> Human beings who spend their lives studying the state of the world, in other words, are poorer forecasters than dart-throwing monkeys, who would have distributed their picks evenly over the three choices.
"He picked two hundred and eighty-four people who made their living “commenting or offering advice on political and economic trends,” and he started asking them to assess the probability that various things would or would not come to pass, both in the areas of the world in which they specialized and in areas about which they were not expert"
This is a little different from what I was suggesting. Predicting the future is an inherently difficult task for experts and I'll give you that.
Rather I'm suggesting that experts set a valuation based on known facts about the company and the state of the industry.
Here's an example. I work in the self driving industry. I always knew based on state of the art tech that Tesla couldn't possibly meet their aggressive timelines of full self driving. I think it's realistic one day, but not on the timeline they claimed a couple years ago. However, the reality is that the public is duped into thinking they'll be on time, and so what happens? I have to base my investment based on what the public thinks of Tesla, rather than what I think of Tesla.
Here's another example. Uber car crashes. NVIDIA stock crashes the next day because the public thought it had something to do with NVIDIA GPUs. I, and all other ML experts should have been able to call this out and freeze the price of NVDA and say "we have it on good authority that NVDA did not do anything to cause that car crash".
The current way the markets work shifts the paradigm from "invest in what you believe in" (which I is the way the world should work) to "invest in what you don't believe in but what everyone else is fooled into believing in".
I mean, the incentive structure is such that I've even bought stocks in many companies I hate, and sold or put stocks in companies I believe in from my scientific background on first principles. Somehow that isn't the way things should work.
> Uber car crashes. NVIDIA stock crashes the next day because the public thought it had something to do with NVIDIA GPUs. I, and all other ML experts should have been able to call this out and freeze the price of NVDA and say "we have it on good authority that NVDA did not do anything to cause that car crash".
See, but you don't actually know that's why the price of nvidia fell, nor do you know that it is an irrational reason for the nvidia stock to fall. Just because you work in the industry doesn't make you an expert on pricing.
There's lots of problems with this idea (it's incoherent wrt supply&demand, why should people closely affiliated with an industry be the one's setting prices, prices should reflect future expectations not current valuations).
> Predicting the future is an inherently difficult task for experts and I'll give you that.
But that's exactly what prices do. Why would I buy any stock at all unless I'm making a forward prediction about what is happening?
> Why would I buy any stock at all unless I'm making a forward prediction about what is happening?
That's exactly what I'm proposing we should change. We should be trading on a company's ability to make the world a better place. That would align incentives across people in a much better way than earnings reports. It also, incidentally, ensure that a company's profit structure is designed to align with its mission rather than align with some arbitrary earnings report deadlines.
I realize that metrics for this are hard to design, but we're in the 2nd millenium A.D. and it's okay for us to revise how we think about money.
I believe in a future of clean energy electric power, and it's f*ed up that I can't always invest in it because there are times when investing in oil gets me more $ that I can use to improve my own personal clean energy efforts while investing in electric would cause me to be at high risk of losing money, and therefore depend on oil because it's cheaper. That's a really messed up, convoluted system right there.
> I realize that metrics for this are hard to design, but we're in the 2nd millenium A.D. and it's okay for us to revise how we think about money.
It's not just that the metrics are "hard" to design. It's a question of why would I buy stocks based on a company's ability to make the world a better place. What makes the "stock" piece of paper worth more when the company does more to make the world a better place. For real-life stocks, it's the expectation of future buybacks/dividends. For your metric, it's unclear why anyone would participate.
> why would I buy stocks based on a company's ability to make the world a better place
I'm precisely proposing that we change the incentive structure in a future version of the economy such that it would benefit you to invest in a company's ability to make the world a better place.
That would eliminate OP's problem, among lots of other things.
I'm proposing that the role of investors should be fully aligned with the role of scientists and engineers, and the incentives should be designed such that that is the case.
Captialism 2.0, if you will. It's a quarter-baked idea, and there are lots of questions to answer, but I'd like to kickstart and encourage more thought and discussion about a future revamped economic system that aligns incentives better than what we have today.
You might have some advantages, but you have 10x more blind-spots.
What you believe about Tesla evaluation based on Self Driving might not matter because even if they miss that timeline, maybe they execute on their scaling or batteries or whatever.
A huge company has many, many different things going and different analysts think different things matter far more.
So you would need a whole group of these 'experts' working together evaluating each aspect.
Now even worse is that expert don't just have hard time predicting the future, but are actually the worst at doing it in many ways. Tesla and electric cars being a perfect example. The amount of car insider that believed electric cars were absolutely not viable as a buissness was tiny.
Even worse is that even if assume you know that some technology X will not hold up to the companies claims, their should maybe still go up based on competitors. I don't believe that Tesla can do Self Driving in the time-frame they claim, but the news about the competitors is even more worrying.
Literally everything is connected in extremely complex way that no experts can understand. The temptation for smart people and expert to think they can is a deceit.
> Uber car crashes. NVIDIA stock crashes the next day because the public thought it had something to do with NVIDIA GPUs.
This is unlikely. More likely the stock crashes because the event can be expected to decrease public trust in and enthusiasm for ML in general.
Blame for a one-off-event in a specific market usually isn’t a big deal. Adverse consumer sentiment with long-lasting impact on your entire bottom line is.
That too is a problem. The public shouldn't be the ones to decide whether or not we trust self-driving cars as a society. Science and hard data should decide that. Valuations should be determined accordingly, based on that science. In this particular case, the valuations of the entire self-driving industry should have been pegged by experts in the industry while the valuation of Uber should have been knocked down for negligence.
Trusting the public to set valuations on a per-trade basis directly results in the heavy hands at the likes of SoftBank scooping up full control.
https://www.newyorker.com/magazine/2005/12/05/everybodys-an-...
> Human beings who spend their lives studying the state of the world, in other words, are poorer forecasters than dart-throwing monkeys, who would have distributed their picks evenly over the three choices.