In the eyes of many, Biden’s Executive Order somewhat overshadowed the UK Summit. The timing was unfortunate. Both events were important milestones. Now that I have had time, here is my analysis of what happened at the UK Summit. As is often the case with such events, there was a lot of talk relative to the amount of action. There was a lot of diplomatic talk, talk of that which everyone agrees upon, relative to the amount of talk of real substance. There were days of meetings that resulted in rather unspicy summaries and resolutions. The language around issues that matter most was softened, the actual mission in danger of being compromised.
A meta questions - how do you keep track of all those things that happen over the span of months, so you can pull them up and reference them when writing the posts? Do you use something like obsidian or another software?
I would like to see more fear and anger in the population, not less. That is not dangerous, that is the only thing that will save human primacy. Large amounts of fear and anger as soon as they can be generated, strong taboos against the use of AI, people treating AI capabilities researchers as if they clubbed baby seals for a living, etc, are the only way out of this.
I understand the summit had to use diplomatic language to get tech involved, but this is not encouraging for those of us who wish to avoid an AI future altogether. I don't consider any statement that calls AI a potentially positive societal transformation to be a good first step, it's basically an acknowledgement of surrender from Day 1. The Day 1 position should be "this should never be built, if you want to convince us to allow it to be built you're gonna have to spend a very long time proving it's desirable for humans, and if you try to build it before we give you the green light you will be stopped by the full range of force available to our government." Europeans employ the precautionary principle for mundane chemicals in plastics, but the most destructive technology ever conceived gets to continue until it reaches some arbitrary danger point?!
Nitpick regarding the Chinese statement: "我们相信" means "we believe". The Mandarin term for existential risk is 生存性风险, at the end of the second paragraph. My Mandarin isn't great and I didn't read the whole thing, but I think the two versions are intended to be word for word translations and GPT-4 is completely off base here.
Andrew Yao is as influential a figure as it's possible to be within Chinese computer science. I don't think you could have gotten a bigger name to sign it -- he is certainly higher-profile than anyone on the US side.
He has an important position at Tsinghua and is held in extremely high esteem. Also he's the founder of the "Yao class" at Tsinghua which is the most prestigious undergrad CS program in China, and whose graduates have a fairly powerful alumni network in the tech industry as well as global CS research.
Whether this translates into actual policy influence for him, I don't really know, but his name carries serious weight.
One thing I'm quite curious about, and am going to try to track, is the degree to which in the US future policy statements begin quoting the Bletchley summit declaration in addition to the EO.
(If yes, that implies that bureaucratic actors think that it is influential inside government.)
At least everything is directionally correct here.
A meta questions - how do you keep track of all those things that happen over the span of months, so you can pull them up and reference them when writing the posts? Do you use something like obsidian or another software?
I would like to see more fear and anger in the population, not less. That is not dangerous, that is the only thing that will save human primacy. Large amounts of fear and anger as soon as they can be generated, strong taboos against the use of AI, people treating AI capabilities researchers as if they clubbed baby seals for a living, etc, are the only way out of this.
I understand the summit had to use diplomatic language to get tech involved, but this is not encouraging for those of us who wish to avoid an AI future altogether. I don't consider any statement that calls AI a potentially positive societal transformation to be a good first step, it's basically an acknowledgement of surrender from Day 1. The Day 1 position should be "this should never be built, if you want to convince us to allow it to be built you're gonna have to spend a very long time proving it's desirable for humans, and if you try to build it before we give you the green light you will be stopped by the full range of force available to our government." Europeans employ the precautionary principle for mundane chemicals in plastics, but the most destructive technology ever conceived gets to continue until it reaches some arbitrary danger point?!
The comments about misinformation, which I do agree with, make me wonder if Elon Musk was ahead of the curve with his blue ticks program?
Maybe once/if we have 10x as many LLM "bots" than people, many more will want to know which is which. Or maybe we will just prefer the bots?
Nitpick regarding the Chinese statement: "我们相信" means "we believe". The Mandarin term for existential risk is 生存性风险, at the end of the second paragraph. My Mandarin isn't great and I didn't read the whole thing, but I think the two versions are intended to be word for word translations and GPT-4 is completely off base here.
Andrew Yao is as influential a figure as it's possible to be within Chinese computer science. I don't think you could have gotten a bigger name to sign it -- he is certainly higher-profile than anyone on the US side.
He has an important position at Tsinghua and is held in extremely high esteem. Also he's the founder of the "Yao class" at Tsinghua which is the most prestigious undergrad CS program in China, and whose graduates have a fairly powerful alumni network in the tech industry as well as global CS research.
Whether this translates into actual policy influence for him, I don't really know, but his name carries serious weight.
One thing I'm quite curious about, and am going to try to track, is the degree to which in the US future policy statements begin quoting the Bletchley summit declaration in addition to the EO.
(If yes, that implies that bureaucratic actors think that it is influential inside government.)