It is impossible for one to act on another’s utility function
Depends on what you mean by this. If a paperclip maximizer pays me $100 to produce a paperclip, and I accept the deal because I want the money, do I “act on paperclip maximizer’s utility function” here?
.
These are some specific mistakes in the text. But there is also a meta-mistake behind them, which is more important, because you are more likely to repeat it in the future. Essentially, you are playing with words (such as “altruism”, “selfish”, “selfless”, “virtue”); you are looking at the map first. Even where you provide a specific example, it seems like you made the conclusion first, and looked for possible examples in the territory later.
How would an opposite approach look like? Perhaps by asking a question “what are examples of things that people call ‘doing good to others’?” and trying to collect many different examples. (Using the word “altruism” already means starting at the wrong place. Is “altruism” the same or not the same as “doing good to others”? That is what needs to be explored. Even “good” is a label, but it is much simpler one than “altruism”.) The more different examples, the better, because with similar examples you are just repeating yourself.
Is it possible to do good without wanting to do good? Accidentally, by a habit, by misjudging the situation. Is it possible to hurt people while trying to do good? What if different people will evaluate “good” differently; perhaps because of conflict of interest (e.g. you helped someone win in a zero-sum game, the loser hates you), or because long-term and short-term consequences point in opposite direction? … This all is just the few things that came to my mind during one minute of writing this comment; there are probably things I didn’t even touch here.
Then, after having various specific examples, we can start examining whether they have something in common. And it is quite likely to conclude things like “X usually Y, but not always (see this weird example of X745)”.
No.
Depends on what you mean by this. If a paperclip maximizer pays me $100 to produce a paperclip, and I accept the deal because I want the money, do I “act on paperclip maximizer’s utility function” here?
.
These are some specific mistakes in the text. But there is also a meta-mistake behind them, which is more important, because you are more likely to repeat it in the future. Essentially, you are playing with words (such as “altruism”, “selfish”, “selfless”, “virtue”); you are looking at the map first. Even where you provide a specific example, it seems like you made the conclusion first, and looked for possible examples in the territory later.
How would an opposite approach look like? Perhaps by asking a question “what are examples of things that people call ‘doing good to others’?” and trying to collect many different examples. (Using the word “altruism” already means starting at the wrong place. Is “altruism” the same or not the same as “doing good to others”? That is what needs to be explored. Even “good” is a label, but it is much simpler one than “altruism”.) The more different examples, the better, because with similar examples you are just repeating yourself.
Is it possible to do good without wanting to do good? Accidentally, by a habit, by misjudging the situation. Is it possible to hurt people while trying to do good? What if different people will evaluate “good” differently; perhaps because of conflict of interest (e.g. you helped someone win in a zero-sum game, the loser hates you), or because long-term and short-term consequences point in opposite direction? … This all is just the few things that came to my mind during one minute of writing this comment; there are probably things I didn’t even touch here.
Then, after having various specific examples, we can start examining whether they have something in common. And it is quite likely to conclude things like “X usually Y, but not always (see this weird example of X745)”.