3 ms·

> Not via selective breeding Yes via selective breeding. Look into Landian analysis of urban centers as “IQ shredders”. Cities attract the smartest people and

by wyager 5y ago

> Not via selective breeding Yes via selective breeding. Look into Landian analysis of urban centers as “IQ shredders”. Cities attract the smartest people and then tank their birth rate. > we've mostly neutered evolutionary pressures on ourselves. In ten thousand years, this might be a minor problem It’s going to be a problem much sooner than that. Look into Yudkowsky’s analysis of evolutionary error correction bandwidth.

wizzwizz4 5y ago
> Cities attract the smartest people and then tank their birth rate. That's pretty simplified – but even if it wasn't, we're not going to run out of “clever genes” because it's not as simple as that. (But even if it were, the selection isn't strong enough for that to happen.) > It’s going to be a problem much sooner than that. If our artificial selection pressures remain the same. Which they won't. Two hundred years ago, our society's “artificial selection” was completely different; in two hundred years, it'll be completely different again.
- wyager 5y ago
  > the selection isn't strong enough for that to happen Your model is busted - the Flynn effect has already reversed and intelligence is decreasing again. You’re right that it’s “not as simple” as my one-sentence summary but my claim is directionally correct. > it’ll be completely different again The risk has nothing to do with what the concrete pressures are, but how much error correction bandwidth we have vs background mutation load rate.
  - TameAntelope 5y ago
    It's far from a conclusion that the Flynn effect has certainly reversed, for what it's worth. The 2018 study most often cited is on a small group of military aptitude tests from families in Norway and hasn't been replicated since.
  - wizzwizz4 5y ago
    > Your model is busted - the Flynn effect has already reversed and intelligence is decreasing again. Which metric are we using to measure intelligence this time? How do you know it gives meaningful results for what you're trying to use it to measure? (Please don't say IQ.) And has the study been replicated? > The risk has nothing to do with what the concrete pressures are, but how much error correction bandwidth we have vs background mutation load rate. But the selection effects you've described are all local and weak. Worst-case scenario – absolute worst-case scenario – the city Americans marry rural Africans and the problem goes away. But it won't get to that point. Eliezer Yudkowsky knows a lot about philosophy of thought, and AI theory, but he hasn't put as much effort into learning about genetics. If the average geneticist isn't worried, chances are that he's made a mistake in his reasoning (or that you're misunderstood his concerns) – and I haven't heard anybody in a biological field worrying about this. --- Wait, are you talking about his “in the absence of selection pressure, we'll get mutations that aren't filtered out”? If so… that's a hundreds of generations kind of problem, and I'm confident that stuff like IVF embryo selection (selecting against the genes for known genetic conditions) is already counteracting that effect. (Yes, this is technically eugenics, yes it could get dystopian if we're not careful, no it doesn't justify what the word “eugenics” makes you think of, no I shouldn't feel the need to write this parenthetical, yes this is the internet we're talking about.) Transposons are likely to be a bigger issue. I read that they've started duplicating themselves more recently, and we don't really know why. But imagine the human genome (for some unidentified reason that hasn't happened in the last thousand years) starts degrading across the entire world, so we only have a hundred years left before no human fœtuses are viable. We could just GM a human embryo to be the genetic child of modern-day humans (whose genome we've already sequenced), and we roll-back any mysterious genetic problems that far-future humans might develop. There is an upper bound to how bad things can get, even imagining something out of a sci-fi novel, so I am highly suspicious of the motives of people who get loudly concerned about the “decline of the human genome”.
krapht 5y ago
I'm pretty sure the future is going to look more like Gattaca than Idiocracy. I wouldn't bet against genetic engineering.
- wyager 5y ago
  This is one possible way of artificially boosting error correction capacity. I hope you are right.

by wyager 5y ago

wizzwizz4 5y ago

> Cities attract the smartest people and then tank their birth rate. That's pretty simplified – but even if it wasn't, we're not going to run out of “clever genes” because it's not as simple as that. (But even if it were, the selection isn't strong enough for that to happen.) > It’s going to be a problem much sooner than that. If our artificial selection pressures remain the same. Which they won't. Two hundred years ago, our society's “artificial selection” was completely different; in two hundred years, it'll be completely different again.

wyager 5y ago
> the selection isn't strong enough for that to happen Your model is busted - the Flynn effect has already reversed and intelligence is decreasing again. You’re right that it’s “not as simple” as my one-sentence summary but my claim is directionally correct. > it’ll be completely different again The risk has nothing to do with what the concrete pressures are, but how much error correction bandwidth we have vs background mutation load rate.
- TameAntelope 5y ago
  It's far from a conclusion that the Flynn effect has certainly reversed, for what it's worth. The 2018 study most often cited is on a small group of military aptitude tests from families in Norway and hasn't been replicated since.
- wizzwizz4 5y ago
  > Your model is busted - the Flynn effect has already reversed and intelligence is decreasing again. Which metric are we using to measure intelligence this time? How do you know it gives meaningful results for what you're trying to use it to measure? (Please don't say IQ.) And has the study been replicated? > The risk has nothing to do with what the concrete pressures are, but how much error correction bandwidth we have vs background mutation load rate. But the selection effects you've described are all local and weak. Worst-case scenario – absolute worst-case scenario – the city Americans marry rural Africans and the problem goes away. But it won't get to that point. Eliezer Yudkowsky knows a lot about philosophy of thought, and AI theory, but he hasn't put as much effort into learning about genetics. If the average geneticist isn't worried, chances are that he's made a mistake in his reasoning (or that you're misunderstood his concerns) – and I haven't heard anybody in a biological field worrying about this. --- Wait, are you talking about his “in the absence of selection pressure, we'll get mutations that aren't filtered out”? If so… that's a hundreds of generations kind of problem, and I'm confident that stuff like IVF embryo selection (selecting against the genes for known genetic conditions) is already counteracting that effect. (Yes, this is technically eugenics, yes it could get dystopian if we're not careful, no it doesn't justify what the word “eugenics” makes you think of, no I shouldn't feel the need to write this parenthetical, yes this is the internet we're talking about.) Transposons are likely to be a bigger issue. I read that they've started duplicating themselves more recently, and we don't really know why. But imagine the human genome (for some unidentified reason that hasn't happened in the last thousand years) starts degrading across the entire world, so we only have a hundred years left before no human fœtuses are viable. We could just GM a human embryo to be the genetic child of modern-day humans (whose genome we've already sequenced), and we roll-back any mysterious genetic problems that far-future humans might develop. There is an upper bound to how bad things can get, even imagining something out of a sci-fi novel, so I am highly suspicious of the motives of people who get loudly concerned about the “decline of the human genome”.

krapht 5y ago

I'm pretty sure the future is going to look more like Gattaca than Idiocracy. I wouldn't bet against genetic engineering.

wyager 5y ago
This is one possible way of artificially boosting error correction capacity. I hope you are right.