What doesn't kill you might make you dumber?

[daylen]

I suspect researchers at any level that are not integrating AI into their workflow are becoming dinosaurs except in exceptional cases. Remember that AI is a much broader category than chatbots. The creative spark is essential but tends to get bogged down in the memorization of facts and trivial calculations. Researchers that offload some memorization and analysis to machines will have more time to think creatively on higher order patterns and problems.

[jacob]

I suspect researchers at any level that are not integrating AI into their workflow are becoming dinosaurs. Remember that AI is a much broader category than chatbots. The creative spark is essential but tends to get bogged down in the memorization of facts and trivial calculations. Researchers that offload some memorization and analysis to machines will have more time to think creatively on higher order patterns and problems.

I'm a dinosaur. I'm not so sure about the usefulness of the higher order patterns insofar they're not backed by lower order patterns. If "theory without practice is empty, and practice without theory is blind" then making either or or the other exogenous to the brain becomes limiting. In the former case, we get people who talk a good game, yet remain incapable of creating anything useful beyond reviews and summaries. The kind that talks a lot but doesn't say much. Having never focused on the practical details, not much of what they say contains actionable insight beyond the trivial. They might even make epic mistakes missing a pertinent detail that "any grunt on the ground floor knows". Much of the metacrisis-community is like that. An example would be a computer scientist who can't program their way out of a wet paperbag.

In the latter case, we get technicians. People following the "advice" of others (whether humans, google searches, or AI) without understanding why any suggestion/advice is a good idea or relevant. Punting theory to authority, we get rule-based thinking and slogan-based justifications. Currently these are the people "who did all their research because they read every article on facebook". Introducing AI would just change the argument to "I asked the AI" from "Google told me" (because if it's found on the internet, it's true).

Basically, there's abstract intelligence, but there's also experiential intelligence. Sternberg probably has a few books on that

I'm with @bsog here. Creative (the last C in the CCCCCC chain) thinking really does require mastering the previous 5. Access to AI, books, professors, can speed this up significantly, but they can not replace it. Since "reference speed" came up ... consider the difference between a brain that already knows the answer to a bottleneck detail... to a brain that either has to figure out what question to ask the AI; or worse doesn't even know that the question is relevant or what the question should be.

The primary difference between a "phd"-brain and a "masters"-brain, is that the latter is good at finding answers to questions; while the former is good at finding questions---basically making sense---of the answers (which it already knows). In FIRE terms, that's the difference between chapter 7 (making sense of the entire parameter space) in the ERE book and the standard blog series that focuses on the optimal portfolio for the 4% rule (finding the optimal technical answer).

IOW, in order to understand the forest, you have to understand what a tree is.

[daylen]

Though if you have two researchers that both understand what a tree is and are trying to understand the forest, then the researcher that collaborates with machine to account for more trees will get a better view of the forest. I suspect this to be much more relevant in biology and sociology than in physics and mathematics. Mileage may vary, though the most productive mathematicians tend not to do mental math (although they probably can).

[jacob]

Well, if you have both ... but I understood the point as being that AI would replace one or the other.

A collaboration between two people (or entities) who each understand half the problem will never ever be as effective as one person who understands both due to the limits of the Q&A channel of the former constellation. Internally the brain can convey concepts of a greater complexity than having to route them through the one-dimensional format of mouth/ear or keyboard/screen.

[daylen]

That's where increasing the channel capacity between brains and computers comes in.

[7Wannabe5]

There is the 'grad level' definition that points to a more granular/specific/exhaustive analysis than required in undergrad. Maybe this is Master's level thinking. AI might be really helpful with that. But 'grad level' can also be though of as a creative synthesis that builds upon and approaches the topic in ways that is different in kind. This requires not only knowing that Master's level analysis and information, but also manipulating it or holding a very large amount of it in mental buffer to work with it. At least, that's a way of framing certain PhD levels of thinking. As I understand it, AI as it currently is can't give you that (or, at least not much of it) - and a hit of a ~5 IQ points could be devastating to this particular skill set. But I don't really keep up with the AI thing...am I way off base?

Well, the way I write my Master's level papers is essentially the same method I use to construct a casserole, so by analogy I believe what you are suggesting is that current AI can't provide enough protein to produce a substantial and unique casserole. Since I am a lazy generalist, it is my tendency to go side-ways to more "unique" rather than forward-ho to more "substantial", so you are quite possibly correct. I did suggest that improvement might be more towards "efficency" than "effectiveness", but in the sense that IQ can be subbed for "competence/productivity at the job", simple efficiency gains may be sufficient. For example, at the juncture when I want to mix in 2 scholarly articles on topic to my paper in the manner I might mix 2 cups of chopped vegetables to my casserole, I could request AI summaries. Or at the juncture when I want to sprinkle a few interesting quotes, like bacon bits on the top of my casserole, I could prompt AI to do a deeper search into the literature than Google search might provide. If I am feeling too brain dead and uninspired to fluidly gush up the special sauce of language to bind all the ingredients together, I could prompt AI to draft a section, etc. etc. etc.

When the very high IQ human Claude (Shannon) was a student, his advisors purposefully chose to push him outside of his current domain to other only semi-related topics such as genetics. It seems like they were hoping that he might achieve some form of synthesis towards consilience. If a human prompter of AI wants to produce an output that is substantial and unique, like the advisors to human Claude, they would first have to ensure substantial knowledge (protein) base in all realms to be combined in the casserole, before attempting to go side-ways with synthesis. I haven't experimented enough with the models to comprehend the limitations to this approach. Maybe it's all still just mostly noodles and sauce.

[ducknald_don]

I think it's increasingly known that many if not most/all autoimmune diseases are various forms of inflammation which in turn cause the associated symptoms (pain in rheumatoid arthritis, etc.), It's true that one can peel the onion indefinitely, arguably all the way back to suboptimal nutrition choices and lifestyle in many/most cases, which then causes gut dysbiosis/gut inflammation, which in turn leads to leaky gut (and leaky BBB) which in turn freaks out the immune system (it's then dealing with foreign substances it was never designed to deal with) and confuses it so it attacks various tissues in the body. My curiosity is whether covid exacerbates these already potentially ongoing problems, or is a separate pathway that can create the maladies independent of an infected person's metabolic health.

I don't think there is much evidence for that, the consensus seems to be most autoimmune diseases are triggered by a virus which causes the immune system to go into overdrive. Often people develop these diseases at a young age when they are otherwise fit and healthy. There is very little correlation between something like type 1 diabetes and weight for instance.

[bostonimproper]

My brother has been experiencing long COVID symptoms for about 4 years at this point. He is currently on a plant based whole food (anti-inflammatory) diet to try and manage it. His take has been the diet has helped him feel healthier generally, but in ways that are distinct from the effects of long COVID (tiring more often, heart palpitations, brain fog). From the outside he seems as intelligent as ever and continues to be able to manage a large workload of complex and novel tasks well at his tech job, enough for regular promotions at least.

[Riggerjack]

On an individual level, 10 points is the difference between levels (engineer vs mechanic vs technician vs janitor), and 3--5 points may be the difference between promotion and stagnation. On a societal level even a few points of difference is a big deal because of the extreme tail effect. Shifting down the average intelligence is not a big deal for the median but shifting the tails substantially increases the number of dumb people while reducing the number of smart people in ways that will be culturally game changing.

Perhaps you can take comfort once you spend a few minutes to realize just how completely untrue this is.

High IQ is mostly a waste byproduct in my culture. There are far, far more high IQ people than there are niches requiring a high IQ. Separating high IQ individuals from competition for high IQ niches is what our education system is for.

A Covid IQ drop will result in slight degradation of the output from those niches, but the high IQ supply still vastly outweighs demand, as it has for generations. And those niches have been producing reduced output nearly as long. The reduced output wasn't critical, a slight degradation of that reduced output, will hardly be game changing.

Of course Covid ain't over, and shows no sign of ever being over. Which I expect to result in significant IQ differences to build over generations between populations that choose to host Covid, and those that don't. Differences the Covid hosts won't notice, naturally.

[Jean]

Our brain are getting dammaged all the time. Fighting against it is like fighting agains a reeceeding hairline.
But we can compensate for it up to a point.
Like how it's much easier to learn a third language than a second one, because you already know how to order the knowledge.

I'm not saying it's not worth trying to avoid things that make it worse, but it's pointless to worry about some of it happening.

So at a societal level, even if covid had an effect, is it really different than what happened regularly, or worse than other phenomenon happening now?

[ducknald_don]

@riggerjack You seem to be assuming the drop is only going to affect people with a high IQ, I would speculate that the opposite is the case. Certainly if you take the median IQ individual a drop of 10 points is going to leave them unable to handle a lot of modern work.

[jacob]

Occupational intelligence studies have been made: https://users.ssc.wisc.edu/~hauser/meri ... mplete.pdf (You want the bar diagrams in the appendix. Any will do.)

It's not the only one but the one I could find quickly. Pertinent is that each occupation has a range. Some occupations large ranges (intelligence doesn't matter so much compared to other qualities) and some have high lower cutoffs (high intelligence is not just a sufficient condition but a required one). Most occupations have a range of 30-40 points.

Another way of saying it is that a 10 point drop corresponds to dropping down a quartile within your occupation. If you're already in the bottom quartile, you're SOL. Job just got too complicated. If you were already a good performer, you'll have some runway, but if you still want to retain relative performance, you need to switch to a less complicated occupation.

Also consider how much easier or harder it is to work with a team (providing instructions, getting things done efficiently, dealing with problem solving, avoiding mistakes) in two teams with a 10 point difference. Here's a long list of the average IQ of college majors. Lets take a team of engineering majors and have them take a class in business administration and have a team of business majors take a class in engineering. The difference in how much each team will struggle learning new material is obvious to anyone who has ever tried this "experiment".

[Riggerjack]

@jacob

Right. So your absolute highest category is Physics, at 133. 133 just isn't very rare. 130+ is about 1% of the population. or about 80 million people on the planet reach this threshold. How many niches are there for high IQ physicists are there, again?

Conversely, 80M high IQ people in the world. How many of them are in niches that require a high IQ? How many of them ever get a chance to use their minds, professionally?

Once again, our education system is a set of obstacles set up to separate these high IQ people from competing for these few niches. Some few of them make it through to the niches. Still, far more than are needed.

What I'm pointing out isn't a criticism, or even a problem. Rather, this seems like an opportunity.

[7Wannabe5]

The difference in how much each team will struggle learning new material is obvious to anyone who has ever been part of this.

Very true. Another example would be how just the 4 point (130 vs 126) IQ difference between a Mathematical Sciences major like me and the many Mechanical Engineering majors I have dated made it difficult for them to fully comprehend my speech, whereas my son who was a Philosophy/Linguistics major has little difficulty, because his working vocabulary is larger than mine. Although I also think his high IQ causes him to drink far too much, and this makes me feel terribly sad. Also, I believe that my sister who kept flip-flopping between Fine Arts and Engineering has a slightly higher IQ than me, because she was able to design and build an animatronic drumming bear from scratch, but her mental health is also sometimes perilous to the extent that she ended up in jail in Detroit after exhibit of "performance art."

IOW, humans tend to become unstable/eccentric as IQ increases beyond 130, so you have to have some more stolid humans around 120 in the mix, because they are better at doing things like holding down a 40 hr/week corporate job, because it's a right-size challenge for them.

[Sclass]

Has the economy collapsed? If we have had a step function decline in IQ is there some kind of system response that we can readily observe?

I’m still not convinced IQ is a great predictor of say GDP. It’s a stretch. I look out my window and people seem as dumb as they were pre pandemic. Same idiocy Macdonalds drive thru for the masses. Still 5% of the people are doing 95% of the complex work. 95% don’t seem to make a big output response…or perhaps their contribution isn’t contingent upon their intellectual acuity.

It’d be nice if we had some easy metric like IQ and we could calculate GDP from it. We all love the predictive power of science. But this isn’t f=ma. It’s muddy.

I’m afraid a tiny input is responsible for most of the output. Perhaps the vast majority of the people have little impact. +/-10 IQ so what? It may not be applied where it counts towards the output. Wealth inequality is an output that gives us glimpses into our system’s actual wiring.

When I don’t really know the laws that govern a system I start by injecting a step function into the input and I measure the output. Historically there have to be some similar examples like economic recoveries after war or famine.

[7Wannabe5]

I’m afraid a tiny input is responsible for most of the output. Perhaps the vast majority of the people have little impact. +/-10 IQ so what? It may not be applied where it counts towards the output. Wealth inequality is an output that gives us glimpses into our system’s actual wiring.

Also the fact that a lot of high IQ folk tend towards being drop-outs. For instance, one of my uncles was a typical highly radical1960s drop-out. He was part of one of the more dangerous anti-Vietnam War groups organized around a major university. He wandered around for many years, and finally married a woman from the Phillipines when he was in his 40s, and they had a couple kids he ended up raising on his own after she died young. One of his kids became very successful in the tech world, and my uncle through tech-world osmosis ended up with the first high-paying job of his life when he was in his late 60s!

ETA:

Gifted students also demonstrate specific negative traits, and the list is extensive.
Davis and Rimm‟s (2004) studies showed that gifted students often exhibit interpersonal
or social difficulties, precocious demeanor, underachievement issues, noncompliant
attitudes, an urge for nonconformity, extreme emotionalism, over activity, edginess,
stubbornness, impatience, absentmindedness, argumentativeness, extreme perfectionism
or extreme sloppiness, self-criticism, or anger. Extreme physical and emotional
sensitivity and idealistic standards also surfaced (Perrone et al., 2007). Other negative
traits included an obsession with justice and fairness, hypersensitivity, extreme detailorientation, unusual sleep patterns, and suicidal fantasies (Cassady & Cross, 2006).
Gifted students may have an inability to finish tasks before starting new ones and oftendemonstrate a compelling need for new material and information (Caruana, 2002). Many
gifted students have a deep need for comprehension and constant mental stimulation
(DeLacy, 2000). They must clearly understand a concept and become frustrated when
they do not. DeLacy‟s findings also support Caruana‟s data showing that gifted students
consistently need new information and less repetition

https://core.ac.uk/download/pdf/58827272.pdf

[chenda]

one of the more dangerous anti-Vietnam War groups

What did they do ?

[Riggerjack]

Certainly if you take the median IQ individual a drop of 10 points is going to leave them unable to handle a lot of modern work.

Well, if they have their certifications, IQ isn't necessary. Or at least that was my corporate employment experience. If they don't have the certification, and the IQ drop prevents them from achieving their certification, someone else will get the certification, and the job. Very, very few jobs are structured so higher IQ is likely to produce significantly different results. Same for lower IQ.

IOW, humans tend to become unstable/eccentric as IQ increases beyond 130, so you have to have some more stolid humans around 120 in the mix, because they are better at doing things like holding down a 40 hr/week corporate job, because it's a right-size challenge for them.

Or perhaps creating an environment with features that attract high IQ, without the constant negative feedback for high IQ my culture generates, would be appropriate. Match the environment to the "problem", rather than trying to solve the "problem" for the default environment.

Although I also think his high IQ causes him to drink far too much, and this makes me feel terribly sad.

Has he ever really had a chance to put his talents to use, and test his capabilities? I also drank far too much before I learned to create my own challenges.

Outliers get less satisfaction from default options, and more negative feedback. Removing default options, replacing them with custom options has been very helpful for me.

[7Wannabe5]

@chenda:

Another SDS faction became known as the Revolutionary Youth Movement (RYM). RYM took its inspiration from a 1965 Bob Dylan song, “Subterranean Homesick Blues,” which had an enigmatic line: “You don’t need a weatherman to know which way the wind blows.” The SDS-RYM faction embraced the name “Weathermen.” The Weathermen hoped to launch a guerrilla insurgency in the United States. As they chanted, “Bring the war home!”, they attempted to assassinate police officers and soldiers, rob armored cars and banks, burn campus Reserve Officers’ Training Corps (ROTC) buildings, and plant bombs in corporate offices.

I'm not sure that he was that violently radical, although I do know that he was arrested for distributing anti-war information in Berlin, and I also know for a fact that he had a very high IQ, because it was a bit of a competitive matter with my father, because my highly intelligent patrician-like grandfather somewhat favored the extremely intelligent baby of the family, even though he was a radical drop-out.

[Riggerjack]

What did they do ?

https://www.amazon.com/Days-Rage-Underg ... 0143107976

There was a time in America, before we coined the term "domestic terrorist", when political violence was just swept under the rug. Easier, when one only has to control a few newspapers and 3 channels.