"It will take 50,000 years to terraform mars" variety.
They present a plan, and sure, if you follow that plan, it will take 50,000 years. Would a different better plan manage to terraform mars faster.
When thinking about future technology, you need to be prepared for someone cleverer than you coming up with a better plan.
There is also the variation that says "X is growing at a rate of Y per year, so in 100 years, it will be 100Y bigger", or similar. Long timescale predictions that implicitly assume no humans or AI will mess with [thing].
Geologists predicting where continents will be millennia into the future. In terms of the geology, sure they may well be correct. The prediction doesn't apply if earth is disassembled to build Dyson sphere.
You could theoretically have an argument where you use evolutionary psychology to construct a complicated explanation for why humans burst into uncontrollable laughter on seeing the color purple. And then briefly mention that, as everyone knows, society recently invented purple paint and started painting everything purple.
Or, to give a more realistic example, you could use evolutionary psychology to explain why humans like sweet fatty food, and combine that with the relatively recent large increases in industrial sugar production in order to explain a recent obesity crisis.
And if you focus on the evolutionary psychology part of that explanation and barely mention the recent increases in sugar production, it can look like a framing error.
Yeah this is fair. I think the thing I’m pointing at is that *it’s suspicious* for an evo-psych result seems to explain the last 50 years and not the time before (or explain results in WEIRD countries and not non-WEIRD countries), and especially when people don’t realize why this is suspicious.
I’d find their explanations much more convincing if they tried to reconcile this by explaining the breaks from both the evolutionary ancestry environment and (as the case might be) even the remaining time between the EAE and 100 years ago.
Fantastic post, and I think you highlight an underappreciated failure mode, although I admit I’m doubtful of my ability to apply this lesson at least beyond what I was already doing intuitively
Regarding your suggestion that people who can find meta street light bias examples, the first example that comes to mind is the claim in most economic analysis of regulation that there are four market failures, which regulation can correct externalities, public goods, information asymmetries, and market power. To me, this seems like a clear case of confusing failures of free markets which are obvious from usual economic model with the actual set of failure that humans care about. To list examples of other possible failures of the top of my head, there human irrationality where peoples willingness to pay doesn’t actually correspond with what we would consider their well-being, things that are valuable, even though people are not willing to pay for them to the full extent of their value because their value is independent of human preference, like animals, not being tortured or the environment, diminishing marginal utility of money, meaning that redistribution can prove good, even if it’s not directly obvious from many economic models. I also think people often even miss things while talking about those four failures which are actually part of those failures but are not so obvious from their usual model. For example, it’s an externality that you might not reproduce even when your child would really be willing to pay a lot to be created, but there is no credible way to set up such a transaction credibly, yet nobody thinks of this problem in this particular way, most of the time, and it’s not what they generally think of as an externality.
A second such example, although one where I’m less sure is the arguments that insect welfare dominates all other causes because this obviously un non-zero possibility, they might be conscious and their share number would cause them to dominate all calculations. I think this is an artefact of a frame where you assume that individual animal brains are the only site of consciousness, and if you went around asking things like whether this or that part of other animal brain is individually conscious, you could equally end up with something else, totally dominating calculations. I get that that is not a great example, but it’s just intuitively obvious to me that it’s an artefact of the way you’re framing the question when estimating how much is expected suffering in the world, and if you didn’t apply assumptions that amount to assuming that consciousness generally takes on a form like it does with our fellow humans, then you would not necessarily get the same overwhelming priority for insect welfare, although I admit I am uncertain, and this might be complete gibberish. I just have this nagging sense that those calculations that prioritise insect welfare so much are just a modelling artefact and in reality, if you admit more of our sheer confusion and consciousness, instead of overestimating things, bye, elevating a particular hypothesis of consciousness in probability by specifically asking for it They don’t dominate expected value calculations to quite that extent, but perhaps I’m not thinking clearly. What do you mean by that? Is that if you said even any Small number for whether an individual animal is conscious, that actually ends up, promoting that particular hypothesis of it being conscious to excess probability because you artificially an accidentally narrow down the scope of theory is considering which gets you an overestimate, but I could be completely wrong about this, and I’m not very confident.
One type of framing error I have seen is the
"It will take 50,000 years to terraform mars" variety.
They present a plan, and sure, if you follow that plan, it will take 50,000 years. Would a different better plan manage to terraform mars faster.
When thinking about future technology, you need to be prepared for someone cleverer than you coming up with a better plan.
There is also the variation that says "X is growing at a rate of Y per year, so in 100 years, it will be 100Y bigger", or similar. Long timescale predictions that implicitly assume no humans or AI will mess with [thing].
Geologists predicting where continents will be millennia into the future. In terms of the geology, sure they may well be correct. The prediction doesn't apply if earth is disassembled to build Dyson sphere.
> Evolutionary psychology (millennia timescale) explaining 50-year behavioral shifts
You could theoretically have an argument where you use evolutionary psychology to construct a complicated explanation for why humans burst into uncontrollable laughter on seeing the color purple. And then briefly mention that, as everyone knows, society recently invented purple paint and started painting everything purple.
Or, to give a more realistic example, you could use evolutionary psychology to explain why humans like sweet fatty food, and combine that with the relatively recent large increases in industrial sugar production in order to explain a recent obesity crisis.
And if you focus on the evolutionary psychology part of that explanation and barely mention the recent increases in sugar production, it can look like a framing error.
Yeah this is fair. I think the thing I’m pointing at is that *it’s suspicious* for an evo-psych result seems to explain the last 50 years and not the time before (or explain results in WEIRD countries and not non-WEIRD countries), and especially when people don’t realize why this is suspicious.
I’d find their explanations much more convincing if they tried to reconcile this by explaining the breaks from both the evolutionary ancestry environment and (as the case might be) even the remaining time between the EAE and 100 years ago.
Fantastic post, and I think you highlight an underappreciated failure mode, although I admit I’m doubtful of my ability to apply this lesson at least beyond what I was already doing intuitively
Regarding your suggestion that people who can find meta street light bias examples, the first example that comes to mind is the claim in most economic analysis of regulation that there are four market failures, which regulation can correct externalities, public goods, information asymmetries, and market power. To me, this seems like a clear case of confusing failures of free markets which are obvious from usual economic model with the actual set of failure that humans care about. To list examples of other possible failures of the top of my head, there human irrationality where peoples willingness to pay doesn’t actually correspond with what we would consider their well-being, things that are valuable, even though people are not willing to pay for them to the full extent of their value because their value is independent of human preference, like animals, not being tortured or the environment, diminishing marginal utility of money, meaning that redistribution can prove good, even if it’s not directly obvious from many economic models. I also think people often even miss things while talking about those four failures which are actually part of those failures but are not so obvious from their usual model. For example, it’s an externality that you might not reproduce even when your child would really be willing to pay a lot to be created, but there is no credible way to set up such a transaction credibly, yet nobody thinks of this problem in this particular way, most of the time, and it’s not what they generally think of as an externality.
A second such example, although one where I’m less sure is the arguments that insect welfare dominates all other causes because this obviously un non-zero possibility, they might be conscious and their share number would cause them to dominate all calculations. I think this is an artefact of a frame where you assume that individual animal brains are the only site of consciousness, and if you went around asking things like whether this or that part of other animal brain is individually conscious, you could equally end up with something else, totally dominating calculations. I get that that is not a great example, but it’s just intuitively obvious to me that it’s an artefact of the way you’re framing the question when estimating how much is expected suffering in the world, and if you didn’t apply assumptions that amount to assuming that consciousness generally takes on a form like it does with our fellow humans, then you would not necessarily get the same overwhelming priority for insect welfare, although I admit I am uncertain, and this might be complete gibberish. I just have this nagging sense that those calculations that prioritise insect welfare so much are just a modelling artefact and in reality, if you admit more of our sheer confusion and consciousness, instead of overestimating things, bye, elevating a particular hypothesis of consciousness in probability by specifically asking for it They don’t dominate expected value calculations to quite that extent, but perhaps I’m not thinking clearly. What do you mean by that? Is that if you said even any Small number for whether an individual animal is conscious, that actually ends up, promoting that particular hypothesis of it being conscious to excess probability because you artificially an accidentally narrow down the scope of theory is considering which gets you an overestimate, but I could be completely wrong about this, and I’m not very confident.