Similar argument was presented as utilitarian argument against quantum immortality by some writers on LW.
They argue: The number of branches of multiverse where I survive is declining after quantum suicide, so the total utility is declining, so I should not care about QI.
Quantum immortality instead is based on the idea of preserving continuity.
Traditional utilitarianism doesn't care about continuity (and also about uniqueness and about complexity of experience). I think that we should update utilitarianism, as it was created as an apoximation of what is really good and should not be taken as religious dogma.
2 Grabby aliens have incentives to send AI-viruses as SETI-attack to naive civilizations and thus be as louder as any spammer. (may be disguised as 1)
3 The same argument as SIA can prove zoo-hypothesis which is very similar to silent aliens (but doesn't require global acasual coordination, it is local). Aliens create or support existence of many independent biological worlds for experimentation or just for pleasure. It also similar to simulation hypothesis.
That is why we can target Andromeda - the distance is good enough so they didn't arrive yet and we can focus on many stars simultaneously and hope that aliens have very large receivers - maybe Dyson sphere-size. Also the main point is to affect local superintelliegce expected utility.
The main point you miss is that the main goal of all this is to get a better negotiation position on Earth with our own superintelligence (and there are high chances that it will be soon and misaligned)- which at early stages will also very uncertain about aliens detraction in the universe. It is like crying into an open window - "a guy, Jack, is in my room and is going to kill me!" You don't know if there is police outside - but you hope to affect Jack's calculus.
After it crosses Saturn orbit, it will take 5.25 years to get to the Sun. Unlikely it will be observed before crossing Saturn orbit as there is no constant monitoring there and it is very dim there.
I am working on sideloding – creating mind models via LLM - which allows performing different experiments including the ones about the ethics of uploading. Will it be interested to you? More: https://github.com/avturchin/minduploading/
One thing about DA is that I have to know that I am selected randmoly from the group of observers. A first AI will know that it is the first, so there is no randomness.
If there are many AIs, then there is no singleton and opinion of each is not important.
Also I think that DA and anthropics is solvable and the most of assumptions are true but in slightly different situations: they answer different questions.
However, such full anthropic solution require a lot of cognitive work to map full anthropic landscape. An AI may use some conservative approach before it is solved adding weights to different solutions.
Of course, I meant not Bronze age collapse, but known plethora of existential risks. But your argument that others will outcompete us is valid - unless the totality of x-risks is a universal Great Filter.
Similar argument was presented as utilitarian argument against quantum immortality by some writers on LW.
They argue: The number of branches of multiverse where I survive is declining after quantum suicide, so the total utility is declining, so I should not care about QI.
Quantum immortality instead is based on the idea of preserving continuity.
Traditional utilitarianism doesn't care about continuity (and also about uniqueness and about complexity of experience). I think that we should update utilitarianism, as it was created as an apoximation of what is really good and should not be taken as religious dogma.
I agree with general logic here, but I see three exceptions:
1. Early civilizations which are under risk of AI takeover have a strong incentive to transmit their values and-or call for help as loud as possible https://www.lesswrong.com/posts/AH4pFd3Sgt6qaoH6h/warning-aliens-about-the-dangerous-ai-we-might-create and https://www.lesswrong.com/posts/2xHhe4EBHAFofkQJf/plan-e-for-ai-doom?commentId=DR29CwhMhKGEnxKt8
2 Grabby aliens have incentives to send AI-viruses as SETI-attack to naive civilizations and thus be as louder as any spammer. (may be disguised as 1)
3 The same argument as SIA can prove zoo-hypothesis which is very similar to silent aliens (but doesn't require global acasual coordination, it is local). Aliens create or support existence of many independent biological worlds for experimentation or just for pleasure. It also similar to simulation hypothesis.
Digital recreation of Igor: https://forum.effectivealtruism.org/posts/JLHxnWxXk4zYwFcHs/recreation-of-ea-pioneer-igor-kiriluk
That is why we can target Andromeda - the distance is good enough so they didn't arrive yet and we can focus on many stars simultaneously and hope that aliens have very large receivers - maybe Dyson sphere-size. Also the main point is to affect local superintelliegce expected utility.
The main point you miss is that the main goal of all this is to get a better negotiation position on Earth with our own superintelligence (and there are high chances that it will be soon and misaligned)- which at early stages will also very uncertain about aliens detraction in the universe. It is like crying into an open window - "a guy, Jack, is in my room and is going to kill me!" You don't know if there is police outside - but you hope to affect Jack's calculus.
Between 1 and 10 years - likely 3
After it crosses Saturn orbit, it will take 5.25 years to get to the Sun. Unlikely it will be observed before crossing Saturn orbit as there is no constant monitoring there and it is very dim there.
I am working on sideloding – creating mind models via LLM - which allows performing different experiments including the ones about the ethics of uploading. Will it be interested to you? More: https://github.com/avturchin/minduploading/
if it is first and knows that it will kill all others
One thing about DA is that I have to know that I am selected randmoly from the group of observers. A first AI will know that it is the first, so there is no randomness.
If there are many AIs, then there is no singleton and opinion of each is not important.
Also I think that DA and anthropics is solvable and the most of assumptions are true but in slightly different situations: they answer different questions.
However, such full anthropic solution require a lot of cognitive work to map full anthropic landscape. An AI may use some conservative approach before it is solved adding weights to different solutions.
Of course, I meant not Bronze age collapse, but known plethora of existential risks. But your argument that others will outcompete us is valid - unless the totality of x-risks is a universal Great Filter.