No, I haven't studied Rawls before, but it sounds like he's invoking (or requiring) a "Right to Anonymity" and we've been grappling with that concept since the mid 60's. Which is, uh, about five years before his proposal. You've heard of Anonymous, I assume?
-
-
Replying to @JediJeremy @consultify and
Oh, OK. It's a "thought experiment". Some principles ARE definitely encoded into modern practice (eg: "de-identification" in EU Data Retention laws) but his basic premise is impossible in our surveillance-heavy world. Govt. knows who you are, and we all know our place in society.
1 reply 0 retweets 2 likes -
I know enough about database queries to know that they are always highly selective. The veil of ignorance is just about making sure that things like ethnicity, gender and faith are not part of the selection criteria determining who gets what.
1 reply 0 retweets 3 likes -
Then you don't know enough about data mining, because those dimensions can always be (statistically) reconstituted through other metrics (where you live, income, interests) if you're unscrupulous. OTOH I'm all for "anti-discrimination" laws that depend on transparency.
2 replies 0 retweets 6 likes -
Replying to @JediJeremy @consultify and
I understand the principle - Justice should be blind - and I agree. But justice _isn't_ blind (well, perhaps selectively) she never was, and pretending it's so has caused endless problems in reality. Especially trying to fix those problems, because it gave them cover.
1 reply 1 retweet 4 likes -
Yes Justice is complicated and difficult, hence your facile demand for a definition of fairness. But the mere fact that we can't seem to get it right should prompt us to try harder, not stop us from trying.
1 reply 0 retweets 4 likes -
That was exactly my point. We have all these amazing new tools, we have an opportunity to do better. Humans and machines, correcting each other biases in a messy world. And I was _not_ being facile about the definition of fairness. If you can't define it, you can't create it.
1 reply 0 retweets 7 likes -
Replying to @JediJeremy @consultify and
It doesn't have to be a simple definition, it can be a thousand pages based on the works of Hume and Kant and Rawls. But the first step to improving anything is being able to measure it, otherwise you can't determine if your "improvements" are working.
1 reply 0 retweets 8 likes -
Replying to @JediJeremy @consultify and
This is what I mean when I talk about the eye-rolling disconnect between the disciplines. When we ask for a definition of fairness, you think we're joking. We don't get any useful answers, and we're left to make it up as we go along. Then you blame us.
4 replies 2 retweets 16 likes -
Replying to @JediJeremy @consultify and
Well. Flipping. Put. This is the spirit in which I devised the notion of Friendly AI, and the parts of academia departing from it to wring their hands about the unsolvability of trolley problems will never produce anything useful.
1 reply 1 retweet 4 likes
You might be entertained by http://arbital.com/p/cev . It's what I came up with to answer all the unanswerable questions in AGI ethics back in 2004, so we could move on to harder problems like how to get a superintelligence to do literally anything without destroying the universe.
-
-
Replying to @ESYudkowsky @JediJeremy and
It's hard for many people to get interested in the control problem, including engineers, yes? In general, no one wants to confront the possibility that their children might try to kill them.
0 replies 0 retweets 0 likesThanks. Twitter will use this to make your timeline better. UndoUndo
-
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.