What justifies this unconventional definition of the word? Random House Dictionary offers three senses of “win”:
to finish first in a race, contest, or the like.
to succeed by striving or effort: He applied for a scholarship and won.
to gain the victory; overcome an adversary: The home team won.
Notice that 2 of the 3 involve a contest against another; definition 2 is closer to what’s wanted, but the connection between winning a competition is so strong, that when offering an example, the dictionary editors chose a competitive example.
This unconventional usage encourages equivocation, and it appeals to the hyper-competitive, while repelling those who shun excessive competition. Why LW’s attachment to this usage? It says little for the virtue of precision; it makes LWers seem like shallow go-getters who want to get ahead at any cost.
Nothing justifies it, really. Like most local jargon in any community, it evolves contingent on events.
In this case, the history is that the phrase “rationalists should win” became popular some years ago as a way of counteracting the idea that what being rational meant was constantly overanalyzing all problems and doing things I know are stupid because I can come up with a superficially logical argument suggesting I should do those things. Newcomb’s Problem came up a lot in that context as well. The general subtext was that if my notion of “rationality” is getting in the way of my actually achieving what I want, then I should discard that notion and instead actually concentrate on achieving what I want.
Leaving all that aside: to my mind, any definition of “win” that makes a win-win scenario a contradiction in terms is a poor definition, and I decline to use it.
Leaving that aside: do we seem to you like shallow go-getters who want to get ahead at any cost, or are we talking about how we seem to some hypothetical other people?
If the word “win” is getting in people’s way because it makes them think of zero-sum scenarios, social status conflicts, or Charlie Sheen, then we can and should find other ways of explaining the same concept.
Rationality should make you better off than you would be if you didn’t have it. That doesn’t imply defeating or outdoing other people. It just means that the way we evaluate candidate decision procedures is whether they create beneficial outcomes — not bogus standards such as how unemotional they are, how many insignificant bits of data they take in, how much they resemble Traditional Scholarship, how much CPU time they consume, or whether they respect naïve ideas of causality.
Yup. Personally, I prefer “optimizers should optimally achieve their goals” to “rationalists should win”. But there exist people who find the latter version pithier, easier to remember, and more compelling. Different phrasings work better for different people, so it helps to be able to say the same thing in multiple ways, and it helps to know your audience.
I have elsewhere argued essentially this, though “optimized” and “optimal” are importantly different. The conclusion I came to reading the comments there was that people vary in how they process the relevant connotations, and arguing lexical choice makes my teeth ache, but personally I nevertheless try to avoid talking about “rational X choosing” when I mean optimizing my choice of X for particular goals.
This is a little speculative and I wasn’t around for the original coining of the LW usage, but American geek slang has contained a much broader sense of “win” for quite a while. Someone conversant in that sociolect might say a project wins if it accomplishes its goals (related to but broader than sense 2 in the parent); they also might say an event is a win if it has some unexpected positive consequence, that a person wins if they display luck or excellence in some (not necessarily competitive) domain, et cetera. I’d expect that to have come from 1990s gamer jargon if you trace it back, but it’s metastasized pretty far from those origins by now.
I’m guessing that “rationalists should win” derives from that usage.
What justifies this unconventional definition of the word? Random House Dictionary offers three senses of “win”:
I apologize, I thought it was appropriate to base my discussion in community terminology. I guess I assumed Eliezer’s definition of “winning” and rationality were a community norm. My bad.
What justifies this unconventional definition of the word? Random House Dictionary offers three senses of “win”:
to finish first in a race, contest, or the like.
to succeed by striving or effort: He applied for a scholarship and won.
to gain the victory; overcome an adversary: The home team won.
Notice that 2 of the 3 involve a contest against another; definition 2 is closer to what’s wanted, but the connection between winning a competition is so strong, that when offering an example, the dictionary editors chose a competitive example.
This unconventional usage encourages equivocation, and it appeals to the hyper-competitive, while repelling those who shun excessive competition. Why LW’s attachment to this usage? It says little for the virtue of precision; it makes LWers seem like shallow go-getters who want to get ahead at any cost.
Nothing justifies it, really. Like most local jargon in any community, it evolves contingent on events.
In this case, the history is that the phrase “rationalists should win” became popular some years ago as a way of counteracting the idea that what being rational meant was constantly overanalyzing all problems and doing things I know are stupid because I can come up with a superficially logical argument suggesting I should do those things. Newcomb’s Problem came up a lot in that context as well. The general subtext was that if my notion of “rationality” is getting in the way of my actually achieving what I want, then I should discard that notion and instead actually concentrate on achieving what I want.
Leaving all that aside: to my mind, any definition of “win” that makes a win-win scenario a contradiction in terms is a poor definition, and I decline to use it.
Leaving that aside: do we seem to you like shallow go-getters who want to get ahead at any cost, or are we talking about how we seem to some hypothetical other people?
If the word “win” is getting in people’s way because it makes them think of zero-sum scenarios, social status conflicts, or Charlie Sheen, then we can and should find other ways of explaining the same concept.
Rationality should make you better off than you would be if you didn’t have it. That doesn’t imply defeating or outdoing other people. It just means that the way we evaluate candidate decision procedures is whether they create beneficial outcomes — not bogus standards such as how unemotional they are, how many insignificant bits of data they take in, how much they resemble Traditional Scholarship, how much CPU time they consume, or whether they respect naïve ideas of causality.
Yup. Personally, I prefer “optimizers should optimally achieve their goals” to “rationalists should win”. But there exist people who find the latter version pithier, easier to remember, and more compelling. Different phrasings work better for different people, so it helps to be able to say the same thing in multiple ways, and it helps to know your audience.
“Lesswrong, a community blog dedicated to the art of optimal goal-seeking, mostly through optimal belief acquisition.”
I kinda like that. It’s wordier, but it loses some of the connotations of “rational” that invariably trip up newcomers.
I have elsewhere argued essentially this, though “optimized” and “optimal” are importantly different. The conclusion I came to reading the comments there was that people vary in how they process the relevant connotations, and arguing lexical choice makes my teeth ache, but personally I nevertheless try to avoid talking about “rational X choosing” when I mean optimizing my choice of X for particular goals.
This is a little speculative and I wasn’t around for the original coining of the LW usage, but American geek slang has contained a much broader sense of “win” for quite a while. Someone conversant in that sociolect might say a project wins if it accomplishes its goals (related to but broader than sense 2 in the parent); they also might say an event is a win if it has some unexpected positive consequence, that a person wins if they display luck or excellence in some (not necessarily competitive) domain, et cetera. I’d expect that to have come from 1990s gamer jargon if you trace it back, but it’s metastasized pretty far from those origins by now.
I’m guessing that “rationalists should win” derives from that usage.
Indeed, one might even say that Less Wrong is for the win.
I apologize, I thought it was appropriate to base my discussion in community terminology. I guess I assumed Eliezer’s definition of “winning” and rationality were a community norm. My bad.