[samizdatum]

> See, in this family, we DON’T just want to maximize the overall quantity of individual happiness (like Bentham). No… In this family, we want to maximize the overall quantity AND the quality of individual happiness (like John Stuart Mill).

But daaad, this now becomes a multi-objective optimization problem, and there are potentially an infinite number of solutions that lie along the Pareto frontier...

[Judgmentality]

I don't know. The solution might be as simple as "make one person very very happy."

https://www.smbc-comics.com/comic/2012-04-03

[samizdatum]

Yes, a utility monster is conceivable under both theories. It's amusing how a utility monster represents a serious objection for utilitarian philosophers, but is a total non-concern for people that solve optimization problems for a living. It would be like software engineers lying awake in bed worrying about what they'd do if they just stopped writing bugs one day.

[adrianN]

Avoiding weird solutions by adding appropriate constraints is extremely important to people who solve optimization problems in practice. The classic example from the inventor of linear programming is the diet problem [1], where the naive LP suggested to eat nothing but bouillon cubes or drink 500 gallons of vinegar.

[1]https://resources.mpi-inf.mpg.de/departments/d1/teaching/ws1...

[Piskvorrr]

It's even a matter of being aware of assumed, implicit constraints. "In the early 1950s [...] the nutritional requirements didn't show a limit on the amount of salt? "Isn't too much salt dangerous?" He replied that it wasn't necessary; most people had enough sense not to consume too much."

[repsilat]

There are certainly cases of weird degenerate solutions to optimization problems. Lots of examples in machine learning (one classic at https://openai.com/blog/faulty-reward-functions/).

In more old-school convex optimization the closest thing is probably insufficiently constrained problems. If you don't say the amount of each food you eat has to be greater than zero, you can get all your dietary needs satisfied for the low low price of minus infinity dollars!

From another perspective, though, perhaps there's a "No True Scotsman" side of this. Is a utility monster a sign of a badly-specified problem, or are they a definitive sign of one? If the former, it stands to reason it's not a "concern" for modellers -- it's a dream!

[AstralStorm]

Utility monster is a case of lack of time constraint on optimization problem and lack of robustness.

For example, the special "Felix" case ignores the case where said guy is stuck by gamma radiation and dies. Over time, the probability of that k or a bunch of other catastrophes ruinning the solution tends to 1.

Therefore, the best solution avoids the most known catastrophes and is updated as new possibilities of those are found. (Tontine lotto, anyone?) Minimax loss optimal. Maximin (maximizing gain without increasing base loss) could be decent as well. Deciding between the two is better left to wizards.

Online stochastic optimization is mathematical black magic anyway so far.

But then, satisfying humans is much easier given all the built in biases we have. Keeping things alive long term is much harder.

[benj111]

Is there not diminishing returns though?

Does the millionth lucky penny feel as special as the first? If you gave Jeff Bezos a million $ would he even notice, let alone feel happier.

I feel that most happiness is gained from being able to feed and house your family, not having your children die from disease. Anything much beyond that is a rounding error and based on standards of the day.

[wazoox]

Notice that in neoclassical economics, value grows because satisfaction of the agents grows and nothing else. Therefore growth can be infinite as long as someone could grow infinitely happy with whatever they get out of economic activities. This implies some form of utilitarianism. This also strikes as complete, unreal fantasy. Hmmm.

[mattigames]

Well that's just a bug in the machine for measuring happiness using the average instead of the median.

[DuskStar]

Unfortunately for Felix, all of his slaves were really bringing down the median happiness. But they're all dead now, and he has robots instead!

Having robots makes Felix really happy. Not as happy as having slaves, but still happier than anyone else could ever be.

[jacobolus]

https://www.utilitarianism.com/nu/omelas.pdf

[samizdatum]

Maximin is probably better for avoiding those pesky "torture/enslave half of humanity" scenarios.

[ucarion]

Not quite -- if a single poor chap starts off with Very Bad torture, maximin will perscribe that everyone else have Medium Bad torture as long as the single worst-off person does a little better.

[waterhouse]

Hmm, really? Maximin = maximizing the minimum, right? Seems like it could also lead to "Find the people with the most difficult-to-cure genetic conditions and/or abusive upbringings that make them extremely unhappy, and enslave the rest of humanity to try to find cures for them." Unless pumping such people with dopamine or whatever is considered a solution. (Or killing them.)

[DuskStar]

If your algorithm accounts for future states - and it'll have to, to be practical - "assigning minimal resources to the really unhappy people to let them die from neglect" is also part of the set of options.

[waterhouse]

I have met at least one vegetarian (who is also a utilitarian) who has said that many species of animals currently suffer so much that it would be better if those species were all dead. I've heard of others who share this opinion. I call it the "Final Solution" to animal welfare; I think the name is quite appropriate.

Anyway, well, then, that seems to leave some big question marks. How is everyone who goes through utilitarian moral reasoning supposed to decide between help and neglect, and hope to come up with the same answer? (Because if two people come up with sufficiently different answers, it may lead to one or both concluding that the expected utility of violently enforcing obedience from the other person is worth the cost, unless they have deontological rules of some sort. Like the kind that the article says Mill supported.)

[emilfihlman]

I believe median*population is a good value to optimise for since then killing everyone but one because counterproductive.

[saagarjha]

Sounds like a sort of inverse Omelas.

[Scarblac]

I love SMBC. He does this kind of comic so well, where it starts out with some interesting premise and then go to complete hilarious absurdity with some "logical" steps. I can't think of anyone else who has the same type of humor.

[Judgmentality]

I've really enjoyed watching him develop over the years. His earlier comics, while good, are just so mediocre compared to how good he is today. While he's obviously always been funny, I guess 10+ years of grinding have really helped him hone his craft. He's probably my favorite comic - and his books are excellent as well.

[ur-whale]

> make one person very very happy

Didn't you just describe North Korea?

[IggleSniggle]

That dude does NOT seem happy, fwiw

[chrstphrhrt]

Are you saying the difference between quantity and quality is zero sum, or maybe more like dividing by zero?

I feel like Pareto efficiency only comes into the picture if the quantities are actually clear, but I don't understand how to make that concrete in ethical (not economic) terms.

The "rule" of "quality" seems like it can easily be abused by whoever is in power in the given situation. E.g. an assertion such as this could be valid: it's okay for police to shoot "innocent" people if "they were disturbing the peace".

Are quality and quantity both vulnerable to being gamed?

[perfmode]

pareto means optimizing one dimension without sacrificing the other.

the frontier is the points in space where neither can be further optimized without impacting the other.

[chrstphrhrt]

Yeah flashbacks from econ 101.. Was referring to the coherence issue of moving away from known quantities/dimensions.