> Andreessen Horowitz, who believe—this is not a strawman—that because technology is good all possible technology is good and it is wrong to regulate it.
That is a strawman. I read the Techno-Optimist Manifesto and it doesn't say that all possible technology is good nor that regulation is always wrong. For what it's worth, I had ChatGPT and Claude read the manifesto and asked whether that's a strawman, and ChatGPT said it's a strawman, and Claude said it's "a significant oversimplification and mischaracterization of the essay's content, leaning towards being a strawman argument."
(I don't mean that LLM judgment is a clinching argument, more just noting that I did the basic diligence of checking whether my objection is easily shown to be crazy before I clog up a comment section with it.)
Somewhat! The manifesto is 5000 words of vague vibes and platitudes, so that's still more specificity than is there. Your phrasing is a plausible extrapolation, but not directly present in the text.
I'm trying to restrain my pedantry, but if you're going to write "this is not a strawman" it seems especially important to hew closely and not-uncharitably to what OP wrote.
So personally I might go with:
"Andreessen Horowitz, who believe—judging from their Techno-Optimist Manifesto—that because technology is good nearly all possible technology is good and it is anti-science to regulate it."
(though I'd double-check whether "anti-science" is the best term, since it doesn't appear in the manifesto)
But actually I'd just try to find a direct quote from the manifesto that states what I wanted.
I will not be writing a letter in support of this bill. The direct effects of SB 1047 are unlikely to cause any harm, so I have also not written a letter in opposition to the bill, but if it does, as this piece suggests, serve as a necessary first step towards the kind of regulation regime that has killed thousands by slowing the distribution of vaccines, I will be regretting my decision to not try help nip it in the bud. There should be no regulation of the specific content which publicly available generative AI is capable of producing. Pornographic deepfakes are morally disgusting, but preventing them is not worth giving up (through the banning of open source image models, for example) the ability for generative AI to serve the interests of individual humans without corporations and governments interfering.
It's great to see people take the time to reach out to social justice activists/left-wingers and address their concerns about longtermism and x-risk activism, especially considering in other posts you were more actively dismissive of SJ/left criticisms of EA (eg in your response to Doing EA Better). However, I don't think this is necessarily the most effective place to do such an attempt, the amount of people receptive of the claims of 'AI ethics' against 'AI safety' reading a niche EA blog are likely to be ~nil.
Relatedly: I do have to wonder if the continued association of AI governance with EA is going to be a liability more than an asset starting on from now, and ought to at least be strategically reduced. The natural allies of AI governance and existential risk prevention in general are, as the endorsements for SB 1047 demonstrate, anti-war, environmentalist, animal rights, women's rights, workers' rights movements (i.e. the political left). But EA is possibly (for a certain amount of reasons) one of the most universally hated contemporary social movements by social justice activists. And a movement seeking to address catastrophic risks of irresponsible use of high technology while being backed by tech millionaires is likely to be taken as seriously as a climate change activist NGO backed by Total, or a public health charity backed by Marlboro would be.
Also, the most common line from the left and 'AI ethics' against EA longtermism is that it's an excuse to not do anything for the poor in the present in the name of saving expected 10^^10 lives in the future. Work done by longtermists is actually of direct interest to the poor in areas like biosecurity, but the argument would be much easier to address if not for treating EA as a single actor that believe that every idea (up to including destroying the entire universe in case electrons can suffer) should be considered seriously, but also that you need a peer-reviewed meta-review of RCTs approved by the scientific consensus of several disciplines before saying crimes against humanity are bad. Both of those epistemological stances could stand on their own separately, but taken together they form a noxious double standard.
(A similar case goes for effective animal activism. It's much harder to address the common argument that antispeciesism entail devaluing humans' unique ability to reason and participate in civic life when your closest allies work from the position that one can work to alleviate Third-Worlders's welfare like one would do for farm animals but one ought not to worry about republican liberty as non-domination/Sen-Nussbaum capabilities. Answering that you actually root your concern for animals in the republican arguments of Kymlicka and Donaldson's Zoopolis is likely to be counterproductive for obvious reasons.)
TBH, I would go the opposite direction. AI governance and animal advocacy both have as much support from the political left as they're going to get, time to reach out to people I (Ozy) personally despise.
"AI governance and animal advocacy both have as much support from the political left as they're going to get" That seems very obviously not true to me? Those issues have more mainstream resonance now than 10, 20, or 50 years ago, and it is very strange to assume that we just hit the ceiling when we're really just barely entering the mainstream, and there is plenty of low-hanging fruit about how we communicate.
I can assume from your earlier more actively dismissive attitudes toward SJ/left views on EA that you have a very negative opinion of their epistemics and think improvement is untractable for that reason. But the 'AI ethics' people are vocal in the court of public opinion but not much of a direct hindrance behind the scenes, so the indirect hindrance the former fact pose does not seem like a stable equilibria.
Done. I'm unclear if doing this as a non-American is at all useful — I'm sure not going to be voting for California governors — but it can hardly hurt. That said, how the ass do you people write a letter in only ten minutes?!!
It seems reasonable to guess that Gavin Newsom has fairly cosmopolitan values, and will care some about what people outside of the US/California think, even though they aren't his constituents as governor. As someone who grew up in California -- Californians like to think of themselves as cosmopolitan, and I think the stereotype is fairly accurate.
Writing a letter doesn't need to be high-effort. I like to think of it as sort of like writing an internet comment, except rather more polite and polished (depending on how polite and polished your typical internet comment is -- probably reasonably good if you're reading this blog). Most likely it will simply be read by a staffer who is tallying up the number of letters for and against the bill. I figure the best letter is one which stands out a little, by making an simple/interesting/memorable point.
> Andreessen Horowitz, who believe—this is not a strawman—that because technology is good all possible technology is good and it is wrong to regulate it.
That is a strawman. I read the Techno-Optimist Manifesto and it doesn't say that all possible technology is good nor that regulation is always wrong. For what it's worth, I had ChatGPT and Claude read the manifesto and asked whether that's a strawman, and ChatGPT said it's a strawman, and Claude said it's "a significant oversimplification and mischaracterization of the essay's content, leaning towards being a strawman argument."
(I don't mean that LLM judgment is a clinching argument, more just noting that I did the basic diligence of checking whether my objection is easily shown to be crazy before I clog up a comment section with it.)
I have edited in a "nearly"; does that address your concern?
Somewhat! The manifesto is 5000 words of vague vibes and platitudes, so that's still more specificity than is there. Your phrasing is a plausible extrapolation, but not directly present in the text.
I'm trying to restrain my pedantry, but if you're going to write "this is not a strawman" it seems especially important to hew closely and not-uncharitably to what OP wrote.
So personally I might go with:
"Andreessen Horowitz, who believe—judging from their Techno-Optimist Manifesto—that because technology is good nearly all possible technology is good and it is anti-science to regulate it."
(though I'd double-check whether "anti-science" is the best term, since it doesn't appear in the manifesto)
But actually I'd just try to find a direct quote from the manifesto that states what I wanted.
I will not be writing a letter in support of this bill. The direct effects of SB 1047 are unlikely to cause any harm, so I have also not written a letter in opposition to the bill, but if it does, as this piece suggests, serve as a necessary first step towards the kind of regulation regime that has killed thousands by slowing the distribution of vaccines, I will be regretting my decision to not try help nip it in the bud. There should be no regulation of the specific content which publicly available generative AI is capable of producing. Pornographic deepfakes are morally disgusting, but preventing them is not worth giving up (through the banning of open source image models, for example) the ability for generative AI to serve the interests of individual humans without corporations and governments interfering.
I read this in the AI forums post? Is this true?
Warning: the Governor's "Contact Us" page (gov•ca•gov/contact) says:
For security, we can’t receive email attachments. If you wish to send an attachment, please do so by mail.
So, if you were planning to send a PDF, I would also send a second email with the letter in the body of the eMail.
It's great to see people take the time to reach out to social justice activists/left-wingers and address their concerns about longtermism and x-risk activism, especially considering in other posts you were more actively dismissive of SJ/left criticisms of EA (eg in your response to Doing EA Better). However, I don't think this is necessarily the most effective place to do such an attempt, the amount of people receptive of the claims of 'AI ethics' against 'AI safety' reading a niche EA blog are likely to be ~nil.
Relatedly: I do have to wonder if the continued association of AI governance with EA is going to be a liability more than an asset starting on from now, and ought to at least be strategically reduced. The natural allies of AI governance and existential risk prevention in general are, as the endorsements for SB 1047 demonstrate, anti-war, environmentalist, animal rights, women's rights, workers' rights movements (i.e. the political left). But EA is possibly (for a certain amount of reasons) one of the most universally hated contemporary social movements by social justice activists. And a movement seeking to address catastrophic risks of irresponsible use of high technology while being backed by tech millionaires is likely to be taken as seriously as a climate change activist NGO backed by Total, or a public health charity backed by Marlboro would be.
Also, the most common line from the left and 'AI ethics' against EA longtermism is that it's an excuse to not do anything for the poor in the present in the name of saving expected 10^^10 lives in the future. Work done by longtermists is actually of direct interest to the poor in areas like biosecurity, but the argument would be much easier to address if not for treating EA as a single actor that believe that every idea (up to including destroying the entire universe in case electrons can suffer) should be considered seriously, but also that you need a peer-reviewed meta-review of RCTs approved by the scientific consensus of several disciplines before saying crimes against humanity are bad. Both of those epistemological stances could stand on their own separately, but taken together they form a noxious double standard.
(A similar case goes for effective animal activism. It's much harder to address the common argument that antispeciesism entail devaluing humans' unique ability to reason and participate in civic life when your closest allies work from the position that one can work to alleviate Third-Worlders's welfare like one would do for farm animals but one ought not to worry about republican liberty as non-domination/Sen-Nussbaum capabilities. Answering that you actually root your concern for animals in the republican arguments of Kymlicka and Donaldson's Zoopolis is likely to be counterproductive for obvious reasons.)
TBH, I would go the opposite direction. AI governance and animal advocacy both have as much support from the political left as they're going to get, time to reach out to people I (Ozy) personally despise.
"AI governance and animal advocacy both have as much support from the political left as they're going to get" That seems very obviously not true to me? Those issues have more mainstream resonance now than 10, 20, or 50 years ago, and it is very strange to assume that we just hit the ceiling when we're really just barely entering the mainstream, and there is plenty of low-hanging fruit about how we communicate.
I can assume from your earlier more actively dismissive attitudes toward SJ/left views on EA that you have a very negative opinion of their epistemics and think improvement is untractable for that reason. But the 'AI ethics' people are vocal in the court of public opinion but not much of a direct hindrance behind the scenes, so the indirect hindrance the former fact pose does not seem like a stable equilibria.
Done. I'm unclear if doing this as a non-American is at all useful — I'm sure not going to be voting for California governors — but it can hardly hurt. That said, how the ass do you people write a letter in only ten minutes?!!
Someone on the EA Forum thinks it's still useful as a non-American:
"You don't have to live in the US to do it. You can help send a powerful message that the entire world is watching California on this issue."
https://forum.effectivealtruism.org/posts/AWYmFwpCrqkknLKdh/how-to-help-crucial-ai-safety-legislation-pass-with-10
Gavin Newsom is well known for leading the charge on gay marriage before it was cool: https://www.latimes.com/politics/la-pol-ca-gavin-newsom-gay-marriage-20180515-story.html
It seems reasonable to guess that Gavin Newsom has fairly cosmopolitan values, and will care some about what people outside of the US/California think, even though they aren't his constituents as governor. As someone who grew up in California -- Californians like to think of themselves as cosmopolitan, and I think the stereotype is fairly accurate.
Writing a letter doesn't need to be high-effort. I like to think of it as sort of like writing an internet comment, except rather more polite and polished (depending on how polite and polished your typical internet comment is -- probably reasonably good if you're reading this blog). Most likely it will simply be read by a staffer who is tallying up the number of letters for and against the bill. I figure the best letter is one which stands out a little, by making an simple/interesting/memorable point.
Also interested in if this is at all useful to do as a non-american