@Richard_Ngo wrote in his quick take: “I also think that AI safety is in a much worse position today than it would have been if it’d taken a more principled truth-seeking stance a decade ago (for example, AI governance would have been less likely to throw in with the Democrats in a way that alienated MAGA). ” As far as I understand Ngo’s position, the counterfactual AI safety which chose the Reps instead of the Dems would have a bigger chance of taking over Trump. Alas, I don’t understand how exactly Trump would be convinced to do things like coordinating with China on pacing the frontier or implementing UBI.
I went into detail on this a while back as it relates to the Sierra Club. America used to have a strong, bipartisan environmentalist movement. In the mid-1990′s, David Gelbaum paid $200 million to them on the condition that they take a hardline pro-migrant stance and deplatform the members that opposed this. The half of the environmentalist movement that disagreed scattered to the four winds, forever turning environmentalism into a partisan issue that would, every 4-8 years, encounter an administration to which it did not have any pre-existing connections.
On balance, Gelbaum singlehandedly crippled American environmentalism beyond what any coal industry lobbyist could ever have hoped to achieve. Half of its potential members were alienated, one of its biggest lobbying organizations fell into infighting and purity spiraling, and it was permanently stapled to an unrelated issue that generally polls unfavorably.
When I see people advocating that Scott Weiner, infamous among all of the right and a good chunk of the center for SB145[1], be openly touted as representative of AI safety, I worry about history repeating itself. There’s lots of outreach that could have happened, and lots of own-goals that didn’t have to be scored.
In the mid-1990′s, David Gelbaum paid $200 million to them on the condition that they take a hardline pro-migrant stance and deplatform the members that opposed this. The half of the environmentalist movement that disagreed scattered to the four winds, forever turning environmentalism into a partisan issue that would, every 4-8 years, encounter an administration to which it did not have any pre-existing connections.
I think that this is a good illustration, but an overblame of Gelbaum. As for Wiener becoming representative of AI safety, my main question would be “who else is a better candidate?” I would suggest Arizona senator Ruben Gallego if we have the ability to contact him.
IMO the most plausible such case is Trump’s position on the issues which aren’t related to AI or UBI conditioned on AI. I would illustrate it by things like Abigail Shrier attacking Newsom for taking measures which, according to Shrier, are evidence that the Dems are making obvious mistakes. Alas, the Dems’ AI-related strategy is NOT an obvious mistake...
It’s dumb imo, but the guy can flip on a dime depending on the vibes. It’s more chaotic, but he might somehow trip into the right basin. He also might have significantly increased the odds of a Democrat president in 2029.
Trump seems less likely to introduce rash regulation that severely curtails the benefits of AI. He doesn’t seem to be influenced by populist job loss concerns (yet).
Trump seems less likely to redirect effort from real alignment problems (ie. instruction following) to fake alignment problems (ie. ”ethics” stuff) through either rhetoric or regulation. For example he won’t want an AI that the president can’t tell what to do, which requires corrigibility (good).
One also gets the feeling that Trump has a flexibility unique among recent presidents—for example, that he would just withdraw from the Iran war right away if he decided he didn’t want to continue. While he is dismissive of most AI safety efforts now, I think he could take unusually decisive action if the situation changed.
I’m very happy with his strong support for data centers as well, while politicians on both sides oppose buildout.
Anybody still want to make the “EA case for Trump”? @Richard_Ngo?
Too early to tell.
@Richard_Ngo wrote in his quick take: “I also think that AI safety is in a much worse position today than it would have been if it’d taken a more principled truth-seeking stance a decade ago (for example, AI governance would have been less likely to throw in with the Democrats in a way that alienated MAGA). ” As far as I understand Ngo’s position, the counterfactual AI safety which chose the Reps instead of the Dems would have a bigger chance of taking over Trump. Alas, I don’t understand how exactly Trump would be convinced to do things like coordinating with China on pacing the frontier or implementing UBI.
I went into detail on this a while back as it relates to the Sierra Club. America used to have a strong, bipartisan environmentalist movement. In the mid-1990′s, David Gelbaum paid $200 million to them on the condition that they take a hardline pro-migrant stance and deplatform the members that opposed this. The half of the environmentalist movement that disagreed scattered to the four winds, forever turning environmentalism into a partisan issue that would, every 4-8 years, encounter an administration to which it did not have any pre-existing connections.
On balance, Gelbaum singlehandedly crippled American environmentalism beyond what any coal industry lobbyist could ever have hoped to achieve. Half of its potential members were alienated, one of its biggest lobbying organizations fell into infighting and purity spiraling, and it was permanently stapled to an unrelated issue that generally polls unfavorably.
When I see people advocating that Scott Weiner, infamous among all of the right and a good chunk of the center for SB145[1], be openly touted as representative of AI safety, I worry about history repeating itself. There’s lots of outreach that could have happened, and lots of own-goals that didn’t have to be scored.
A very controversial bill aimed at decriminalizing adults having sex with minors in certain circumstances.
I think that this is a good illustration, but an overblame of Gelbaum. As for Wiener becoming representative of AI safety, my main question would be “who else is a better candidate?” I would suggest Arizona senator Ruben Gallego if we have the ability to contact him.
What is the counter-counterfactual case for voting for trump in real life, given that the world is arranged such that it is?
IMO the most plausible such case is Trump’s position on the issues which aren’t related to AI or UBI conditioned on AI. I would illustrate it by things like Abigail Shrier attacking Newsom for taking measures which, according to Shrier, are evidence that the Dems are making obvious mistakes. Alas, the Dems’ AI-related strategy is NOT an obvious mistake...
It’s dumb imo, but the guy can flip on a dime depending on the vibes. It’s more chaotic, but he might somehow trip into the right basin. He also might have significantly increased the odds of a Democrat president in 2029.
Trump seems less likely to introduce rash regulation that severely curtails the benefits of AI. He doesn’t seem to be influenced by populist job loss concerns (yet).
Trump seems less likely to redirect effort from real alignment problems (ie. instruction following) to fake alignment problems (ie. ”ethics” stuff) through either rhetoric or regulation. For example he won’t want an AI that the president can’t tell what to do, which requires corrigibility (good).
One also gets the feeling that Trump has a flexibility unique among recent presidents—for example, that he would just withdraw from the Iran war right away if he decided he didn’t want to continue. While he is dismissive of most AI safety efforts now, I think he could take unusually decisive action if the situation changed.
I’m very happy with his strong support for data centers as well, while politicians on both sides oppose buildout.