Benquo comments on Should Rationalists Tip at Restaurants? - Less Wrong

8 Post author: Mass_Driver 12 July 2011 05:28AM

You are viewing a comment permalink. View the original post to see all comments and the full post content.

Comments (29)

You are viewing a single comment's thread.

Comment author: Benquo 12 July 2011 05:43:03AM *  4 points [-]

EDIT: Summary of my comment: noticing whether you would get caught costs a lot of mental energy that would likely be more productively used elsewhere.

On one hand:

  • You get to do something better with the money. Given that possibility, it's wasteful to spend it on something like tips instead.

  • Most people (at least in the US) tip, and therefore are motivated to come up with arguments for tipping. This means we have probably heard more arguments for than against, relative to the true merits, and should discount the case for tipping appropriately.

On the other:

  • Getting caught sends a terrible signal to possible allies that you will defect when you think a rule is unenforceable. It's hard to know when you will get caught, so the mental energy required to track likely consequences is likely to far exceed the value of the money saved from not tipping.

  • Human beings often value deontological or virtue ethics, not consequentialism. You will feel (and therefore act, i.e. give of signals) like a bad, guilty person. This effect will vary a lot depending on your disposition. It also might be a poor trade for your utility function, again varying by individual values and dispositions.

  • TDT-like considerations, a desire to gain cooperation from other TDT agents on single-shot prisoners' dilemmas.

I consider only my first argument for tipping to be a strong one, but on its own I think it is strong enough that I would be quite surprised if a rational sociopath (e.g. Quirrelmort) failed to tip.

Comment author: endoself 12 July 2011 06:05:10AM 0 points [-]

Human beings often value deontological or virtue ethics, not consequentialism. You will feel (and therefore act, i.e. give of signals) like a bad, guilty person. This effect will vary a lot depending on your disposition. It also might be a poor trade for your utility function, again varying by individual values and dispositions.

It might also motivate you to be extra careful to use the money well so you can justify your actions.

TDT-like considerations, a desire to gain cooperation from other TDT agents on single-shot prisoners' dilemmas.

Most waiters aren't TDT agents. TDT's prisoner's dilemma algorithm is to cooperate if and only if your opponent will cooperate if and only if you cooperate.

I consider only my first argument for tipping to be a strong one, but on its own I think it is strong enough that I would be quite surprised if a rational sociopath (e.g. Quirrelmort) failed to tip.

Agreed.

Comment author: Benquo 12 July 2011 12:22:55PM 1 point [-]

I don't think the TDT argument is a very strong one, but it is slightly stronger than you seem to think.

By tipping you're not only cooperating with the waiter (who provided good or bad service before the tip was revealed) but also with past and future customers (the tipping practices of past customers may influence the quality of waiter you get and the quality of service they provide, and your tipping practice affects the outcomes experienced by future customers).

Comment author: endoself 12 July 2011 06:16:16PM 0 points [-]

I don't think most customers are TDT agents either.

Comment author: Benquo 12 July 2011 09:33:13PM 0 points [-]

Agreed, that's why the argument is weak.