If you take the business part of the business and separate it from the Ponzi part of the business, it’s not a Ponzi scheme. But apparently it was a package deal—at least until Alameda were to weather the crypto crash? Like, we just need to siphon FTX customer funds over to Alameda until they recoup $8-10 billion over time in profits, which have not been forthcoming lately. But oops, we “forgot” that when we transferred that money over, it affected our financials—there it is, fake books. And that’s not even getting into the laughable valuations of their other “shitcoin” assets, or even of FTT itself, just the flow of funds!
This also speaks to further impairment of FTX’s value by management—if you separate the business part of the business from the management part of the business...and you can to some extent, but the damage has been done. Who is going to trust FTX as an exchange going forward even with a new structure and management team?
Finally, it’s not like Madoff vaporized the money and SBF/FTX/Alameda didn’t. If anything, it’s the opposite; Madoff was a far better steward, making the assets recoverable. SBF/FTX/Alameda simply gambled them away. Put differently, the non-Ponzi part of the business was a bigger share of Madoff’s fund than of SBF’s bundle. Obviously the valuation in either case was fake, based on a multiple of both real assets and Ponzi accounting over time, but Madoff skimmed customer funds while SBF/FTX straight-up embezzled them to prop up the husk of Alameda. In its death throes, FTX was in straight-up Ponzi mode, making withdrawing customers whole to maintain the facade at the expense of those last in line.
I think the comparison is quite apt, and the points of contrast are more interesting than absolving. I was initially hesitant to say there was some fraud/Ponzi beyond just accidentally falling into borderline insolvency, but by this point, especially with how enormous the hole, it looks much more intentional, and Matt Levine has even dropped the P-word because of it.
Disagree that “[a]t best, these theories just do not bring much new to the table” but agree that the over-emphasis on these theories is “extremely unhelpful.” That is, they provide good insights and explanatory power, and even substantial explanatory power on the margin sometimes, but having non-zero R2 is not the same as being the explanation with the highest R2. Assuming the latter is often not accurate. Ironically, I find myself mostly agreeing with a post that Razied disagrees with while also mostly agreeing with Razied—the human value function is not solely, or even primarily, status seeking.Likewise, I also disagree with “[on social media], it is well known people are playing their meaningful status games and doing false signalling on a high simulacrum level’ because Ape in the coat is doing exactly what they are complaining about. Certainly this happens, and in some cases dominates any other reasons for people’s behavior on social media, but again, most of why people engage on social media is likely not to be about obtaining status, unless status is to be so broadly defined as to include things like “personal satisfaction.”Getting meta, here I find myself engaging in a social medium that even has a very clear social approval mechanism. Yes, I want lots of yummy upvotes (is it because they accord me status somehow, or because the social approval provides peer review for the validity of my hopefully-rational take?), but mainly I’m posting because I think my opinion is right, worthwhile to share, and hopefully conveyed in a way that can, if not persuade, update.
Nicely done!The question is what should be the denominator of the risk?For societal aggregate worries: Day? Sure.For personal risk: Day? Nope. Person-day? Yes!
He guessed I had an allergic reaction and threw 5 different antihistamines
Not so much dumb luck after all! Allergic reactions often cause inflammation, and it’s the inflammation that is uncomfortable. Sure, it’s not very controlled (could have suggested one, then another, then another, until you got to Boswellia, preferably in order of prior belief for each more-specific hypothesis), and other things could cause inflammation, but it’s not completely luck either. (Though this did not detract from my enjoyment of the post!)
Some quick Googling says,
Current research showed that 3-O-Acetyl-11-keto-beta-boswellic acid (AKBA) is the one boswellic acid with strong pharmacological activity; for example, AKBA has a powerful inhibitory effect on 5-lipoxygenase (5-LOX)
The tissue, animal model, and animal and human genetic studies cited above implicate ALOX5 in a wide range of diseases...chronic inflammatory conditions such as rheumatoid arthritis, atherosclerosis, inflammatory bowel disease, autoimmune diseases
Helpfully, this study compared AKBA vs. NDGA, so maybe that could be a useful way to test the mechanism of action (they also differ in their specifics within that too, so that’s yet another question mark). Obviously not medical advice and just a curious wondering.
To the contrary, johnswentworth’s point is not that the experiments have low external validity but that they have low internal validity. It’s that there are confounds.
Ironically, one of my quibbles with the post is that the verbiage implies measurement error is the problem. Not measuring what you think you’re measuring is about content validity, but the post is actually about how omitted variables (i.e., confounders) are a problem for inferences. “You are not Complaining About What You Think You Are Complaining About.”
Adam’s whole position here, to me, is rather silly, even if we limit ourselves to use cases where the Twitter poll is being used only to try and extrapolate towards national sentiment.
I agree except with the last part (it’s not silly when thinking about extrapolating to national sentiment). The key is to what extent is it evidence of [insert thing], and of course if you’re interested in learning more, what are the factors that affect the extent to which it is evidence of [insert thing]? In other words, what are you trying to generalize to, and what interesting things are limiting your ability to generalize to [insert other thing]?
Often we are comfy with generalizing from sample to appropriately-defined population (sample of Zvi Twitter noticers to Zvi Twitter noticers), but when we don’t define the scope of our generalization properly, we get uncomfy again (sample of Zvi Twitter followers to US general population). Often we are interested in the limits of generalizability (e.g., this treatment works for men but not women, isn’t that interesting and useful!), unless the those boundaries are trivial (e.g., vasectomies work for men but not women, gosh!) or we already don’t see them as boundaries (e.g., “what if you had changed the wording from ‘YOU in particular’ to ‘YOU specifically’?).
Interestingness is in the eye of the beholder. Concede to Adam for the moment that the boundaries are not interesting because they are well-known limits to generalizability (selection, wording). Then, is it “bad evidence?” Depends on what you’re trying to generalize to (what it is purported to be evidence of)! Adam waves between Twitter polls being “meaningless” and “does not generalize at all” as in worthless for anything at all, which is obviously mostly false (it should at least generalize to Zvi Twitter noticers, though even then it could suffer from self-selection bias like many other polls), vs. “not representative of general views,” which is not silly and is far more debatable (it’s likely “weak” evidence in that Twitter polls can yield biased estimates on some questions [this is the most charitable interpretation of the position]; it’s possibly “bad” evidence if the bias is so severe that the qualitative conclusions will differ egregiously [this is the most accurate interpretation of the position seeing as he literally wanted to differentiate it from weak evidence] - e.g., if I polled lesbian women on how sexually attractive the opposite sex was to infer how sexually attractive the opposite sex is to people generally). So overall, the position is rather silly (low generalizability is not NO generalizability, and selection and wording ARE interesting factors relevant for understanding people), except on the very specific last part, where it’s not silly (possibly bad evidence) but it is also still probably not correct (probably not bad evidence).
Re: Physical World Modeling
I’m not surprised by those South Africa vaccine efficacy numbers, since they are broadly in line with releases we’ve been getting over the last 9 months. We already knew VE vs. infection for the monovalent vaccine was very low and that VE vs. severe disease would be higher but still lower than vs. Delta. We already knew they provide about 3-4 months of “good” protection. We already knew VE vs. BA.4/5 was lower than vs. BA.1/2.
But yeah, it’s pretty easy to square, seeing as Dr. Rivers tweeted the paragraph whose last sentence is “need for vaccines to incorporate variants of concern.” Table 1 is stuff we know about old vaccines, and future public health decisions will not be using those. An annual booster of bivalent vaccine timed like the flu shot is a different story! It’s the one being told (e.g., Zvi’s points 2 and 5), and it’s the one not shown in Table 1. Seeing as we are getting a bivalent booster in about 10 months, in part thanks to uncertainty in how the FDA was going to treat approvals, getting it down to the 6 month lead time of the flu vaccine seems in the realm of possibility for future years (potentially contra Zvi’s point 3). We will “need for vaccines to incorporate variants of concern” for annual shots to make sense, and that’s exactly what we’ll be doing.
Right, when you go to argue the merits, you ask “well, if there were to be a phase change, what would the phase change look like?” And the original estimate was derived from not much effort in calibrating the numbers, and the reply was that even if we saw an utterly shocking phase change, we’d get nowhere close to 20%. You can do varying degrees of in-depth analyses to get to that point (good on you), or you can do like I did and rely on a semi-informed prior.Here’s US growth from 1947. Imagine all the things that happened since then that could have induced mild phase changes to the growth trajectory. ASSUMING that there will be a substantial phase change (again, see Cameron Fen’s thread), 20% is still ludicrous.https://fred.stlouisfed.org/graph/?g=TE7B
Not a comment on the argumentation or anything, I know we want to be rationalists and worry about the arguments (so thank you for posting about the disagreement that actually offers some analysis), but just registering my initial reaction to the 20%/year in 10 years claim:Anyone with a cursory understanding of the history of economic growth (I don’t even mean professors who have spent their careers studying growth economics) will know that number is facially ridiculous. My first thought was, great, now I know this person has no idea what they are talking about and can be safely ignored. As a communication device, that prediction failed miserably for me because it did not make me want to assess the underpinnings or energize me to research further about economic growth, which is the least such a tweet should have done if not prompt an update. Though good luck with the latter because, again, if you have an iota of knowledge about the area, your prior distribution for the realm of US economic growth possibilities is correctly more narrow than “let me throw out a shocking round number to myself, see if I can live with it, and then see if others take it seriously.” Am I being too harsh? Well, no, he literally “didn’t put much effort into calibrating the numbers.”Now back to the regularly scheduled programming of arguing about the merits.
That’s a charitable interpretation. But the steps for microwaving and not microwaving formula are the same, just in a different order. If you forget to check the temperature of overheated formula, you may burn your baby, regardless of the heating method.
It’d be like recommending against making instant oatmeal in the microwave...but you just need to heat the water separately!
I’m with Davidmanheim here, it seems this idea could benefit from reading in measurement theory, or at least recognizing a discrepancy that undermines the analogy. I’ll get into that a bit, but to start, the post was definitely positive food for thought.
If you’re measuring actual temperature, you have some measure options there too, but fundamentally it’s a quality of the material under study. If you’re measuring “the” perceived temperature, it’s an interaction between “the average person” and the material, and sticking fingers in is probably a good measure. Yes, temperature and perceived temperature will correlate, but if the thing you’re measuring exists only in someone’s head, you’re going to have to go to their head for the measurement (also see psychophysics).
“Train[ing] a net to replicate human reports” is not obviously less useful than “actual” scales. Human reports may in fact be the most construct-valid measure. (Though I do agree that leaving these reports in the form of natural language rather than attempted quantifications would indeed be ambiguous, and if we lack face-valid quantitative measures, we will have to develop them from somewhere, probably with those open-ended responses as a foundation.)
Although human reports may be noisy, so are all measures. The thermometer has an implicit +/- margin of error. It seems very precise to us, but human judgments of attributes can also be reliable (in that lots of people agree) and precise (in that the error bars are narrow). For example, if I asked a lot of people to rate the perceived precision of various measures on a scale of 1=extremely noisy to 100=extremely precise, I expect there to be a decent amount of consistency in the rank ordering of those ratings, for thermometers to score highly, and for at least some of the average perceived precisions to flash pleasantly narrow error bars.
But because even the lowest-variance perceptions vary a lot between people (vs. the variability in temperature readings from a thermometer), I do suspect you’re not going to get readings that are “approximately-deterministically” useful indicators for lots of perceptual domains, such as alignment. But you’ll get indicators that “far-from-deterministically-but-reliably” predict variance in criterion variables. In the end, we’re pessimistic and optimistic about the same things; I just don’t think it’s because human reports are inherently the wrong tool, it’s because the attribute of interest is a psychological construct rather than a conveniently-precisely-measurable-physical property. Again, the post was good food for thought—just as measurement of temperature has improved and gotten more precise (touch it → use mercury → use radiation), maybe the methods we use for psychological measurement will develop and improve, with hope for alignment.
One quibble, there was a little bait and switch from someone with a well-calibrated model whose calibration just hasn’t been well-evidenced, to...
You’ll hear people saying that X will definitely fuck everything up very soon.And it doesn’t.And when the catastrophe doesn’t happen, don’t over-update.Don’t say, “They cried wolf before and nothing happened, thus they are no longer credible.”
These people ARE no longer credible as they are not estimating 5% chances but 95% chances, and the lack of an event, rather than being consistent with their model, is inconsistent with their model.
Your point is still well-taken, and I think the switch is a natural reflex given the infrequency of pundits attempting to make well-calibrated or even probabilistic judgments. For example, it has been noticeable to me to see Jamie Dimon publicly assigning probabilities to differing recession/not-recession severity bins rather than just sticking to the usual binary statements often seen in that space.
Great point on trust. Here’s a recent paper by Eric Budish reflecting on the issue, via Tyler Cowen.
Thanks for doing your own research and laying out clearly what you think Bitcoin offers.
All these features of Bitcoin make it an attractive candidate for being a store of value and medium of exchange.
I think you’re mostly right on what features Bitcoin has, but I think you’re mistaken that they make it a good currency.
Bitcoin is scarce. No debate there.
Bitcoin has no counterparty risk. In theory, sure, but in practice as you note in your parenthetical, it will remain intermediated, even if it could deliver some gains.
Bitcoin is durable, portable, and divisible. No debate there.
Bitcoin is backed by real assets. It takes resources to create BTC, but that doesn’t mean it is backed by real assets. Look no further than the US Mint—it takes resources to create and maintain fiat currency.
If BTC has 1, 2, and 3, do those make it a good currency?BTC maximalists and goldbugs lament printing dollars (if the Fed is achieving its stated goal, the dollar should depreciate against a general basket of products at 2% per year, i.e., it only buys 98% of what it would the previous year). In a world where production is constant, stabilizing the supply of money should be sufficient to deliver a “sound currency” (one that buys the same general basket of products every year, i.e., 0% inflation). But we don’t live in that world; production increases. Thus to deliver a “sound currency,” the money supply needs to increase (albeit not as much as it needs to in order to achieve a 2% inflation target [aside: money demand is also important for these macroeconomic dynamics, but not critical to the discussion here]). A constant money supply (assuming BTC would be used as money) in the face of increasing production will lead to an appreciating currency (i.e., deflation [aside: there are also macroeconomic frictional reasons to prefer inflation to deflation but again, not relevant]). So much for storing value as a “sound currency,” its value is moving! That said, people can plan pretty well under a transparent monetary regime, so we don’t need 0% inflation, we just need a consistent inflation rate (hence the 2% target). If there are real production shocks, it would be nice to implement shocks to monetary policy as well to preserve the nominal economy; it needs to not just be a store of value but also a unit of account. A fiat currency is superior to BTC in this regard. Structural scarcity doesn’t necessarily make for a wonderful store of value that can also be used as a unit of account in the currency context. I agree with Sam Bankman-Fried that BTC can store value in the same way other assets store value. Gold bars store value (we don’t want to use them as currency). Company shares store value (we don’t want to use them as currency). BTC stores value. But that doesn’t stem particularly from its scarcity. The value comes from the beliefs about its scarcity and value; see GME stock in 2021. In this regard, BTC is just another asset in that its value derives from belief, albeit belief that must be cultivated rather than ordered out of the implicit barrel of a gun. But, the stability of its supply hampers its effectiveness as a unit of account in an economy susceptible to real shocks. And the fact that the perceived value of BTC is either as pure speculation (sell before everyone decides there’s no there there) or as the promise of useful currency in the future, differentiates it from other assets that have value even though we don’t think of them as currency. A gold bar can be used for physical things, a company share is a residual claim on the company’s assets. A dollar is backed by fiat/guns. A BTC-that-will-never-be-a-currency may not be completely devoid of value (novelty, medium of exchange in small circles, etc.), but it’s of very limited value or a game of musical chairs. The case for tremendous value and not being a game of musical chairs requires the existence of BTC-that-might-eventually-be-a-currency; because I believe that to be unlikely, you can surmise that I believe BTC to have little value (that doesn’t mean you can’t make -or lose- money trading it!).Blockchain (and central bank digital currencies [CBDCs]) offer the potential to further reduce frictions and risks in the intermediation structure of money (assuming they get transactions/second to compete, which is a big assumption). BTC gets to ride this wave, so that is good for its prospects as a medium of exchange. But practically, a sovereign will not relinquish its monopoly over the currency; if BTC looks like it will replace the dollar, the sovereign can regulate a market structure into existence that neutralizes the benefits of BTC vs. the fiat currency (see CDBCs, or something to make BTC operate less usefully).
Yes, BTC is as durable as, more portable than, and more divisible than current digital dollars, so that is good for its prospects as a medium of exchange. But as above, I don’t see these as truly unique selling propositions; these are advantages that can be eroded away by improvements to current payment networks.
So overall, BTC offers marginal benefits over digital dollars in intermediation, durability, portability, and divisibility that in the future can be competed or regulated away. It fails as a unit of account since its scarcity is ironically a problem, not a benefit. It stores value based on belief, but the collective belief is on shakier ground than many other assets that store value. It has features, but they’re not enough. IMHO
Related: probabilistic negotiation (linking to my comment). Because of asymmetric information about demand schedules in the individual one-off context, either you’re guessing or accepting their self-reports (i.e., I agree with Kokotajlo and Shlomi). As nice as probabilistic negotiation is in theory, practically you just hope to converge to splitting the surplus, and giving-in happens for whomever tires of the negotiation first. Depends on how much you know about your counterpart.
It’s much easier to set market prices where you have repeated transactions across participants, so “market” demand schedules (i.e., multiple unitary reservation prices) can be “learned” and the “market price” that enables value-maximization reveals itself. I appreciate that it’s harder at the individual level—bringing in probability allows working with individual demand schedules (i.e., multiple probabilistic reservation prices rather than a single unitary reservation price), but bringing in probability doesn’t exactly solve the problem because probabilities can only be learned through being furnished knowledge of the generating mechanism (e.g., Yudkowsky and Kennedy) or through repeated observation, the exact things that we assume we lack in this situation and that make this a problem in the first place.
Previously after Pfizer’s 11⁄5 interim report on Paxlovid in high risk patients, you said an 89%, 95% CI: [64, 97] (n=774) was certain enough to conclude efficacy but not certain enough to stop the trial because the CI was uncomfortably wide.After they had their 12⁄14 final report on Paxlovid in high risk patients, you said an 89%, 95% CI: [72, 96] (n=1379) looked good.At that time they also shared the interim report on Paxlovid in standard risk patients, showing 70%, 95% CI: [-8, 92] (n=854).Now after their 6⁄14 final report on Paxlovid in standard risk patients, you’re saying 51%, 95%: [-44, 83] (n=1145) is just an issue of sample size?
I’ve really appreciated all of your analysis and curating for us here, but I feel like we’re missing a lot about your internal model of evidence for these kinds of studies.
It was at that point I thought, “we’ve rediscovered Kant’s categorical imperative.”
Transitivity is a fundamental axiom necessary for a consistent utility function, which is central to rational choice theory. Sure, the potential for resource loss makes it more problematic for the agents you’re studying, but if you don’t have a consistent utility function to support your modeling in the first place, it’s already problematic for your studying of the agents. Put another way, you don’t even need to “reach” the coherence argument if you can’t get over the consistency bar.
Why is a resource central here? Consider (if it helps, also change “upgrade” to “switch”):
Let’s start with the simplest coherence theorem: suppose I’ll pay to upgrade pepperoni pizza to mushroom, pay to upgrade mushroom to anchovy, and pay to upgrade anchovy to pepperoni. This does not bode well for my bank account balance. And the only way to avoid having such circular preferences is if there exists some “consistent preference ordering” of the three toppings—i.e. some ordering such that I will only pay to upgrade to a topping later in the order, never earlier. That ordering can then be specified as a utility function: a function which takes in a topping, and gives the topping’s position in the preference order, so that I will only pay to upgrade to a topping with higher utility.
Surely one can notice the circularity problem without using a measuring stick.
On fluvoxamine, the FDA’s report includes additional analyses that even go beyond what I talked about regarding the NIH’s. Though I will say their discussion of the meta-analysis seemed a little disingenuous (though some comments in peer review can feel the same, so) - garbage in, garbage out is always a potential problem, and one should never hope for a meta-analysis to “substantially alter the assessment of the individual trials,” so failing to deliver on that is just par for the course and should not be viewed as a negative.But even just taking the meta-analysis at face value, the summarized evidence of efficacy is kinda weak (a reduction in severe disease that is consistent with reductions ranging from 0-50%...consistent with 0% means “While the FDA has concluded that the existing clinical data are insufficient to support the issuance of an EUA, these data suggest that further clinical investigation may be warranted.” At least it’s not consistent with, say, −10%-40%, in which case the FDA would presumably not have thrown that bone).