As an add-on, I found this LW article today that captures the essence of “first pleasure”
clippy interjects from the balcony: like number of paperclips created!
I’m not quite sure I got this part, could you please elaborate on it?
why seek pleasure at all? Wouldn’t it be better to measure success directly based on meeting long-term goals
Here, I would argue that the feeling of the second pleasure is essential to meeting long-term goals. Feeling good about accomplishing sub-tasks will keep someone working towards an important goal, especially if it requires a long period of sustained effort.
Thus, meeting short-term goal → success → second pleasure → working on next short-term goal; with enough iterations, that will lead to the meeting of long-term goals and thus success.
https://wiki.lesswrong.com/wiki/Paperclip_maximizer is the canonical example of over-simplified goal optimization. I bring it up mostly as a reminder that getting your motivational model wrong can lead to undesirable actions and results.
Which leads to my main point. You’re recommending one type of pleasure over another, based on it being more aligned with your non-pleasure-measured goals. I’m wondering why you are arguing for this, as opposed to just pursuing the goals directly, without consideration of pleasure.
Ah, now I’ve got what you mean. Thanks for referring me to that thought experiment, I don’t have much prior knowledge on the field of AI so that was definitely a new insight for me.
I see now that my original shortform did not explicitly state that my terminal value was indeed the fulfillment of important goals. I was reflecting more on the distinction between pleasurable feelings that led to distraction & bad habits, vs ones that led to the actual fulfillment of goals. It was a personal reminder to experience the latter in place of the former, as much as I can.
Now, I hold the view that pleasure can be a useful tool in the pursuit of my goals. It is a means to that end. An important caveat that your response reminded me of, though, is that sometimes pursuing goals might not be immediately pleasurable, so it might be wise not to naïvely expect pleasure from every part of that process.
As an add-on, I found this LW article today that captures the essence of “first pleasure”
I’m not quite sure I got this part, could you please elaborate on it?
Here, I would argue that the feeling of the second pleasure is essential to meeting long-term goals. Feeling good about accomplishing sub-tasks will keep someone working towards an important goal, especially if it requires a long period of sustained effort.
Thus, meeting short-term goal → success → second pleasure → working on next short-term goal; with enough iterations, that will lead to the meeting of long-term goals and thus success.
https://wiki.lesswrong.com/wiki/Paperclip_maximizer is the canonical example of over-simplified goal optimization. I bring it up mostly as a reminder that getting your motivational model wrong can lead to undesirable actions and results.
Which leads to my main point. You’re recommending one type of pleasure over another, based on it being more aligned with your non-pleasure-measured goals. I’m wondering why you are arguing for this, as opposed to just pursuing the goals directly, without consideration of pleasure.
Ah, now I’ve got what you mean. Thanks for referring me to that thought experiment, I don’t have much prior knowledge on the field of AI so that was definitely a new insight for me.
I see now that my original shortform did not explicitly state that my terminal value was indeed the fulfillment of important goals. I was reflecting more on the distinction between pleasurable feelings that led to distraction & bad habits, vs ones that led to the actual fulfillment of goals. It was a personal reminder to experience the latter in place of the former, as much as I can.
Now, I hold the view that pleasure can be a useful tool in the pursuit of my goals. It is a means to that end. An important caveat that your response reminded me of, though, is that sometimes pursuing goals might not be immediately pleasurable, so it might be wise not to naïvely expect pleasure from every part of that process.