whacky Web that doesn't mean the pattern matching will always pick the correct decisions (or that it won't hallucinate non-existent ones). We've already seen ChatGPT written briefs get caught by judges, mostly because of hallucinations. This is the problem with all software, not just AI models. Before you implement any program, you have to have a strong quality control and testing plan for how to verify that it works well enough for production release. Obviously testing is almost always imperfect, how much you have to do and how well you do it is driven by product requirements. In the case of an "AI" legal advise bot, the potential costs of bad output are high (not as high as a medical bot, but still high) and you need to have a way to verify that questions are answered correctly almost all the time. That's why I suggest that building one should be a on a state by state basis (to avoid contamination and hallucination from other states different law sets), and output should be rigorously tested with actual legal experts.

Without quality control, your outputs will be untrustworthy and useless.

I agree with a lot of this. I think general purpose pattern matching models (colloquially called "AI" now) are terrible, we've all see their rough edges and hallucinations. My example is asking ChatGPT for legal advice and it confidently tells you terribly incorrect things about the laws in your state. But I think if you can create smaller models focused in your area of expertise and do a lot of testing/QA to validate their answers, you can create some powerful tools out of them. Getting back to my legal example, feed your model all of the laws of your state, case histories, common law, precedents, etc, but don't let anything else contaminate it from the world wide web, then work with a team of paralegals and lawyers to come up with a strong testing regime to find and fix any mistakes, and maybe you have something that works as well as asking any lawyer who passed the bar. That would be hugely valuable product for the average joe in your state.

Here is a little more color on FTX recovery. TLDR: Its likely but not certain claimants will get paid back all their money. They have regained a lot of asset value from increasing crypto prices for BTC and Solano, there was a lot of real estate and other assets Sam stole from FTX and gave to himself, friends and family that has been recovered, and then the Anthropic gain. That last paragraph says that the franchise value is still $0. It turns out that, still, nobody wants to trade on the FTX exchange; its stream of trading fees is extinguished forever. But the pile of magic beans and real estate that FTX had at the end of 2022 turned out to be worth more than its liabilities? The roughly $8 billion hole that FTX’s affiliated trading firm, Alameda Research, blew in its balance sheet got … filled up? The customer money all turned out to be there? FTX was worth less than $0 in November 2022, but has somehow become far more valuable in bankruptcy? It’s weird. There are several explanations for what has changed between November 2022 and now. For one thing, crypto prices collapsed alongside FTX, and by the time Bankman-Fried was looking for buyers, the value of its crypto holdings was low. Now crypto prices are up, which helps on the asset side. It should hurt more on the liabilities side — if FTX owes customers Bitcoin, that debt is worth more now than it was in 2022 — except that that’s not how the bankruptcy accounting works. FTX’s bankruptcy planprovides for “the valuation of claims in U.S. dollars as of the Petition Date” (Nov. 11, 2022) and payment in cash. If FTX owed you one Bitcoin before it collapsed, now it owes you roughly $17,000 (the value of a Bitcoin on Nov. 11, 2022). If FTX had one Bitcoin before it collapsed, now it has about $43,000 (the value of a Bitcoin today). If FTX had enough Bitcoins to pay off half of its customers' claims in 2022, now it has enough to pay off all of them. I’m not sure that’s the main mechanism. By the time of its collapse FTX didn’t really have much Bitcoin; its crypto holdings were largely “Samcoins” associated with Bankman-Fried that have not recovered nearly as much value. But a big chunk of FTX’s holdings were in Solana, which has rallied a lot. And FTX’s bankruptcy estate apparently dumped $1 billion of the Grayscale Bitcoin exchange-traded fund this month, profiting from the rise in Bitcoin prices since 2022. So rising crypto prices have definitely helped. For another thing, FTX has, like, posthumously pivoted to AI? In April 2022, Bankman-Fried invested $500 million of FTX/Alameda’s money into Anthropic, a somewhat obscure artificial intelligence startup. That was kind of a reckless thing to do with customer demand deposits, putting them into an illiquid speculative equity investment in futuristic technology. It did work out, however: AI is huge now, Anthropic is a big player, and FTX’s stake is probably worth billions of dollars. Bankman-Fried's whole schtick, for most of his career, was about taking terrifying risks that somehow worked out. “Let’s take our customers’ money and secretly put it into venture investments in an AI startup” is a terrifying risk, an insane thing to do, and yet it worked out! Not for Bankman-Fried, though; he’s in jail. Also there’s the real estate? Generally speaking, FTX seems to have siphoned off a ton of money into fancy apartments for its executives, big vanity investments in their friends’ companies, inflated endorsement deals with celebrities, and other reckless spending. When Bankman-Fried was shopping FTX to buyers in 2022, he didn’t count this stuff as FTX assets, because it mostly wasn’t: The money had been paid out to realtors and friends and celebrities. But the bankruptcy team generally argues that this stuff does belong to FTX, that it ended up in the hands of FTX’s executives and friends and family illegitimately, and they are trying to get it back. They might succeed with some of it. If your model of FTX is “Sam Bankman-Fried stole a lot of customer money and used it to buy luxury apartments,” that’s actually pretty good news for customers, because you can seize the apartments and sell them and give the money back to the customers. Whereas “Sam Bankman-Fried lost customer money gambling on crypto” would mean it’s probably gone. Also there is just a due diligence layer here that probably matters for valuation. The bankruptcy executives who currently control FTX are humorless types who care a lot about careful accounting and who report regularly to a court. When they say things like “we found $300 million of assets under a couch cushion,” people are inclined to believe it. When they want to sell stuff, they run a slow and heavily lawyered sales process that lets multiple buyers kick the tires, and the buyers can have some confidence in what they are buying. When Sam Bankman-Fried tried to sell FTX for $0, he had a slapdash and terrifying spreadsheet that emphasized the “Hidden, poorly internally labeled ‘fiat@’ account” with a balance of negative $8 billion. If you are a potential buyer, and you get that spreadsheet along with a text message like “hey FYI I need your final binding bid in two hours or it’s off to jail with me,” you might not be inclined to make an aggressive bid? You might think “hmm negative $8 billion is a suspiciously round number, what if there are other hidden liabilities here that I don’t know about and can’t find out about in two hours?” You might just pass. It’s possible that FTX really was worth at least $0 the whole time, that even at its lowest point it had enough stuff lying around to pay back all its customers. It’s just that Sam Bankman-Fried was in no position to make anyone believe that

He is cute, goddamnit. There should be a law that no one can become a billionaire until they are an old saggy man like myself.

Anthropic, which I already mentioned, is a huge win with at least a $2B paper profit. You already pointed out that he gambled on BlockFi, Voyager, Alameda, Robinhood etc with customer funds, so its exactly like taking customer funds to Vegas.

FTX had legal agreements with clients. Some deposits could be loaned out, most could not. FTX loaned out almost all of the deposits. If you had withdrawn the extra funds and spent them, yep, that's fraud. If you left them in the account until it the situation was fixed, no.

Lewis's book has been heavily debunked, he was clearly dazzled by SBF to the point where he was unable to ask tough questions or accept negative information about SBF. And if you steal $100,000 from work to buy lottery tickets, and one of them pays off $100,000 so you can repay your boss in full, did you really steal? The answer is of course, yes. SBF took customer deposits he wasn't allowed to lend out, and lent them to his own hedge fund Alameda to cover their trading losses. He did this because while FTX had a lot of long term assets of potential value, the only assets (liabilities really) that had liquidity were customer funds. In one case in 2023 he threw $500M into Anthropic AI startup, which at its latest proposed valuation ($18B) would be worth around $2.5B. But how can FTX shareholders monetize that stake? Anthropic is likely to be cool to the idea of helping FTX sell it because it will hurt the valuation of their current round (assuming it hasn't closed yet). So the lawyers more likely will have to hire a bank to find investors who missed out of the latest round who might pay $1.5B to $2B. Still a nice gain, but essentially he made a high risk investment and its so far paid off, using other peoples money. Its no different than what Martin Shkreli did, where he lied and stole from investors, but before he lost all their money hit on one big investment win large enough to pay them back. He then tried to claim he did not defraud any clients since they all got their money back, but a judge saw through that to the tune of 7 years in prison.

I do have a macro viewpoint on interest rates, that they historically always return to the 5-7% level, and given our record breaking national debt load, may even go higher in the long run. But again, it almost never comes into play in my investment decisions. I own no real estate, and never run any DCFs. I might use Graham's formula, and I would use a historic interest rate instead of near zero rates when we had them. But that's it. If I was going to buy real estate I'd much rather buy it when interest rates are high, but either way, if the cash flow works out to offer a reasonable return I'd probably do the deal. I would always treat appreciation as a bonus, my only concern is the property pay for itself with a reasonable margin of safety. So in RE like in stocks, I'd say no to a hundred opportunities before saying yes to one.

Getting out requires a catalyst, and I can't get on boards. So I pick carefully from net-nets that have already announced a catalyst, or I think are likely to announce one shortly. The second category was very painful for me in 2nd half 2023 however, and I've mostly pulled back to first category. When a sale is announced usually volume spikes and spreads narrow substantially, and even if they don't you can just wait for close. Going private transactions are one of my favorite catalysts. Even if its an outright theft where management is only paying 60% of book, that's still a 20% gain in 6-9 months if it closes. There is likely a lot of truth in what you wrote here. I get paid quarterly and could abuse my clients by throwing a little cash at the ask at the end of day on last day of quarter to show better results. But besides being blatently unethical and wrong, now do I have to do it end of every quarter just to stay even? And if the business gets sold for below that ask price, I've just manufactured a loss. Its hard enough to beat the market why also juggle a bunch of lies and manipulated prices on top of that for a short term gain. So while I do not do any manipulation, I find significantly higher return opportunities in nano and micro caps. If I'm able to outgrow them I'll be very excited because it will mean I have hundreds of millions under management. And the skills I apply here are applicable to larger caps, the returns will just be a bit lower.

Its really hard to get a woman to respond to me on Tinder at my age, I finally got one but she moved to Oahu. If you need a short-term house-sitter it would guarantee me at least one date before she blocks me;) In the mean-time, February is the test of the next Starship prototype. Assuming they've solved the water-hammer effect in the fuel lines that blew up SuperHeavy last test, probably get to see them land a 230 foot sky scraper in the ocean which would be super cool. Only problem is its likely that we only find out a day or two ahead of time when the test is going to happen, and even then it could be delayed days for any pad/assembly/fueling issues or weather. So when they announce I'm planning to hop in my car in Phoenix and drive straight through to Boca Chica, Texas and sleep on the beach with my dog.

If we are going to weigh anecdotal evidence like that one boat builder then I have to share that my ex works for a large national home builder in the Phoenix area (epicenter of almost every housing bust), and her office is very busy. As usual I can't make a dime wasting time making any macro guesses about the economy. I can only keep reading annual reports until I find an attractive investment at a great price offering outstanding future returns. If macro changes hurt its business, then I will probably still make good returns, if macro changes help it then my returns may turn out to be incredible.

Sisu might have something to say about that.

I predict that stocks will be up strongly in 2024, except for all the ones that won't be.