What’s the worst systematic bias in thinking on the future? My guess: too much abstraction. The far vs. near mode distinction was first noticed in future thinking, because the effect is so big there.
"population, technology and so forth have evolved with increasing speed throughout history" -- Population is a number. It is nonsensical to say that population evolves. Technology is a product of reason; it therefore does not evolve. Gelfer thinks "evolve" means "grow".
"Given how much perceptions around gender have changed in the past 50 years, if we accept a pattern of exponential development in such perceptions..." -- Now we see Gelfer thinks "evolve", "grow", "develop", and "improve" are all synonyms. He has no understanding of evolution, and a serious misunderstanding of culture and ethics. Basically your typical circa-1910 view that "everything always gets better" describes and explains all social change.
But the question Marshall Bolton raised is how knowing that they are in a simulation alters the experience (even if the simulation is locally flawless).
I'm not claiming that ems won't know they are in a simulation. I'm claiming that if they want to they can be in sims where they can't in practice tell it is a sim.
Hmm. I'd been envisioning a virtual world with many layers of abstraction for different purposes, like we see in video games today. Having re-read your 'How To Live In A Simulation' it seems you expect a more unified virtual physics, with the main simplification being a sharp cutoff in time and space at the boundary of what's relevant.
Having done a little investigation it seems like the technology for more general low-level physics simulation is more advanced than I had realised: https://www.youtube.com/wat...(And of course there's another century or so for this tech to improve, according to your Age of Em timing prediction.)
I'm still not sure this kind of simulation technique will be so outright superior, and therefore ubiquitous, that that the em world will always be indistinguishable from our world, or that future civilisations would be able to create simulations that totally fool their inhabitants without putting much effort towards achieving this goal in particular. But these do seem more plausible than I had previously realised.
Thanks for taking the time to discuss this, I feelstrongly about it and adore finding out additional on this topic. If feasible,as you acquire expertise, would you thoughts updating your blog with much moredetails? It is extremely useful for me.
Cartoony worlds mainly just seem more pleasant, rather than cheaper. Only tiny fraction of people would ever notice if a sim violated its claimed laws of physics, and for those people you can just wait until they notice something, then back up, fix the problem, and rerun them again without the problem.
Can you expand on that? Seems to me that if our world is a simulation, the abstractions it uses to model the laws of physics are far more watertight than we would expect without a LOT of work having been put into making them so. Even in the most technologically advanced video games today it's trivially easy to produce visual bugs, making objects clip through each other for example. And that's with an extremely restricted set of possible actions. In reality we can perform precise experiments on arbitrary pieces of matter, and every single time we do so we get the answer that physics tells us to expect. Even if it's possible to create a simulation as perfect as that, why would it be cheaper than a brain emulation, let alone 100x cheaper?
Also, didn't you argue in The Age Of Em that em virtual realities would be more cartoony than our reality, due to it being much easier to simulate a cartoony world than a detailed realistic world?
The cost to create a virtual reality that our minds can't ordinarily distinguish from reality is quite low (< 1% of) compared to just the cost of running our minds.
This seems obviously true to me, if only because it surely wouldn't be worth making em environments as detailed and seamless as the real world.
If we're in a simulation, it's an inordinately expensive one compared to what's necessary to make someone feel sufficiently comfortable to forget about their environment and be able to get on and do things in it. Whoever built this simulation apparently saw tricking us into thinking it's real as so important that they were willing to spend almost endless sums in order to fully achieve this goal.
It seems to me marshall bolton's point is (or should be) that ems would know the virtuality of their experience. That's the difference from us, even if you think the "we are in virtual reality" hypothesis is plausible.
Maybe you don't, but I do (know that I am not in virtual reality this very moment). The burden of evidence would also rest with you as it is such a nonsensical notion, you are promoting.
EMS are software. The brain processes information in a reasonably modular fashion, so someone unscrupulous could insert a little script that triggers religious awe or sexual desire in a captive EM anytime they interact with their "master" or perform more subtle memory deletion or rewiring of reward systems so that the Em will have no choice but to want to perform whatever task is set to them.
Morality aside, we are prevented from creating this sort of New Soviet Man in the flesh by our inability to make sufficiently precise changes to the brain's structure and chemistry, and by the certain knowledge that even a small mistake risks lobotomizing the subject. Software edits, however, can be perfectly precise, and if we don't like the result, we can simply reload the subject from before the procedure. You don't need a very high success rate of you can try each step however many times it takes to get it right. The advantage to slaves isn't that they can be forced to perform a task, it is that they can be edited in ways no free Em would accept, until they want to perform it.
Ems aren't much more easily reprogrammed that are humans. And your imagined slave still costs resources to create, and has to compete with free workers. How sure are you that slaves are more productive?
Why do we assume Ems would even be treated as people? In the real world, a person is time consuming to make, difficult to reprogram, and impossible to duplicate. Plus people with valuable skills tend to have family, friends and co-workers who will notice if they go missing and stand up for their rights.
EMS, by contrast, can be made instantaneously, reprogrammed easily (although a given attempt may not be successful), and duplicated without effort. More to the point, anyone with an isolated computer and low morals can create a customizable slave that can never escape and that no one has any reason to suspect is there.
Different, easily exploited, and valuable is a a very dangerous set of circumstances. A world where Em slavery is economically viable is one where a great many Ems will be slaves.
Thanks for your response. Sidenote: and thanks for your writings, especially on signalling, which has changed how I understand many of the workings of the world.
> I assumed he intended more content that that.
Hopefully we see a response from Joseph Gelfer that clarifies the matter.
>"could be different" is the weakest possible claim to make about a parameter estimate.
In this post, you quote your book: "one gender might end up supplying proportionally more workers than the other", which could be characterized as "the weakest possible claim to make about a parameter estimate" by the same process. Unfortunately I do not have access to you book, so I can not see what elaborations you make upon that issue.
With what I can see, it seems like you made some weak claims in your book (which is totally fine) and criticized for Gelfer reading beyond what you explicitly said (which is totally fine), but then you switch to reading beyond what Gelfer explicitly said on the basis that he made a weak claim.
I feel it is likely that I'm missing something about your thought process.
"population, technology and so forth have evolved with increasing speed throughout history" -- Population is a number. It is nonsensical to say that population evolves. Technology is a product of reason; it therefore does not evolve. Gelfer thinks "evolve" means "grow".
"Given how much perceptions around gender have changed in the past 50 years, if we accept a pattern of exponential development in such perceptions..." -- Now we see Gelfer thinks "evolve", "grow", "develop", and "improve" are all synonyms. He has no understanding of evolution, and a serious misunderstanding of culture and ethics. Basically your typical circa-1910 view that "everything always gets better" describes and explains all social change.
But the question Marshall Bolton raised is how knowing that they are in a simulation alters the experience (even if the simulation is locally flawless).
I'm not claiming that ems won't know they are in a simulation. I'm claiming that if they want to they can be in sims where they can't in practice tell it is a sim.
Hmm. I'd been envisioning a virtual world with many layers of abstraction for different purposes, like we see in video games today. Having re-read your 'How To Live In A Simulation' it seems you expect a more unified virtual physics, with the main simplification being a sharp cutoff in time and space at the boundary of what's relevant.
Having done a little investigation it seems like the technology for more general low-level physics simulation is more advanced than I had realised: https://www.youtube.com/wat...(And of course there's another century or so for this tech to improve, according to your Age of Em timing prediction.)
I'm still not sure this kind of simulation technique will be so outright superior, and therefore ubiquitous, that that the em world will always be indistinguishable from our world, or that future civilisations would be able to create simulations that totally fool their inhabitants without putting much effort towards achieving this goal in particular. But these do seem more plausible than I had previously realised.
Thanks for taking the time to discuss this, I feelstrongly about it and adore finding out additional on this topic. If feasible,as you acquire expertise, would you thoughts updating your blog with much moredetails? It is extremely useful for me.
Cartoony worlds mainly just seem more pleasant, rather than cheaper. Only tiny fraction of people would ever notice if a sim violated its claimed laws of physics, and for those people you can just wait until they notice something, then back up, fix the problem, and rerun them again without the problem.
Can you expand on that? Seems to me that if our world is a simulation, the abstractions it uses to model the laws of physics are far more watertight than we would expect without a LOT of work having been put into making them so. Even in the most technologically advanced video games today it's trivially easy to produce visual bugs, making objects clip through each other for example. And that's with an extremely restricted set of possible actions. In reality we can perform precise experiments on arbitrary pieces of matter, and every single time we do so we get the answer that physics tells us to expect. Even if it's possible to create a simulation as perfect as that, why would it be cheaper than a brain emulation, let alone 100x cheaper?
Also, didn't you argue in The Age Of Em that em virtual realities would be more cartoony than our reality, due to it being much easier to simulate a cartoony world than a detailed realistic world?
The cost to create a virtual reality that our minds can't ordinarily distinguish from reality is quite low (< 1% of) compared to just the cost of running our minds.
This seems obviously true to me, if only because it surely wouldn't be worth making em environments as detailed and seamless as the real world.
If we're in a simulation, it's an inordinately expensive one compared to what's necessary to make someone feel sufficiently comfortable to forget about their environment and be able to get on and do things in it. Whoever built this simulation apparently saw tricking us into thinking it's real as so important that they were willing to spend almost endless sums in order to fully achieve this goal.
It seems to me marshall bolton's point is (or should be) that ems would know the virtuality of their experience. That's the difference from us, even if you think the "we are in virtual reality" hypothesis is plausible.
Maybe you don't, but I do (know that I am not in virtual reality this very moment). The burden of evidence would also rest with you as it is such a nonsensical notion, you are promoting.
EMS are software. The brain processes information in a reasonably modular fashion, so someone unscrupulous could insert a little script that triggers religious awe or sexual desire in a captive EM anytime they interact with their "master" or perform more subtle memory deletion or rewiring of reward systems so that the Em will have no choice but to want to perform whatever task is set to them.
Morality aside, we are prevented from creating this sort of New Soviet Man in the flesh by our inability to make sufficiently precise changes to the brain's structure and chemistry, and by the certain knowledge that even a small mistake risks lobotomizing the subject. Software edits, however, can be perfectly precise, and if we don't like the result, we can simply reload the subject from before the procedure. You don't need a very high success rate of you can try each step however many times it takes to get it right. The advantage to slaves isn't that they can be forced to perform a task, it is that they can be edited in ways no free Em would accept, until they want to perform it.
Ems aren't much more easily reprogrammed that are humans. And your imagined slave still costs resources to create, and has to compete with free workers. How sure are you that slaves are more productive?
Why do we assume Ems would even be treated as people? In the real world, a person is time consuming to make, difficult to reprogram, and impossible to duplicate. Plus people with valuable skills tend to have family, friends and co-workers who will notice if they go missing and stand up for their rights.
EMS, by contrast, can be made instantaneously, reprogrammed easily (although a given attempt may not be successful), and duplicated without effort. More to the point, anyone with an isolated computer and low morals can create a customizable slave that can never escape and that no one has any reason to suspect is there.
Different, easily exploited, and valuable is a a very dangerous set of circumstances. A world where Em slavery is economically viable is one where a great many Ems will be slaves.
You don't know that you aren't in virtual reality this very minute.
Thanks for your response. Sidenote: and thanks for your writings, especially on signalling, which has changed how I understand many of the workings of the world.
> I assumed he intended more content that that.
Hopefully we see a response from Joseph Gelfer that clarifies the matter.
>"could be different" is the weakest possible claim to make about a parameter estimate.
In this post, you quote your book: "one gender might end up supplying proportionally more workers than the other", which could be characterized as "the weakest possible claim to make about a parameter estimate" by the same process. Unfortunately I do not have access to you book, so I can not see what elaborations you make upon that issue.
With what I can see, it seems like you made some weak claims in your book (which is totally fine) and criticized for Gelfer reading beyond what you explicitly said (which is totally fine), but then you switch to reading beyond what Gelfer explicitly said on the basis that he made a weak claim.
I feel it is likely that I'm missing something about your thought process.