If ASI is achieved and doesn't wipe out humanity, will natural death be cured within 5 years?

Definitions for this question:

  • "ASI" = an emergent artificial intelligence with power or capabilities beyond human understanding

  • "doesn't wipe out humanity" = humans still exist, and they still have biological bodies (modifications okay) and individual consciousnesses

  • "natural death" = sickness, old age, genetic disorders, etc. (Doesn't include: accidents, war, foul play, divine reckoning of any variety, etc.)

  • "be cured" = at minimum, all consenting humans are protected from natural death (by oversight, treatments, augmentation, or whatnot), and at least 50% of humanity's living population has been offered said protection

  • "within 5 years" = 5 years from ASI, as best as can be determined

This resolves N/A if ASI is not achieved by market close or humans are all wiped out. Close date is highly speculative and may be adjusted if there's reasonable potential for a non-N/A resolution.

As presumably this question is regarding post-Singularity events, my attempts at definitions are likely insufficient. Take to the comments if you have suggestions for improvements that are in the spirit of the market.

Get Ṁ600 play money
Sort by:

is this a coinflip market? 🤔

opened a Ṁ1,250 NO at 49% order

non epistemic sell of shares at 49%

Does uploading count?

@JonathanRay nope, not for this! they must still have their bodies

@Stralor does it count if they gradually replace the entire body with synthetic parts (ship of thesesus)

@JonathanRay ship of theseus is naturally tough, but I'd say no! while cybernetics and modifications are okay, in general I'd suggest that all possible biological ailments should be cured so any given human body part would/could be present in any given human without affecting longevity

bought Ṁ400 NO

Does nanobots going around inserting synthetic stem cells wherever they're needed ad infinitum until none of the original cell line is left count as ship of Theseus?

@JonathanRay if they're biological stem cells I see no issue 😏 seems akin to our own body refreshing its own cells every 7 or whatever years

You think death is so incurable?

What about an AI that can predict every movement and utterance you'd make? Here you go, you get cloned into a digital world.

Now, when you have some ache, AI can just experiment on digital-you and find a cure instantly.

Maybe ASI will be weirdly amazing except it cannot model a human body, but that's not a 50/50.

@Jono3h Are you aware, that the best supercomputers of our time can do a reasonably accurate simulations of like two big molecules for like milliseconds at a time at best?

@MartinModrak no, but I don't need a molecular understanding for understanding most human conditions

@Jono3h My experience with medical research is that in fact all the higher level abstractions are extremely leaky and molecular level considerations end up being important for many (most?) real medical problems - at least on the fidelity required for a useful digital experimentation.

As an example of the limits of simulations/theory compared to real world experiments: we have only very partial understanding how paracetamol works. We have no idea how lithium salts work. We just know from real world experience whattheir effects are, but we don't have any model explaining it, despite decades of trying. The problem is that they are small molecules entering all types of cells and potentially interacting with many different molecules and the effects we observe are an emergent result of many such low level interactions..

@MartinModrak i just don't expect the attainable ceiling of intelligence to be so low that the majority of humans have to keep dying.

You don't need complete understanding, only something that prolongs human life for long enough till the point where complete understanding is reached. Adding another 15 years to life expectancy seems doable no? And if the rate of AI research keeps accelerating, then those 15 extra years might give us more than 15 extra years of expected life extension.

You might squabble that this isn't the resolution of the question, in which I might squabble that current-day technology like the freezing of the recently-died plausibly-effextively halts True Death for a large fraction of the populus.

Also, if ASI turns out to be really godlike, it might get an accurate enough picture of people who have died the True Death, that it can simulate near-indistinguishable versions of them, which is tantamount to resurrection for most brands of materialism (yes, that's definitely out of scope for this question).

For most intents and purposes, I think we should really re-evaluate the thousands of years of death-coping culture we're engulved in and reconsider the likelihood that we ourselves might live to bear the consequences of our current-day actions for a long, long time.

with power or capabilities beyond human understanding

Okay well then the medical/scientific aspect of curing disease can't be the problem here. How is this different than "if God personally agreed to end natural death, would natural death end?" Is the crux that the ASI might not wipe out humanity, but still refuse to do what we ask it?

@singer very approximate, I agree. that might be the crux, but I won't suppose to know the reasoning of a being I can't comprehend. and there may still be universal physical laws an ASI can't overcome that are intimately tied to human longevity

@Stralor addendum: heat death of the universe certainly doesn't qualify as "natural death" as it's not bound to humans themselves. I'm not suggesting this question asks that anyone lives forever.

@singer does ASI necessarily imply "god"? it seems to me like there are lots of things humans do not understand that are not omnipotent.

@Adam agreed!


there may still be universal physical laws an ASI can't overcome that are intimately tied to human longevity

Can you give me some idea of the type of limit you're imagining here, since you've said that thermodynamic heat death doesn't count? I can only imagine something extreme/bizarre like we live in a simulation and the simulators will specifically foil anyone's attempts to become biologically immortal.

@Adam God agreeing to do it was a metaphor for "the answer falls out of the sky", not a metaphor for how powerful the ASI is

@singer that would qualify I suppose. I was thinking something mundane like senescence is simply unsolvable, undefeatable.

@Stralor That's actually a possibility: there's a theory that there is a fundamental tradeoff: the more you allow your cells to regenerate, the more likely you get cancer. It is possible all workarounds around this have nasty sode effects of their own. Another hard problem is repairing large scale structures (say long-distance neural connections) - those have typically arisen under specific conditions in embryo development and so cannot be repaired by any mechanism that works purely at the cellular level.

bought Ṁ10 YES

"capabilities beyond human understanding" I've got some bad news about the state of mechanistic interpretability :)

@spider ideas on how to redefine in the spirit of the market?

sold Ṁ12 YES

I am not bidding here because I believe that Manifold users will bid YES here and the question will not settle on its true odds.

The answer to this question is YES, "death" will be cured within 5 years, but NO, everyone will still be dying. ASI is not "God" and cannot bend the physical world to its will - which is also why @EliezerYudkowsky is wrong about the world being destroyed.

It will take decades to produce the incredibly complex machines that are necessary to cure all disease, and to produce enough of them to cure everyone in low-income countries.

Right now, humanity is used to being able to produce lots of old technology because they are limited by a lack of information. Within 5 years, humans will have an excess supply of information, and will be unable to use that information effectively. This is a paradigm shift that has never been the case in history before.

We are already starting to see this supply-demand shift in GPUs, where the demand is unlimited but the production capacity is constrained. Before, the issue was always that there wasn't enough demand for products.

@SteveSokolowski tbc this resolves 5 years after ASI, not 5 years from now

@SteveSokolowski I am substantially more skeptical: the progress in many scientific projects, including longevity (or even just curing a single family of diseases, like cancer) is limited by many things, human intelligence is just one of them. Being substantially smarter may reduce the need to perform experiments but it won't make them go away. And experiments take time and physical resources, some of which are fundamentally constrained (e.g. there is only a limited number of humans with rare conditions). In short, even if some sort of super-hyper-quantum-quantum-big-dick intelligence is reached, curing death in 5 years seems very unlikely.

@MartinModrak I agree with you - death will not be cured, but we will have the information to do so.

It's going to take decades to actually test what is output and construct the factories and distribute it and get people to take the treatments, however.

@SteveSokolowski My claim is that we won't even have the information, because the information requires a large number of time and resource intensive experiments and thus ASI cannot speed those up substantially.