I don’t necessarily endorse MichaelAnissimov’s take on this particular issue, but Tumblr’s a big site. You can find everything from cat picture blogs to literal Nazis on it if you look. That doesn’t mean it’s disproportionately cats (plausible) or Nazis (very improbable), though, nor that people are talking about cat pictures or antisemitism when they complain about Tumblr culture.
More specifically, there are basically two things you can easily use Tumblr for: image sharing and text microblogging. The former lends itself well to porn, the latter to radical politics, and the communities built up around these use cases don’t overlap all that much.
I’m at work and probably shouldn’t click that link, but once again, I’m not saying that MichaelAnissimov’s analysis is right; that’s a rather complicated question that touches on some deep and AFAIK unresolved issues in identity formation, though priors point to “no”. I am saying that Tumblr can accurately be called a major vector for radical identity politics without contradiction with its other use cases. It’s both full of straight (and non-straight) erotica and full of LGBT and feminist activists, and the latter are the more relevant when we’re discussing its political influence.
Tumblr can accurately be called a major vector for radical identity politics
Tumblr is just a publishing platform. The rise of the popularity of a platform does not imply that those who migrate to it en masse grew more numerous, only that they prefer this particular form of expression. The rise and fall of MySpace did not imply the rise and fall of indie bands.
The rise of the popularity of a platform does not imply that those who migrate to it en masse grew more numerous, only that they prefer this particular form of expression.
Groupthink is a thing. Emergent communities are a thing. Tumblr isn’t just a publishing platform, it’s a social platform, and that makes it meaningful to talk about “Tumblr” as a social entity rather than a purely technical tool. And when looked at in that light, Tumblr (the community, or more specifically the radical politics community as opposed to e.g. the catpic community) almost certainly radicalized many people who wouldn’t otherwise have gotten involved; naturally there would have been other centers of radical politics and other vectors of radicalization if David Karp had been hit by a bus back in 2004, but that’s not the world we live in. Similarly, MySpace played a role in the rise of indie music by exposing people to it through social-media channels.
We’re typing this on Less Wrong, dude. I can hardly think of a clearer example of online movement-building. Do you really think that this community—movement, cult, hugbox, whatever you want to call it—would look much like it does now if Eliezer had stayed on the SL4 list, or gone straight to publishing a book?
So do you think there is a causal connection between the number of unique pageviews on LW and the state of AI research (size of programs, funding, etc) out in the real world?
AI research in general, only marginally. That’s a big field with a lot of other things driving it. AI safety research, or e.g. awareness of the heuristics-and-biases program among LW’s target demographic? Yes, absolutely.
I don’t necessarily endorse MichaelAnissimov’s take on this particular issue, but Tumblr’s a big site. You can find everything from cat picture blogs to literal Nazis on it if you look. That doesn’t mean it’s disproportionately cats (plausible) or Nazis (very improbable), though, nor that people are talking about cat pictures or antisemitism when they complain about Tumblr culture.
More specifically, there are basically two things you can easily use Tumblr for: image sharing and text microblogging. The former lends itself well to porn, the latter to radical politics, and the communities built up around these use cases don’t overlap all that much.
Precisely. So correlating popularity of Tumblr to gender dysphoria leads one straight to here.
I’m at work and probably shouldn’t click that link, but once again, I’m not saying that MichaelAnissimov’s analysis is right; that’s a rather complicated question that touches on some deep and AFAIK unresolved issues in identity formation, though priors point to “no”. I am saying that Tumblr can accurately be called a major vector for radical identity politics without contradiction with its other use cases. It’s both full of straight (and non-straight) erotica and full of LGBT and feminist activists, and the latter are the more relevant when we’re discussing its political influence.
The link is SFW :-)
Tumblr is just a publishing platform. The rise of the popularity of a platform does not imply that those who migrate to it en masse grew more numerous, only that they prefer this particular form of expression. The rise and fall of MySpace did not imply the rise and fall of indie bands.
Groupthink is a thing. Emergent communities are a thing. Tumblr isn’t just a publishing platform, it’s a social platform, and that makes it meaningful to talk about “Tumblr” as a social entity rather than a purely technical tool. And when looked at in that light, Tumblr (the community, or more specifically the radical politics community as opposed to e.g. the catpic community) almost certainly radicalized many people who wouldn’t otherwise have gotten involved; naturally there would have been other centers of radical politics and other vectors of radicalization if David Karp had been hit by a bus back in 2004, but that’s not the world we live in. Similarly, MySpace played a role in the rise of indie music by exposing people to it through social-media channels.
We’re typing this on Less Wrong, dude. I can hardly think of a clearer example of online movement-building. Do you really think that this community—movement, cult, hugbox, whatever you want to call it—would look much like it does now if Eliezer had stayed on the SL4 list, or gone straight to publishing a book?
So do you think there is a causal connection between the number of unique pageviews on LW and the state of AI research (size of programs, funding, etc) out in the real world?
AI research in general, only marginally. That’s a big field with a lot of other things driving it. AI safety research, or e.g. awareness of the heuristics-and-biases program among LW’s target demographic? Yes, absolutely.