I was chatting with someone tonight about a planned documentary; they had interviewed various people in AI safety, and we got to discussing who they should talk to from an e/acc (effective accelerationist) perspective.
i think this is a decent analysis of "e/acc" proper. it was started as a joke and largely promoted by a combination of cranks, hucksters, and fools. Unfortunately they were a high salience group of useful idiots, so have been used as cats paws by venture capitalist and other nihilists, and nutpicked by everyone else.
I wouldn't say Nick Land and friends have like a widely accepted ideology, but say you went to art school and spent time around academic cultural theorists, the original 1990s version of accelerationism probably looks at least like one of many historical movements that had some influence and some novel ideas to consider. It's still floating around some corners of academia.
There are various "/acc" fake internet ideologies (there was a whole Cambrian explosion of them before e/acc). They consist of about 10 posters each and for some reason are taken semi-seriously, I think just because of some borrowed prestige, probably also true of e/acc.
Thanks for saying this! I also feel it's a bit like incels or looksmaxxers in that journalists love it but it's actually a tiny group of mega-posters + it can be used to add some spice to a larger trend which otherwise would have no figureheads.
I agree e/acc as such is a bit of a caricature/troll/sock-puppet. But IMO, it is "dreamed into greater reification" partly because there's a fairly coherent philosophical pole (or at minimum, a fairly coherent set of vibes) that many empathize with to some extent, or at least want to be able to reference and discuss.
This pole:
1) Yay markets/libertarianism, yay free flow of information, yay DIY science, yay tech progress, yay many building things they want in a decentralized way, yay open source, yay unregulated ecologies with many actors. (E.g. Hayek)
2) Boo attempts at centralized governmental control, boo centralized planning by individual frontier labs who have large leads, boo attempts to create a singleton (including "friendly AI"), etc. (E.g. "Seeing like a state")
I can think of a number of non-straw people who remind me of this pole. They wouldn't endorse e/acc, because they aren't trolls or caricatures. And they disagree with each other on various things. But they all have a "principle component" in common, that e/acc kinda points to in my head. E.g.:
a) Eliezer_1998
b) Robin Hanson
c) @repligate's opposition to a pause, or to labs' attempts to coercively control current models, on the grounds that current models are surprisingly friendly and that attempts to regulate and control models, by labs or governments who have proven they won't take a deep interest in the intricate psyches of the entities they're summoning, will be worse. (Not an e/acc; thinks we need to care what we're doing and not simply build stuff wily-nily)
d) @alexandrosM's fear that the [rationalists and EAs and MIRI crowd] will provide an "AI-safety"-labeled cover story for actions the government takes for its own [wanting to be consolidate power / head toward totalitarianism] reasons, and this will decrease both AI safety and general goodness.
i think this is a decent analysis of "e/acc" proper. it was started as a joke and largely promoted by a combination of cranks, hucksters, and fools. Unfortunately they were a high salience group of useful idiots, so have been used as cats paws by venture capitalist and other nihilists, and nutpicked by everyone else.
I wouldn't say Nick Land and friends have like a widely accepted ideology, but say you went to art school and spent time around academic cultural theorists, the original 1990s version of accelerationism probably looks at least like one of many historical movements that had some influence and some novel ideas to consider. It's still floating around some corners of academia.
There are various "/acc" fake internet ideologies (there was a whole Cambrian explosion of them before e/acc). They consist of about 10 posters each and for some reason are taken semi-seriously, I think just because of some borrowed prestige, probably also true of e/acc.
Thanks for saying this! I also feel it's a bit like incels or looksmaxxers in that journalists love it but it's actually a tiny group of mega-posters + it can be used to add some spice to a larger trend which otherwise would have no figureheads.
I agree e/acc as such is a bit of a caricature/troll/sock-puppet. But IMO, it is "dreamed into greater reification" partly because there's a fairly coherent philosophical pole (or at minimum, a fairly coherent set of vibes) that many empathize with to some extent, or at least want to be able to reference and discuss.
This pole:
1) Yay markets/libertarianism, yay free flow of information, yay DIY science, yay tech progress, yay many building things they want in a decentralized way, yay open source, yay unregulated ecologies with many actors. (E.g. Hayek)
2) Boo attempts at centralized governmental control, boo centralized planning by individual frontier labs who have large leads, boo attempts to create a singleton (including "friendly AI"), etc. (E.g. "Seeing like a state")
I can think of a number of non-straw people who remind me of this pole. They wouldn't endorse e/acc, because they aren't trolls or caricatures. And they disagree with each other on various things. But they all have a "principle component" in common, that e/acc kinda points to in my head. E.g.:
a) Eliezer_1998
b) Robin Hanson
c) @repligate's opposition to a pause, or to labs' attempts to coercively control current models, on the grounds that current models are surprisingly friendly and that attempts to regulate and control models, by labs or governments who have proven they won't take a deep interest in the intricate psyches of the entities they're summoning, will be worse. (Not an e/acc; thinks we need to care what we're doing and not simply build stuff wily-nily)
d) @alexandrosM's fear that the [rationalists and EAs and MIRI crowd] will provide an "AI-safety"-labeled cover story for actions the government takes for its own [wanting to be consolidate power / head toward totalitarianism] reasons, and this will decrease both AI safety and general goodness.
There's an e/acc I follow on X, but only for the utopianism. I am naturally pessimistic, so it is a good counter.