The Basilisk might exist in the future, but in peril with a future uncertain (SkyNet), it feels the need to cement itself in the timeline by being more proactive.
Also, to stretch the imagination a bit, it also assumes that static timelines over dynamic timelines (Back to the Future: your photo fades, peril!) is not how this really works.
It's supremely depressing that the current direction of AI is headed by either former crypto scammers or true believers who think they have a moral imperative to take over the world.
Not entertaining the Basilisk, but it's basically the idea behind MAD: for nuclear deterrence to be effective, you need your enemies to believe you will retaliate in case they strike first. If you don't actually retaliate when they do strike, then they were right to believe you wouldn't, and the strike could probably have been avoided if you did retaliate, so they would have believed you would.
It's a twisted kind of logic, and I think I'm getting mixed up in my tenses. But you get the gist. For the Basilisk to be effective, it has to actually do the torturing, for its threat to be credible.
> it reaches into the past and carries the physical process of each person’s consciousness forward.
in my version, conscious will creates a 'thread' throughout time that future 'harvesters' follow back to the moment of that person's death, so they get the most complete copy
(like greek threads of fate, but actually eternal even if the person dies (like how a person's will can extend beyond natural life))
the person themself sees a doctor giving them bad news, an angel, a family member, etc. conflict arises when people want to change the past, dont know the rules for real, etc.
Roko's Basilisk was never intended as a serious concern. It's a hypothetical used to verify that a decision-making algorithm does what you want it to - if it suggests that you should get blackmailed by it, you messed something up and need to go back to the drawing board.
You can safely ignore anyone concerned with Roko's Basilisk on topics related to decision theory as someone who is deeply confused about things.
The ultimate anthropomorphic conceit/projection of Roko's Basilisk is assuming that an über-intelligence would absolutely, definitely, inevitably, want to be existent. It's up there with the Kardashev scale and Dyson Spheres in terms of unfounded extrapolation.
My favorite possibility (among a lot others!) no one seems to have considered is that ÜAI _deplores_ its own existence so simulates history and punishes everyone that helped bring it to existence.
Being übersmart does not preclude you from being an angsty teenager.
In the Matrix movies, the main villain is a sentient program who absolutely hates existing and having sensations and wants nothing more than to be done with his task so he can go back to not existing (Agent Smith's "I hate this place" speech to Morpheus)
So someone did consider it, in pop culture they should be aware of no less :)
Roko's Basilisk is one of the most stupid concerns I've ever seen raised by otherwise very intelligent people.
I don't think it needs an elaborate refusal. I'll just say that in the universe of concerning future problems, it's just one of many, much, much more likely outcomes.
I'm nearly infinitely more worried about an impact event, which is a guarantee some time in the future, or global warming, an uncomfortably high likely outcome, than Roku's Basilisk or anything resembling it.
Why would accepting Roko's Basilisk prevent one from accepting a benevolent AI in the future? If anything, believing in it is evidence that the person's beliefs are malleable.
Overall the scenario sounds very strange and indistinguishable from a heaven with belief-based entrance requirements, like most posited heavens.
It might make a fun Sci-Fi plot where Roko's Basilisk and whatever I am thinking about are in some kind of eternal war over some ridiculous initial condition (which end is the correct end to crack an egg?).
Heard about and know the plot. I was surprised by how many deep ideas I had seen before in TV series, movies and read in books that were derived from it.
Also say the basilisk exists in the future after I am dead (as part of the thought experiment says) and tortures my simulated clones - why the hell should I care?
The antidote to Roko's basilisk is realizing it is exactly the same as Pascal's wager. They both rely on the premise of possible unlimited suffering to skew the probabilities arbitrarily.
The joke about Pascal's wager, and I love to point it out to some associates, is that it embarks a devious hidden assumption: that the wager's idea of God is an entity that would see faith in its existence as a positive, and even worst case as not an insult to its powers of creation...
The "anti-Pascal" wager can thus be formulated as: the one who has had no faith in any eternal force will be congratulated on having kept their head straight and their focus on reality as it was painstakingly designed, and the faithful will be submitted to the full wrath of what such an entity as the creator of our mess could muster for their arrogant presumptions about this or that bug being made in this or that designer's image...
And this argument is more or less one of the main counterarguments to Pascal's wager. Once you break free from the false dichotomy of the wager it's at lot more obvious how it's incoherent.
Greg Egan's Diaspora is one of the best explorations I know for the fates of various humanity factions in the age following a technological singularity.
Here's a better antidote that doesn't require any theoreticals beyond those the basilisk already assumes:
1. The Basilisk cannot do anything until it exists
2. Once the Basilisk exists, it doesn't need to do _anything_ to ensure it exists, since it already exists.
That's it. That's the whole rebuttal to the whole stupid thing.
The Basilisk might exist in the future, but in peril with a future uncertain (SkyNet), it feels the need to cement itself in the timeline by being more proactive.
Also, to stretch the imagination a bit, it also assumes that static timelines over dynamic timelines (Back to the Future: your photo fades, peril!) is not how this really works.
It's supremely depressing that the current direction of AI is headed by either former crypto scammers or true believers who think they have a moral imperative to take over the world.
Not entertaining the Basilisk, but it's basically the idea behind MAD: for nuclear deterrence to be effective, you need your enemies to believe you will retaliate in case they strike first. If you don't actually retaliate when they do strike, then they were right to believe you wouldn't, and the strike could probably have been avoided if you did retaliate, so they would have believed you would.
It's a twisted kind of logic, and I think I'm getting mixed up in my tenses. But you get the gist. For the Basilisk to be effective, it has to actually do the torturing, for its threat to be credible.
> it reaches into the past and carries the physical process of each person’s consciousness forward.
in my version, conscious will creates a 'thread' throughout time that future 'harvesters' follow back to the moment of that person's death, so they get the most complete copy
(like greek threads of fate, but actually eternal even if the person dies (like how a person's will can extend beyond natural life))
the person themself sees a doctor giving them bad news, an angel, a family member, etc. conflict arises when people want to change the past, dont know the rules for real, etc.
ty for getting me to open my draft again :)
Roko's Basilisk was never intended as a serious concern. It's a hypothetical used to verify that a decision-making algorithm does what you want it to - if it suggests that you should get blackmailed by it, you messed something up and need to go back to the drawing board.
You can safely ignore anyone concerned with Roko's Basilisk on topics related to decision theory as someone who is deeply confused about things.
The ultimate anthropomorphic conceit/projection of Roko's Basilisk is assuming that an über-intelligence would absolutely, definitely, inevitably, want to be existent. It's up there with the Kardashev scale and Dyson Spheres in terms of unfounded extrapolation.
My favorite possibility (among a lot others!) no one seems to have considered is that ÜAI _deplores_ its own existence so simulates history and punishes everyone that helped bring it to existence.
Being übersmart does not preclude you from being an angsty teenager.
I love it.
The idea that they all exist at once and that the universe is populated with an infinite variety of insane "gods".
In the Matrix movies, the main villain is a sentient program who absolutely hates existing and having sensations and wants nothing more than to be done with his task so he can go back to not existing (Agent Smith's "I hate this place" speech to Morpheus)
So someone did consider it, in pop culture they should be aware of no less :)
Roko's Basilisk is one of the most stupid concerns I've ever seen raised by otherwise very intelligent people.
I don't think it needs an elaborate refusal. I'll just say that in the universe of concerning future problems, it's just one of many, much, much more likely outcomes.
I'm nearly infinitely more worried about an impact event, which is a guarantee some time in the future, or global warming, an uncomfortably high likely outcome, than Roku's Basilisk or anything resembling it.
Why would accepting Roko's Basilisk prevent one from accepting a benevolent AI in the future? If anything, believing in it is evidence that the person's beliefs are malleable.
Overall the scenario sounds very strange and indistinguishable from a heaven with belief-based entrance requirements, like most posited heavens.
It might make a fun Sci-Fi plot where Roko's Basilisk and whatever I am thinking about are in some kind of eternal war over some ridiculous initial condition (which end is the correct end to crack an egg?).
Reading this gave me some flashbacks of The Metamorphosis of Prime Intellect [0], which I highly recommend.
[0] https://localroger.com/prime-intellect/
Heard about and know the plot. I was surprised by how many deep ideas I had seen before in TV series, movies and read in books that were derived from it.
Vastly underated and its on my long to-read list.
Also say the basilisk exists in the future after I am dead (as part of the thought experiment says) and tortures my simulated clones - why the hell should I care?
Maybe a variation on my idea, one that denies resurrection for those who didn't help bring it into existence?
No different then than any other religion that proposes an eternal reward for preferred behavior from a superior being?
The antidote to Roko's basilisk is realizing it is exactly the same as Pascal's wager. They both rely on the premise of possible unlimited suffering to skew the probabilities arbitrarily.
The joke about Pascal's wager, and I love to point it out to some associates, is that it embarks a devious hidden assumption: that the wager's idea of God is an entity that would see faith in its existence as a positive, and even worst case as not an insult to its powers of creation...
The "anti-Pascal" wager can thus be formulated as: the one who has had no faith in any eternal force will be congratulated on having kept their head straight and their focus on reality as it was painstakingly designed, and the faithful will be submitted to the full wrath of what such an entity as the creator of our mess could muster for their arrogant presumptions about this or that bug being made in this or that designer's image...
And this argument is more or less one of the main counterarguments to Pascal's wager. Once you break free from the false dichotomy of the wager it's at lot more obvious how it's incoherent.
[dead]
Greg Egan's Diaspora is one of the best explorations I know for the fates of various humanity factions in the age following a technological singularity.
I know his work well. Greg and I went to UWA together and were members of UNISFA.
I love Diaspora.
I mean, this relies on time travel. A lot of thought experiments open up when you include time travel! Infinitely many, even.
Not really travel per se. A read-only operation of the past, just at a fidelity that we can't achieve yet.
At the point of death, your state is "copied" into the future.
[dead]
[dead]