0 sats \ 8 replies \ @elvismercury OP 16 Nov 2023 \ parent \ on: _Broken Money_ book club, part 4 bitcoin
This is an interesting addendum to @siggy47's point, above. If your friend has scienced the shit out of some market stuff, and gets alpha from it, is that not an argument that an entity [presumably] possessing an order of magnitude more data than your friend's company could get even more alpha?
You can nitpick the shit out of this hypothetical, and I have no interest in that, as I'm sure you don't either. But I'm very interested in why some orgs (your friend's company) can do better than the theoretical optimum, and other orgs (the govt) are generally recognized as not being able to?
Some deep theory-of-the-firm stuff here.
(meta: what will happen to this reply when your post disappears?)
This reminds me of the quant trading craze. Maybe I'm just a luddite invisible hand guy, but I don't think the variables of a market will ever be quantifiable.
reply
I don't think the variables of a market will ever be quantifiable.
The nature of the invisible hand is that people making better predictions in the market are favored. So even if the market as a whole will never be quantifiable, the market is competing on approximations that aim to quantify parts of it, and the invisible hand will guide the market closer and closer to making it quantifiable - even if it never gets there exactly.
Having said that, this leaves a bunch of room for pseudo-scientific claims of quantifying markets better and I'm on board with fully leaning into skepticism of it.
reply
Your response reminds me of the whole theory of technical analysis. I was interested in stock trading in a former life. I never trusted TA, but it is intriguing that human herd emotion can be charted and used as a predictive tool. Of course, "models" fail all the time, so these strategies never approach the level of science.
reply
TA is another beast all together I think. The kind of quantification I'm talking about doesn't make price predictions as much as it tries to say "who/why/how/etc are people most unhappy when using products in this market category."
It's related to TA in that it's trying to help you draw conclusions that are otherwise hard to draw though.
reply
You are doing this in a way when designing SN algorithms and even features, I guess. I remember when you gave Car a little jab: "See, you wanted bookmarks, but you dont use them." Predicting what people want and how they will react.
I'm very interested in why some orgs (your friend's company) can do better than the theoretical optimum, and other orgs (the govt) are generally recognized as not being able to?
This is too deep for me to tackle let alone right after I've met it. Everything obvious is probably something you've already considered.
In this specific case at least, their methodology isn't so much 'do something different and better with existing data' as it is 'other people collect data that is easy to collect but meaningful conclusions cannot be drawn from it.' Their meta-alpha is ignoring everything that's normally used to segment people and instead focusing on the intent of subjects and the fruitfulness of their intent, and only then drawing conclusions.
I don't know anything about this field but their methodology sounds like some level of innovation. So the question could be rephrased as why isn't the government good at innovation?
Governments were great at space stuff, but that's when governments were the only place space people could do space stuff.
As a free, exceptionally talented person, why would you go work for a government when the expectation is less pay and less impact? Governments can short employees on one or the other and probably stand a chance at competing for talent, but both? It's like choosing to move to the soviet union near its end.
(meta: what will happen to this reply when your post disappears?)
Replies will be preserved. The text of the comment will just disappear.
reply