Question: What is 1 + 1
Answer: “what would jesus do?”....
Not helpful is it… Wouldn’t it be better to have a cognitive model that knows how to process data rather than reaching for the cheat button?
Even if the question was “If a man was drowning etc etc” the answer “what would jesus do?” is never going to be as effective as having a data processor that can custom build an answer for the exact question...which isn’t what X would do but what is the right answer.
Yes, that would be ideal, but a current human brain is not going to work for that. Until there is practical brain augmentation or otherwise accessible advanced AI, a set of role-models would help.
No i’m not perfect and I have biases come to my attention and fly under the radar etc- but I don’t ask myself what I would do. I don’t ask myself what someone else would do. I literally have no role models and can’t think of any I ever had. I do make decisions as they come up and if I ever was to base one off the fact that “that’s what Dan would do” then that throws up a red flag to me. It says ask the question again and find a real answer because maybe I don’t have a real reason.
What you are describing to me sounds like a short cut to a nasty bias that self perpetuates- telling you to never question anything just follow the status quo.
What the narratives would do would be to give you time to consider those situations and resolutions without actually being forced to do so on the spot. If you read it and understand that it is what you would want yourself to do in that situation, then you will have that solution on hand without your extensive on-the-spot calculations. If you think the resolution you just read in the Least Wrong was complete crap, then you would try to figure out what a better solution would be, again without the time-pressure. DO question the status quo, if you disagree with it. The point is to make them so good that you wouldn’t disagree to begin with, and would be happy to have the help that (mortals not blessed with infinite and instantaneous cognitive resources) could use in uncommon but important decisions.
I guess i don’t “want” myself to do anything. I don’t decide what is right in advance because if i do anything to predetermine my answer before a question arises then i’m starting off with a bias.
In a way what i’m saying is 1+1 could equal 2 tomorrow (which it can’t) and I will still probably get the answer right because i didn’t decide to stick with the answer 1+1=2 before the question was asked (therefore before this mysterious universe switched the answer).
I’ll sound completely biased and unbelievable when i say this but i’ll say it anyway- in my experience of breaking down my expectations for who “Dan” is and what “Dan does” i’ve made really good choices for the good of everyone around me. People around me have a model of what Dan apparently is which is empathetic, nice, generous etc. I’m always the first to point out a bias such as racism or nonfactual emotional opinions etc. I don’t have to see myself as any of those things though. All I have to do is keep asking questions properly and at the right time and then output a response. No i’m not a calculator but the results are good according to everyone I meet and interact with.
“The point is to make them so good that you wouldn’t disagree to begin with”
The problem with that is if you fix an answer like cement in to your brain based on one set of data- even if the data changes later you will have this cement lump in your head saying it’s “so good that you wouldn’t disagree” and so you don’t recalculate. I mean why would you calculate an answer you already know?
What you really need to do is not make accurate biases to pre-determine or influence your answer but work on removing all your layers so you make the calculations properly and unbiased. That way you won’t have to worry about if you dance or not- which ever one is right will be determined when the question comes up.
Again this isn’t just a theory of what i think you should do- this is what i do so don’t tell me it isn’t possible.
+1 for sharing; you seem the sort of person my post is aimed at: so averse to being constrained by self-image that you turn a blind eye when it affects you. It sounds to me like you that you are actively trying to suppress having beliefs about yourself:
People around me have a model of what Dan apparently is which is empathetic, nice, generous etc. I’m always the first to point out a bias such as racism or nonfactual emotional opinions etc. I don’t have to see myself as any of those things though.
I’ve been there, and I can think of a number of possible causes of this aversion:
Possibility #1: You see that other people are biased by their self-images in harmful ways, so you try not to have any self image that might resemble one that they would have. What you end up with is something like a “moral calculator” self-image, or a “really objective guy” self-image:
All I have to do is keep asking questions properly and at the right time and then output a response. No i’m not a calculator but the results are good according to everyone I meet and interact with.
This distinguishes you from others in a way that doesn’t activate your “don’t screw up like them” alarm bells.
Possibility #2. You are mildly disgusted by human biases and limitations, and find using the story-like heuristics of “common people” quaint but distasteful. This gives you a “too good for that silly human-think” self-image, which biases you to ignore methods of thinking that are especially useful for humans if employed correctly (i.e., as moderate-bias-high-accuracy estimators). No one is saying go think like all your wrong friends now or stop having real-time assessments of things, and the fact that you interpreted the post in that way suggests that you are somewhat sensitive to this issue. I’m saying to spend some time understanding the strengths of common emotional heuristics like narrative, not just their weaknesses, so you can make a better decision about when and how to use them.
One final comment:
I do make decisions as they come up and if I ever was to base one off the fact that “that’s what Dan would do” then that throws up a red flag to me.
It should. This should also throw up a red flag:
I don’t decide what is right in advance because if i do anything to predetermine my answer before a question arises then i’m starting off with a bias.
You are not going to escape having to cache some of your thoughts. Computers do it, AIs are going to do it, people do it, and you do it. When I learned linear algebra, I made myself re-derive every theorem and its dependencies, back to the field axioms, in my head every time I used them… but eventually I had to stop in order to follow seminar talks that’ll use 5 major results results in a span of 10 seconds. It was inevitable. And really, you don’t add 12 to itself 12 times every time you compute 12x12, even if you feel like a calculator. You don’t re-derive the distributive law from first principles every time you use a multiplication algorithm. And if you do, you’re going to be unnecessarily—dare I say irrationally—slower than otherwise ;)
The best thing to do is accept this fact, so that you can start caching instructions like keep an eye out for the following exception to this other cached instruction or watch out I don’t think I’m a calculator and assume I’m immune to biases arising from my own self-image.
Okay, I guess that makes you the first member of the Vocal Opposition.
I am not going to try to deny your subjective assessment of your own mental processes, but even in the event that you are capable of judging a situation from the ground up every moment of your life, surely you must be aware that very few other current humans share this ability.
The only reason that I can see for your opposing this idea would be to maintain your superiority by preventing access to a simpler method which would work nearly as well. I suspect that the favorable traits in your personality as your independent research has reported would disappear with a larger sampling size, as well.
Question: What is 1 + 1 Answer: “what would jesus do?”....
Not helpful is it… Wouldn’t it be better to have a cognitive model that knows how to process data rather than reaching for the cheat button?
Even if the question was “If a man was drowning etc etc” the answer “what would jesus do?” is never going to be as effective as having a data processor that can custom build an answer for the exact question...which isn’t what X would do but what is the right answer.
Yes, that would be ideal, but a current human brain is not going to work for that. Until there is practical brain augmentation or otherwise accessible advanced AI, a set of role-models would help.
But you are saying I don’t exist.
No i’m not perfect and I have biases come to my attention and fly under the radar etc- but I don’t ask myself what I would do. I don’t ask myself what someone else would do. I literally have no role models and can’t think of any I ever had. I do make decisions as they come up and if I ever was to base one off the fact that “that’s what Dan would do” then that throws up a red flag to me. It says ask the question again and find a real answer because maybe I don’t have a real reason.
What you are describing to me sounds like a short cut to a nasty bias that self perpetuates- telling you to never question anything just follow the status quo.
What the narratives would do would be to give you time to consider those situations and resolutions without actually being forced to do so on the spot. If you read it and understand that it is what you would want yourself to do in that situation, then you will have that solution on hand without your extensive on-the-spot calculations. If you think the resolution you just read in the Least Wrong was complete crap, then you would try to figure out what a better solution would be, again without the time-pressure. DO question the status quo, if you disagree with it. The point is to make them so good that you wouldn’t disagree to begin with, and would be happy to have the help that (mortals not blessed with infinite and instantaneous cognitive resources) could use in uncommon but important decisions.
I guess i don’t “want” myself to do anything. I don’t decide what is right in advance because if i do anything to predetermine my answer before a question arises then i’m starting off with a bias.
In a way what i’m saying is 1+1 could equal 2 tomorrow (which it can’t) and I will still probably get the answer right because i didn’t decide to stick with the answer 1+1=2 before the question was asked (therefore before this mysterious universe switched the answer).
I’ll sound completely biased and unbelievable when i say this but i’ll say it anyway- in my experience of breaking down my expectations for who “Dan” is and what “Dan does” i’ve made really good choices for the good of everyone around me. People around me have a model of what Dan apparently is which is empathetic, nice, generous etc. I’m always the first to point out a bias such as racism or nonfactual emotional opinions etc. I don’t have to see myself as any of those things though. All I have to do is keep asking questions properly and at the right time and then output a response. No i’m not a calculator but the results are good according to everyone I meet and interact with.
The problem with that is if you fix an answer like cement in to your brain based on one set of data- even if the data changes later you will have this cement lump in your head saying it’s “so good that you wouldn’t disagree” and so you don’t recalculate. I mean why would you calculate an answer you already know?
What you really need to do is not make accurate biases to pre-determine or influence your answer but work on removing all your layers so you make the calculations properly and unbiased. That way you won’t have to worry about if you dance or not- which ever one is right will be determined when the question comes up. Again this isn’t just a theory of what i think you should do- this is what i do so don’t tell me it isn’t possible.
+1 for sharing; you seem the sort of person my post is aimed at: so averse to being constrained by self-image that you turn a blind eye when it affects you. It sounds to me like you that you are actively trying to suppress having beliefs about yourself:
I’ve been there, and I can think of a number of possible causes of this aversion:
Possibility #1: You see that other people are biased by their self-images in harmful ways, so you try not to have any self image that might resemble one that they would have. What you end up with is something like a “moral calculator” self-image, or a “really objective guy” self-image:
This distinguishes you from others in a way that doesn’t activate your “don’t screw up like them” alarm bells.
Possibility #2. You are mildly disgusted by human biases and limitations, and find using the story-like heuristics of “common people” quaint but distasteful. This gives you a “too good for that silly human-think” self-image, which biases you to ignore methods of thinking that are especially useful for humans if employed correctly (i.e., as moderate-bias-high-accuracy estimators). No one is saying go think like all your wrong friends now or stop having real-time assessments of things, and the fact that you interpreted the post in that way suggests that you are somewhat sensitive to this issue. I’m saying to spend some time understanding the strengths of common emotional heuristics like narrative, not just their weaknesses, so you can make a better decision about when and how to use them.
One final comment:
It should. This should also throw up a red flag:
You are not going to escape having to cache some of your thoughts. Computers do it, AIs are going to do it, people do it, and you do it. When I learned linear algebra, I made myself re-derive every theorem and its dependencies, back to the field axioms, in my head every time I used them… but eventually I had to stop in order to follow seminar talks that’ll use 5 major results results in a span of 10 seconds. It was inevitable. And really, you don’t add 12 to itself 12 times every time you compute 12x12, even if you feel like a calculator. You don’t re-derive the distributive law from first principles every time you use a multiplication algorithm. And if you do, you’re going to be unnecessarily—dare I say irrationally—slower than otherwise ;)
The best thing to do is accept this fact, so that you can start caching instructions like keep an eye out for the following exception to this other cached instruction or watch out I don’t think I’m a calculator and assume I’m immune to biases arising from my own self-image.
Okay, I guess that makes you the first member of the Vocal Opposition.
I am not going to try to deny your subjective assessment of your own mental processes, but even in the event that you are capable of judging a situation from the ground up every moment of your life, surely you must be aware that very few other current humans share this ability.
The only reason that I can see for your opposing this idea would be to maintain your superiority by preventing access to a simpler method which would work nearly as well. I suspect that the favorable traits in your personality as your independent research has reported would disappear with a larger sampling size, as well.