It's quite similar to what you describe. AI fizzle on the left, and the singularity and paper clips are on the far-right. The other ones are center-right. Futurama and singularity are up, dystopia and paper clips are down.
I really like another of your essays! And I also consider myself a double optimist in the top right quadrant. However, that doesn’t mean the outcomes feared by the other quadrants cannot happen, or that are mutually exclusive —the fears and hopes of all four can happen at the same time! That’s why times of crisis are important. A crisis cannot be solved by logic or trying to anticipate the future. It’s going to happen what it’s going to happen. I believe our fate is decided already, and we passed the Judgement Day like last year or whatever. I loved your last sentence: “When AI enables us to colonize the stars, it will be because AI researchers worked for decades to create AI that aligns with mankind. “
I think this general belief that "good triumphs" is related to a more general question of whether you believe that intelligent agents tend to cooperate. I would guess that people who are less worried that AI will be hostile are also less likely to believe that the US and China are destined to fight a war and so on. I don't have any data to back this up but this is what I'd guess.
I also thought "doomer" was generally used to refer to the Yudkowsky-types in the bottom-right with P(doom)>50%, not to EA and other such people with P(doom) of around 10%. Maybe if you are in some bubble then people with P(doom) of 10% are called doomers.
There's also something to be said for separating prediction and conviction. Sometimes the numbers don't look good. What are you gonna do then? Give up and build a bunker? Actual solutions come from conviction, not from prediction.
> It is faith in people and Goodness, and faith in unbounded technological progress.
Love this! Especially the idea that def/acc people are "double optimists", we believe in the powerful future of technology, and we believe in the ability of humans to steer this technology towards good outcomes for humanity.
On the bottom right quadrant being described as "doomed future", I do think we should slow down, pause, or stop specific technological developments. E.g. it is good that we are trying to prevent the development of mirror life, at least until we're confident we can protect humanity and civilization against its harms.
Given the evidence we have today about the dangers AI could post to humanity, I think it's also reasonable to make the same claim about AI.
Therefore, I wouldn't label the bottom right quadrant as "doomed future". Perhaps the y-axis is measuring "how much do you prioritise accelerating defenses vs decelerating offensive technologies", and the bottom right quadrant might then be labelled something like "vulnerable world" while the top right is "resilient world".
I believe good will triumph, but that's because we will both accelerate defenses and decelerate offensive capabilities.
Yes! I think this is partially captured here: "Others [in the upper right] would prefer to slow frontier AI development, and choose to accelerate defense only because this seems unlikely to happen. Their ideal is probably closer to differential development."
I think you are in the upper right quadrant, maybe close to where Claude or ChatGPT place themselves. (On some level, it is optimistic to believe mankind *could* work together to delay development.)
Have you ever seen this post?
https://scottaaronson.blog/?p=7266
It's quite similar to what you describe. AI fizzle on the left, and the singularity and paper clips are on the far-right. The other ones are center-right. Futurama and singularity are up, dystopia and paper clips are down.
Hadn't seen this –– interesting taxonomy, I like it.
I really like another of your essays! And I also consider myself a double optimist in the top right quadrant. However, that doesn’t mean the outcomes feared by the other quadrants cannot happen, or that are mutually exclusive —the fears and hopes of all four can happen at the same time! That’s why times of crisis are important. A crisis cannot be solved by logic or trying to anticipate the future. It’s going to happen what it’s going to happen. I believe our fate is decided already, and we passed the Judgement Day like last year or whatever. I loved your last sentence: “When AI enables us to colonize the stars, it will be because AI researchers worked for decades to create AI that aligns with mankind. “
Go off, Olivia
I think this general belief that "good triumphs" is related to a more general question of whether you believe that intelligent agents tend to cooperate. I would guess that people who are less worried that AI will be hostile are also less likely to believe that the US and China are destined to fight a war and so on. I don't have any data to back this up but this is what I'd guess.
I also thought "doomer" was generally used to refer to the Yudkowsky-types in the bottom-right with P(doom)>50%, not to EA and other such people with P(doom) of around 10%. Maybe if you are in some bubble then people with P(doom) of 10% are called doomers.
I think the upper right quadrant can have a high sans-intervention p(doom), but still have faith in the interventions.
Well I was thinking about actual existential doom not airborne Ebola. Of course I’d put you on the top right.
There's also something to be said for separating prediction and conviction. Sometimes the numbers don't look good. What are you gonna do then? Give up and build a bunker? Actual solutions come from conviction, not from prediction.
I agree with this.
> It is faith in people and Goodness, and faith in unbounded technological progress.
Love this! Especially the idea that def/acc people are "double optimists", we believe in the powerful future of technology, and we believe in the ability of humans to steer this technology towards good outcomes for humanity.
On the bottom right quadrant being described as "doomed future", I do think we should slow down, pause, or stop specific technological developments. E.g. it is good that we are trying to prevent the development of mirror life, at least until we're confident we can protect humanity and civilization against its harms.
Given the evidence we have today about the dangers AI could post to humanity, I think it's also reasonable to make the same claim about AI.
Therefore, I wouldn't label the bottom right quadrant as "doomed future". Perhaps the y-axis is measuring "how much do you prioritise accelerating defenses vs decelerating offensive technologies", and the bottom right quadrant might then be labelled something like "vulnerable world" while the top right is "resilient world".
I believe good will triumph, but that's because we will both accelerate defenses and decelerate offensive capabilities.
Yes! I think this is partially captured here: "Others [in the upper right] would prefer to slow frontier AI development, and choose to accelerate defense only because this seems unlikely to happen. Their ideal is probably closer to differential development."
I think you are in the upper right quadrant, maybe close to where Claude or ChatGPT place themselves. (On some level, it is optimistic to believe mankind *could* work together to delay development.)
Hard to capture this in a 2D graph.