At the risk of necro’ing and old thread, I think we may want to reassess in light of the latest Wait But Why article.
Clearly, at this point cryonics is not an effective way to save a life: costs of up to $250k per life are a far cry from the effectiveness of for example AMF. The moral implications are also not clear: is cryopreserving people morally equivalent to saving people from treatable diseases, for instance? Do we care about aggregate happiness, or suffering? How does cryonics compare to XRisk causes?
There are all very difficult questions, and certainly some of these will have very personalised answers. Still, even for those who consider a successful resuscitation as a result of cryonics of equal or greater value to a life saved by other means (perhaps even if that life then reaches the longevity escape point) it remains matter of basic effectiveness.
I propose that the value of cryonics for EA may be in something beyond its immediate effectiveness: its economy of scale, which relates to its neglectedness. Currently cryonics is a heavily neglected cause; the total number of people signed up in the world don’t even measure into the ten thousands. This is unfortunate, because cryonics is an intervention that benefits greatly from improvements in scale; both by reducing costs, and by improving techniques (thereby increasing the probability of success). This article by Alcor suggests that a world scale cryonics network could make saving a life cheaper than even AMF. Of course, these numbers are likely fairly biased in favour of cryonics; nonetheless, we should take note of this possiblity.
Instead of paying for (impactful) people to be frozen (a recurring cryonics proposal in EA), the real value of cryonics as a cause may be to work on its expansion and ‘normalisation’. This will be a challenge, and the weirdness associated with cryonics may negatively impact public perception of EA, especially since the most effective intervention may involve attempting to change this public perception. This is definitely something to keep in mind. Moreover, the project may not be worth its resources if anti-aging technologies reach the longevity escape point before cryonics becomes sufficiently ‘normal’, although it will still have uses in saving people from then-uncurable diseases.
Estimation is rather complicated; the difficulties are reminiscent of XRisk calculations, although cryonics lacks the feature of the destruction of ‘a hypothetical overwhelming number of future sentients’; instead it covers only a moderate amount of sentients, scaling inversely with advances in medical technology. Despite this, the effectiveness of ‘cryonics normalisation’ (for lack of a better term) as a cause may still be worth discussing, simply because of its tremendous potential.
EDIT: A relevant blog post by Robin Hanson discussing both scale and other charitable aspects of cryonics
I like this view. I think it agrees with my intuition that morality is just a function of whatever a society has decided it cares about. This can make practical sense in the case of ‘murder’, and perhaps not so much in the case of ‘hands on table’. Of course one might then wonder if they ought to care about things making practical sense, which in turn depends on whether that’s a thing we care about, or whatever underlying(/meta-)goal applies; turtles all the way down.
I like that this perspective takes some of the mysticism out of morality by explicitly noting the associated goals/desires. Also generalising obligations to things you should do according to at least one of every self-consistent moral framework is pretty neat. The latter is mainly playing with definitions obviously, but it makes sense as far as definitions go (whatever that means).
I wonder if there exists some set of obligations that caring about/following maximises CEV. Assuming we care about achieving CEV (by definition we might?), this seems like a strong candidate moral framework for everyone to agree on, if such a thing is at all possible.
Possible problems: 1) current volition is not extrapolated volition, so we may not want to care about what we would want to care about, 2) extrapolated volition of different sentients may not converge (then maybe look for the best approximation?).
I’m not that well read in these issues, so please do tell me if I’m clearly missing something/making an obvious mistake.