greyenlightenment.com
greyenlightenment investments: META/FBL, TSLA, TQQQ, TECL, MSFT ...
1yr ago
(greyenlightenment.com)
Edited 1yr ago
3045 thread views
Nassim Taleb is likely wrong about IQ and talent
- 30
- 11
Jump in the discussion.
No email address required.
Notes -
I am going to link to this post as well because it's related and expounds on the above [Refuting common arguments against IQ and talent
](https://greyenlightenment.com/2022/07/13/refuting-common-arguments-against-iq-and-talent/)
Arguments alleging the unimportance of IQ or low correlations between innate factors and outcomes, have two flaws, those being such arguments ignore individual preferences and conditional probabilities. If you want to be good at something, you are going to be competing against other people who also aspire to be good at that very thing, not just people plucked at random. "Goodness" also means relative ability, not only absolute ability/improvement. If you go to gym and lift weights, it's reasonable to assume you will get stronger or stronger compared to someone who has never lifted, but eventually you'll want to know how good you are relative to others, too. If you want to be better relative to other people, this means having to compete against people who may have more aptitude for that very thing you are trying to get better at. In the case of weightlifting, this means people with a more favorable bone structure, favorable muscle insertions, or more fast-twitch muscles. Running? That would be lactic acid threshold, VO2 max, etc.
Correlations do not mean as much as people think they do and are misleading. When you analyze large groups of people, you tend to get low correlations. Short people can succeed at basketball, even in the NBA, but they face much more insurmountable odds compared to tall people. Most tall people do not play in the NBA (the correlation between being over 6 ft 6 and playing professional basketball is indeed small) but if one aspires to play in the NBA, being tall is pretty much a prerequisite despite this low correlation. Same for the correlation between wealth/income vs. IQ, which is low, but it's not like all high-IQ people aspire to high-paying professions. Some are writers, performance artists, painters, or don't want to move up the corporate ladder.
Of course there are benefits of doing things even if one does not aspire to being competitive. The health benefits of exercise may make it worthwhile in and of itself, even if you never get that good at it.
Successful, high-status people are not just good at things: they are good relative to other people too, which matters more. It's like great, I can do math at a 6th grade level; yes, this a major improvement over a 1st grade level, but this is not good enough to get hired at NASA. A 160 on the LSAT is an improvement over 140. Will that get you into a top law school ? likely not (unless everyone else also scores as poorly or worse). Just getting better at something is not good enough. If I sold a course promising a 20-point improvement on the LSAT, a lot of people would be interested, only to feel let down when they learn that this 20 point improvement is compared to guessing.
Second, the alleged low correlation between job performance and IQ is conditional on getting the job. Being hired typically involves two screening stages: sending the resume/inquiring about the job, and if successful, an interview. For good-paying jobs, the screening process means that people with low or average IQs are often weeded out. Likely there is a major cliff in competence between someone with an IQ of 95-100 vs 115-120. So employers want to ensure they are choosing from a pool of applicants who have credentials that correlate with at least + 1-sigma of IQ. But the 130-140 IQ guy may not be much better than the 120 guy, but the 120 guy is waaay better than the 95 guy.
My contra-argument to this would be that more strength has no value in the modern society. You need some strength to do most activities but as we have different power-tools, average worker is as much productive as the Olympic level athlete. And even if it would matter, the strongest man is only marginally better than an average male (in good physical health and training status). The group work is more important than individual strength since immemorial times. Even in stone age hunting mammoths required teamwork more than brute force.
Intelligence however is very different from physical strength. Maybe it has more value in the modern society. Again, some geniuses may make important discoveries that can benefit us all. But that may be very unpredictable and hard to measure anyway.
Taleb doesn't say that IQ doesn't matter at all. Definitely some people are smarter and therefore more successful. But the correlation of IQ with success probably maxes at certain limit.
It's possible, although imho very unlikely, advances in AI will render human intelligence redundant.
It does not matter much if you don't control for individual preferences or job screening. But people tend to compete against each other, such as for job openings or credentials (like grad school or law school). If I want to be a top physicist or mathematician or whatever, I am going to have to compete against the 130+ IQ guy, not the 90 IQ guy. It maxes out at a low threshold if one preference's are aligned in such a way that a high IQ is not an asset, such as a 120+ IQ person at working at fast food and refusing all promotions and never changing jobs.
Just an interesting observation from my pharmacist class in the UK. The biggest earners now are those who did average or even were on the verge of failing. It is just my anecdote. Obviously you need to be smart enough to pass the exam to get a licence. But once you pass this bar, in real life to be on top of academic knowledge is less important than perseverance and drive to succeed and social skills. IQ tests do not measure these things.
I have reason to doubt that the studies showing IQ correlation with success are true. It is very difficult to get clean data and eliminate bias. Just see how hard it was to prove that masks are/are not effective in containing covid. Scott deeply delved into this issue but got it wrong – final Chinese experience shows that masks are useless for all practical purposes (outside some very strictly controlled environments, like hospitals) and we were fooling ourselves with mask mandates (still current in some places in Europe). Probably studies with IQ are similarly unreliable. My trust in them is very low.
I’m very against masks (always thought evidence was insufficient to move past priors and believed there were huge downsides). Can you share the Chinese studies though? Would be interested in reading those.
There is no Chinese study I would trust. I based it on observation of great number of infected. It appears that the current Chinese wave is due to lifting lockdown restrictions but from the data Zvi has provided, it seems the opposite. The number of infected was rising and the government realized that it makes them look stupid and they decided to lift lockdown to shift the blame.
Going back to more reliable data – I remember that article on marginalrevolution.com where they showed that good data doesn't need complicated statistics to show that vaccines work. The graph of infected without vaccine was clearly so different from the one for recently vaccinated. It seemed so convincing. And yet, now we know that this conclusion was wrong. The vaccine protection from infection didn't last long, only 3-4 months on average. It might be protecting from severe disease and death (by about 90%) but that wasn't what this graph was trying to prove. It turns out you need a good and thorough statistics and study design even for things that seem obvious.
As a pharmacist I was trained to back up all my assertions by evidence. If you give a recommendation to the patient or a doctor to use a specific drug, they might always ask – what is the evidence? I should be able to answer this questions without delay. But when the masks started to be recommended, the best evidence I had seen was Scott's article. When I tried to ask doctors and others if they have more evidence about effectiveness about masks, they said that you don't need evidence for parachutes. The many reversals in this pandemic showed that many things that initially seemed like parachutes were not like them, therefore the question about evidence remained.
In the UK we usually rely on NICE guidelines. Not all evidence is of equal strength but NICE as the organization put a lot of effort to evaluate and list all available studies for given intervention. I am in favour of regulatory barriers for all new medicines. The only problem is that the regulatory process is too slow and might fail us during pandemics when quick decisions are required. I know that Scott is very critical of regulatory institutions but I cannot share his criticism because pharma companies try to push too many costly medicines with only a marginal benefit (especially for cancer). Sometimes even regulators make mistakes and approve doubtful medicines (like aducanumab) but they err on the opposite side – not approving useful drugs – only in very rare cases. I can only agree that the regulatory process is too slow, cumbersome and inflexible in some cases but the main service they provide is very important.
During covid pandemic nothing of this seemed to matter. Regulators were disregarded, politicians introduced measures without any proof of effectiveness. They even introduced vaccine mandates that we were previously taught to be a bad policy. My native country bought too many vaccines and now they are donating them to poor countries – again we were taught this to be a bad policy. Those countries most likely simply destroyed the donated vaccines. Usually it works much better to donate money so that they can buy what they need and not the excess stock. Other countries are unable to accept donations of medicines because they usually are not in the form their doctors are trained, they might not have labels in their local languages, they might not use the particular brands or strengths and supply chains might not be appropriate. For example, Pfizer had mRNA vaccines to be stored in -70°C. Then they made another variant that could be stored in -20°C, and then again another one to be stored in fridge (2°–8°C). The options in poor countries are more limited and if they have settled for one specific cold chain, all other vaccines that need a different conditions will never make to final recipients. This distribution is hard. We learned this when lockdowns disrupted our supply chains in the west. In poor countries supply chains are even more fragile and simply donating medicines have very little benefit if we don't consider how they will fit in their supply chains. But that didn't matter for politicians, they only needed to save their faces and pretend being good humanitarians.
More options
Context Copy link
More options
Context Copy link
More options
Context Copy link
More options
Context Copy link
More options
Context Copy link
More options
Context Copy link