I think you're going to need quite powerful AI to make serious progress on aging. In fact, what you're really going to need is systemic tissue replacement, and those replacements are going to take time to grow.
I think that in turn makes it pretty likely we get bigger warning shots than we've gotten thus far.
My best guess is that results in net negative sentiment towards accelerationism. If you look at the general public's answers to questions like "would you rather have a million dollars of a 50% chance of $100 million" and you'll see that the average person is quite conservative. I'd be a little surprised if they want gamble on AI-enabled radical longevity if they think there's a serious danger of AI job loss and/or death.
My bigger concern at the moment is that we might not get another warning shot because AI is become situationally aware enough to falsify logs. And we seem to be getting pretty close to RSI.
Agreed that people seem very conservative with respect to risk and disruption, and this probably tips the scales towards being fearful of AI rather than going full e/Acc. And furthermore, indeed -- making "serious" progress on aging is probably really hard, and even earlier glimmers -- random cancer cures, promising results-in-mice, some drug that can marginally slow some subcomponent of aging (like what people hope rapamycin might be) -- will take time to show up and then get proven out and affect people's views. Which again favors the forces of AI safety over e/Acc.
But on the other hand, it's uncertain / hard to say. Maybe the idea that a new technology ("it's just math!" it's just computers!") could truly have agency and pose an active threat to mankind, is just fundamentally bizarre to people, compared to the idea that a new technology could deliver a bunch of cancer cures or etc. (Normal people are not going to be aware of the crazy gompertz-curve fact about how even eliminating all cancer would only raise expected lifespans by like 2 years -- many folks would probably perceive "we get a bunch of impressive cancer cures in short succession" as nearly equivalent to curing aging.) Or maybe some powerful group, like -- *checks twitter* -- whoever's President of the United States, or CEO of Nvidia, or leader of whatever AI startup seems poised to make a lot of medical breakthroughs, decides to put a ton of effort into boosting this message.
Maybe the net effect of the two arguments fully entering the public consciousness is indeed favorable to AI safety. But a longevity + e/Acc preference cascade could still probably cause a lot of trouble -- eg, further polarizing AI-related issues, perhaps enough to make a strong, coordinated, swift response to AI emergencies much more difficult.
Also agreed that if we're already really close to RSI, all these considerations probably don't matter very much. They seem more important in longer timelines (whether because AI development naturally still has a long way to go despite appearances, or because we successfully pace / pause development).
I think an anti-anti-aging movement is more likely, on the model of the anti-data-center movement: "stop the rich from creating a sinister transhumanist world"...
I'm not so sure- I think anti-longevity arguments like that tend to be mostly a social performance that wouldn't survive the shift in social incentives that would come with the public actually thinking anti-aging technology was possible.
Currently, most people see the idea of anti-aging technology as absurd, and therefore strongly associate it with grifters, crazy fringe people, and fictional tropes about evil people seeking immortality as a demonstration of their hubris and/or rejection of the Christian afterlife. When the subject of anti-aging research comes up, people are therefore very strongly incentivized to perform opposition to the idea- they have to come up with counter-arguments on the spot to demonstrate that they aren't taken in by the charlatans and protect their social credibility.
Often, they'll turn to thought-terminating cliches like "death is a part of life" or "there's nothing more selfish than wanting to live forever". Other times, they'll come up with very poorly thought-out arguments about overpopulation, immortal dictators or cultural stagnation. Sometimes, especially on the Left, they'll turn the low status of the idea into a weapon and use it to attack a political target like the Silicon Valley technology brothers.
These arguments fall apart with a few minutes of honest thought, but that doesn't usually matter much to the people making them- the arguments work just about as well as a social performance of opposition to charlatans whether true or false.
The arguments may also be in part a way of coping with a fear of death, though I suspect that's very secondary to worry over loss of status in practice; distant existential worries tend to be a lot less emotionally salient than immediate social fears.
If there were a bunch of medical breakthroughs that convinced most people that anti-aging technology might become possible in the near future, however, then being associated with that idea would no longer be a status risk. I think those arguments- including the political attacks- would mostly then be abandoned (except by a handful of people who had accidentally invested way too much social capital in them).
My mental model for social acceptance of anti-aging drugs is: obesity and Ozempic.
The moment we got a drug that's good enough to actually put a dent in obesity, a lot of "healthy at any weight" and "I'm fine the way I am" went out of the window - in favor of weekly injections.
It's not that people were somehow unaware that being overweight is actually bad for you. It was pretty obvious. It's just that they didn't see the point in fighting it. The moment they saw that there's a drug that actually works? Saw a coworker drop 20 pounds, or a celebrity go from "gaining weight" to "lean and thin"? The effort of "lose weight" went down, the return went up - and suddenly, there was no point in trying to cope your way into obesity not being all that bad.
I expect something very similar to happen to aging. If a drug that actually puts a dent in aging hits the shelves, we'll see the same effect: the true preference for "I don't want to grow old and frail" would override the "I can't fight aging, so let's think of a way in which aging is good" cope.
This is, however, contingent on the "miracle anti-aging drug" having an obvious, undeniable effect. Aging is slow. If you have a drug that simply slows aging by 20% for as long as you take it, with no immediately visible effect? It might make ripples in the medical community, gym crowds, etc - but fail to get a broad societal impact. If we got a "miracle anti-aging drug" that visibly improves skin quality and deletes wrinkles? That makes a "preference cascade" much more likely.
Even though I emphasized in my post that preference cascades often happen in a jerky, suddenly-all-at-once way, I also think that there are scales to these things: there will probably be many mini preference cascades, some larger than others. (eg, the current AI safety moment isn't the first time that a wave of new people have suddenly started worrying about AI x-risk, nor will it be the last.)
Bit's post from last month seems to be arguing that even the (IMO kind of scam-adjacent and unvirtuous, even though both are probably sincere) efforts of Bryan Johnson and David Sinclair constitute anti-aging ideas moving into the mainstream. To me, that feels like a tiny drop in the bucket and not at all what I was thinking about by preference cascade. (Like, okay, some random people on twitter got interested, and now the longevity community has tripled in size from "tiny" to "still tiny, but now bolstered by some people who fall for twitter hype"... seems quite a bit smaller than this AI safety moment which is getting tons of front-page headlines, attention from current and former presidents of the united states, etc.)
An "ozempic for aging" -- some treatment that undeniably de-ages / rejuvenates people in some clear way -- would indeed give rise to probably the biggest and most thunderous cultural pivot on this issue. But I think even something more minor (like a bunch of undeniable impressive cancer cures) or less clear-cut (like a drug that slows aging by 20%), might still have a lot of preference-cascade potential. Sure, you aren't going to recognize an overnight change in your elderly neighbor who's taking the new 20% drug. But if the results are firm enough to convince the scientific community on academic questions like "Is this really targeting the core mechanisms of actual aging? Rather than just (like metformin or ozempic or statins) producing an apparent lifespan-enhancing effect simply by treating ailments, like metabolic dysfunction or heart disease, which are very common?", then I could imagine that producing quite a vibe shift among medical researchers, doctors, regulators, etc. Which would then percolate out to 1. science journalists, the New York Times, etc, and then 2. elite upper- and upper-middle-class folks who take their health relatively seriously and read the journalism, and then 3. some indeterminate but significant portion of the wider, less-elite population.
I like the framing of anti-aging efforts vs. pausing AI, because it helps us estimate the floor of human suffering caused by delaying technological progress (like achieving LEV). So whenever someone argues for pausing AI, their argument should, imo, address the fact that millions of people will die each day aging cure is not found/deployed.
Cryonics also seems particularly relevant here as a shovel-ready plan that acts as a “Plan B” alternative vision for longevity.
If cryonics is the next rationalist hobbyhorse to hit the mainstream, that'd be amazing. Maybe we gave up on raising the sanity waterline too quickly!
I agree there's a potential preference cascade here. Worth remembering that the most reliable voters are the old, who will be much more willing to risk short timelines!
I think anti-aging is very different from ai safety in one perspective: it's affected by social wealth distribution. Better ai safety results in AI causing less trouble like Huggingface incident, and less x-risk, so everyone benefits from it. However the anti-aging cure could be expensive, and limited to only the riches and those in charge of political power.
I think this difference will significantly affect how the public views it.
That's not how pharma works.
For most interventions, we see the same pattern of "R&D is expensive, drug is cheap". So there's a massive financial incentive to sell the resulting drug as widely as possible.
For rare disorders, the uptake has a very hard limit, so the drugs targeting them can stay expensive quite easily.
Aging? It's the opposite of a rare disorder.
For anti-aging treatments to be gated behind "social wealth distribution" for any extended period of time, they must rely on something that doesn't scale well - things like invasive surgery or expert input. And with advanced AI, even "expert input" is not the blocker it once was.
e/acc is very unlikely to go mainstream, as their worldview is seen as crazy. Don't forget that even a paced frontier is still singularity material, just a singularity that can maybe be harnessed rather than nanobots eating the sun~
I liked this post, except for the section where you seemed to suggest having crazy plans to reshape the government in favor of longevity even if you don't believe them or that they'll work? I'm okay with somewhat strategically mentioning something you happen to believe, and not spending resources enacting a plan that you endorse if it's not worth it, but I definitely don't endorse endorsing a plan that you don't actually think is a good idea for strategic reasons. Not saying you necessarily meant this - just that it read like you might've.
I happen to think that a medicine futarchy would be a-mazing, but that getting it to actually happen probably would be a waste of resources unless you had better knowledge of politics and government and thought you could make it real, and even then your time is probably better spent on AI x-risk prevention. If you genuinely agree, then I don't think it's a problem to have strategic reasons influence you mentioning the genuine truth, but lying about your endorsed plans is a step too far. Again, not saying you necessarily meant this, it could just have been poorly phrased or maybe I'm reading especially uncharitably.
First, I think this is a good article. However, I have a question about one of your premises:
“What is the problem?? How can you fail to recognize this! Do you not AGREE that DEATH IS BAD?”
Has anybody done a second-level analysis of what happens if we discover immortality tomorrow? While I'm not personally crazy about dying tomorrow, I think that, viewed from the perspective of an entire race, the answer is not so obvious. Despite the fact that Malthus was early with his prediction of widespread disaster, there is still some finite carrying capacity of the Earth. We're already seeing clear effects from climate change with 8 billion people. If nobody dies, either we stop having children, or the population balloons very rapidly, putting more strain on our resources.
I'm old enough to have lived through the Green Revolution that provided fertilizer, seeds, and pesticides that dramatically increased what a farmer could produce. There are also plenty of other examples of big improvements in technology as well, so I don't rule out the possibility that we could support an order of magnitude more people on the planet, but I also think it's far from a given.
There is also the issue that you see in various research fields, politics, etc. Progress is said to happen one funeral at a time; if none of the senior people die, how do the younger ones ever move forward? Do we really want a thousand-year-old politician (of any stripe) still sitting in office? On the flip side, will the average worker who only does the job to make the pay want to continue working for centuries? If they don't, who supports them? Do we end up like Logan's Run and enforce maximum age limits?
I don't know if any of these things are likely to be real problems, but failure to plan is planning to fail. So, I'm wondering if there are any good papers, articles, etc. that examine the societal and planetary consequences of immortality.
A good way to show anti-aging bona fides is to fund Intervene Immune. It has a very promising treatment, and is very funding constrained. Claude thinks rapamycin trials are more promising, but it thinks that Intervene Immune's treatment would only improve thymus function. I see much broader benefits, based on unpublished evidence.
I wrote more about this here.
I'd argue that at least with current AI, anti-aging research is an extremely ill-suited problem, precisely because the feedback loops are so long. This is an issue with current, human-run aging research programs. If you want to see whether X extends lifespan in mice, you kinda have to wait for the mice to grow old. There are some biomarkers that can give faster feedback, but the usefulness of them is contested - ultimately, a ground truth signal takes on the order of an organism's lifespan. For homo sapiens, that's the better part of a century.
I think it'd be very worthwhile for someone to do an AI 2027 style narrative forecast of life extension trajectories, if only so that it becomes easier to talk about scenarios where people might be more willing to slow down AI development because they reasonably come to believe they're going to live long enough to see the benefit of powerful AI anyway.
I think AI 2040 already makes this point in abundance. (An AI slowdown will not actually feel like a slowdown at all)
There's a major difference between the two.
AI Safety went mainstream because it reinforced a preexisting bias against AI - people afraid for their jobs, tired of slop and convinced that it's bad for environment. There was already a significant (even if highly fragmented) anti-AI movement before the general population started seriously considering x-risk.
Longevity won't be benefitting from the preexisting bias against death. Culturally there's no such thing. We don't have people protesting against death and writing articles about how bad death is.
What we have are preexisting biases against medical interventions (including well-proven ones, like vaccines), "playing God", wealth concentration and gerontocracy. Longevity will have to fight an uphill battle to overcome all of those.
In a basic sense people’s apparent “preference” for aging I think comes from the underlying sense that whilst immortality might be great for me it might be bad for society in general.
All over the world we already have huge issues with wealth distribution favouring the elderly and political systems entrenching that as the age distribution tends older. 60% olds outvote 40% youngs and so it goes.
Conservatism has a lot of these kinds of ”chesterton’s fence“ style scenarios. There are plenty of taboos which are difficult to rationally explain because they exist as a result of third, fourth, fifth order effects. (There are also plenty of taboos that make or no longer make sense).
To add to this, as an example that is a bit off the original thread - Ozempic is brought up above. There is a taboo, at least in some quarters, on taking weight loss drugs. I think that this taboo exists because people believe that weight control is a willpower thing and that we want more people with willpower and fewer people without willpower and that makes a better society.
Anti-aging sentiment might go rapidly mainstream, in the same way AI Safety just did.
Recently, the AI safety community has been enjoying a massive preference cascade that has rapidly moved AI x-risk concerns into mainstream political discourse
Why did this happen?
But this isn’t the only preference cascade waiting in the wings, as we continue to scale AI capabilities. Consider another throwback rationalist classic:
The labs are probably working towards this.
Just like the idea of AI x-risk, the hope of using AI to unlock medical cures and other transhumanist goods is widespread among both employees and corporate leaders at the frontier AI labs, rationalists, etc.
Insofar as they aren’t devoting 100% of their energy towards steering to RSI + trying to do some critical safety work along the way (these two things seems to be most of what they’re doing), Anthropic also seems to be steering somewhat out of their way to push medical capabilities – consider for instance how centrally medical science features in Dario’s essay “Machines of Loving Grace”, their detailed coverage of the biology capabilities of each new model, their development of a project-glasswing-style “Life Sciences Verification Program” for qualified access to bioweapons-research-capable models, how they've allegedly opened a wet lab for biology research, etc. Deepmind also seems to have also devoted a disproportionate amount of effort towards biology research, via efforts like AlphaFold and many others.
Lots of smart people seem to believe that AI-for-longevity is possible, and they’re working hard to make it happen. Maybe they’ll pull it off!
Although perhaps a cascade wouldn’t happen or would come too late to matter.
Maybe anti-aging is just fundamentally too weird to enter the overton window?
Evidence for the tractability of anti-aging medicine might build very gradually and smoothly.
A longevity x e/acc movement might simply take too long to gather steam, thus arriving too late to be relevant.
A longevity preference-cascade would be great! Except, by default, the energies unlocked might flow directly into e/acc :(
This cultural earthquake, or avalanche, or dam-breaking, catalyzing massive societal energies for the fight against aging, is something that the longevity community has dreamt about for for years. Unfortunately, the anti-aging preference cascade might, in practice, channel all its energies into AI-accelerationist sentiment.
Before our current AI safety moment, surely there was nevertheless widespread anxiety about humans being replaced by AI. (See eg the dramatic unpopularity of anything AI-themed in various polls.) But since the idea of AI x-risk wasn’t being discussed publicly, it strikes me that those worries were essentially being “sublimated” (in the psychological sense) in favor of somewhat vague and incoherent anger about datacenters, flock cameras, copyright infringement, et cetera. This sublimation process artificially boosted those movements but (until now) greatly weakened public enthusiasm for AI x-risk mitigation.
Right now, humanity’s terror at the prospect of death is similarly sublimated in a variety of ways, since after all there doesn’t seem to be anything practical we can do about it. But it’s also such an intense terror that it arguably drives a huge percentage of human motivation.
Meanwhile, today’s e/acc movement is honestly kind of pathetic. Its main argument within the current overton window is just that AI is going to make a lot of money. (“...though not necessarily for YOU, average american worker – I was more thinking about my silicon-valley investor friends.”) This is enough to convince Trump (at least for now) that opposing AI safety is essential for making sure the stock market keeps going up. There are a few true e/acc believers who understand that the biggest potential prize is longevity (alongside many other transhumanist goods, like intelligence enhancement). This has little traction with the public, and it seems kind of kooky to even imagine that ordinary people (people like my mom! my neighbors!) might suddenly transform into anti-aging transhumanists.[2] But this was exactly the state of AI safety discourse a few months ago, when a few experts cared deeply about x-risk but public discussion was an incoherent swirl of mere datacenter skepticism and it seemed crazy that “notkilleveryoneism” would go mainstream any time soon.
Historically, the strength of “e/acc x longevity” has been limited by the fact that most people smart enough to grasp the allure of this idea (eg AI lab CEOs, rationalists, etc) have also been persuaded by the argument for AI risk, and thus for caution despite the allure of e/acc x longevity. But a future preference cascade among the mass public (cross-pressured by anxieties over superintelligent AI and bioweapons, versus enthusiasm over medical progress) might not play out the same way. So it seems to me that the combination of e/acc ideas, and a mass public that for the first time truly believes in imminent cancer cures, aging treatments and other medical miracles, could end up going in some pretty crazy directions.
IMO the (small, very inside-baseball) controversy over Nick Bostrom’s paper “Optimal Timing for Superintelligence” reflects AI safetyists’ understanding of this dynamic. “Don’t speak too loudly about this very selfish, dangerous idea!” people seem to be saying. “If anyone else hears, it might set off the latent anti-aging preference cascade!” Personally, I don’t think that analytic-philosophy essays (or, for that matter, LessWrong posts like this one) are going to be the match that starts this particular fire. But the kindling is building up.
Of course I’m not saying we should oppose medical cures!
I’m in a weird position here, I guess. I’m not saying we shouldn’t try to use AI to develop some medical cures. Medical cures are great! (Other things equal, of course. Where “other things” includes x-risk from both reckless AI scaling and from enabling easier bioweapons development.) Even Eliezer Yudkowsky thinks we should try to go for the medical cures, if we can do so in a responsible way: in his Time open letter of 2023, he says “The moratorium on new large training runs needs to be indefinite and worldwide... If I had infinite freedom to write laws, I might carve out a single exception for AIs being trained solely to solve problems in biology and biotechnology, not trained on text from the internet, and not to the level where they start talking or planning; but if that was remotely complicating the issue I would immediately jettison that proposal and say to just shut it all down.” Even as he also says that it would naturally be “incredibly selfish and evil” to roll the dice on misaligned AI as described in Bostrom’s “Optimal Timing” paper.
So then what SHOULD we do?
idk! Here are some thoughts for AI safety groups and funders:
Beyond that, I think AI safety folks might also want to establish some anti-aging bona fides well ahead of time. Some ideas for how to do that:
All the above stuff, taken together, sounds like I’m pitching a maximalist agenda for basically transforming EA/rationality into a longevity-focused movement, which would be absurd. I’m instead just saying that we should consider doing some of this stuff, a little bit more than we already are.
The overarching idea here is that AI safety should be ready to present an alternative vision of radical longevity progress – one that isn’t reliant on breakneck general AI scaling. This plan doesn’t necessarily need to explain how it’ll actually OUTPERFORM a plan of breakneck scaling at achieving medical progress (which might be unrealistic). But during a preference cascade, we’ll need to address the issue in some way. (Just like how it behooves Republicans to have some kind of notional plan even on issues that are unfavorable for them, like healthcare. And similarly behooves Democrats to make some sorts of commitments on immigration, even if they’d mostly rather people not talk about the issue since it’s unfavorable for them.) Lest we end up looking like the e/accs do in the current AI safety moment – losing the argument in the eyes of the public, reduced to weak rebuttals, weird ad hominems, and transparent attempts to change the subject.
The “support for FDA reform and scientific research” stuff earlier is important not just so we have a track record to point to, but also because after a longevity preference cascade, there will be strong public demand for taking action on aging. People will want a roadmap, they’ll want big initiatives being launched and funded, and so on. So, EA / rationalist groups should perhaps fund some of these things in small-scale now, with a view to being able to impressively scale them up later when the demand arises. Cryonics also seems particularly relevant here as a shovel-ready plan that acts as a “Plan B” alternative vision for longevity. Ideally the idea of a crash program to improve the science and scale up the practice of cryonics might be capable of relieving some of the desperate urgency and time pressure that the public might feel about the problem of aging after a preference cascade, and would answer some of the newfound political will to immediately “do something about aging” in a way that doesn’t involve further scaling AI.
What about the labs themselves; what should they do?
Final thought: in all the above discussion I’ve made it seem like it’s important to have all these plans ready (to channel the energy unlocked by a longevity preference cascade) for the purpose of performing PR judo and deftly deflecting a movement that might damage the cause of AI safety. But obviously there’s another good reason to have such plans ready in advance: properly directed, a longevity preference cascade could be really great! It would be amazing to use all that energy for good, using it to fuel more ambitious medical progress in a way that isn’t existentially dangerous. So, for that basic object-level reason, it would clearly be great to have some shovel-ready plans scoped out for rapidly scaling up ambitious medical research.
Longevity fans will complain that progress on aging is actually more tractable than the public thinks. I agree! But look at the history of the field: resveratrol? senescent cells? some mouse trials? None of this stuff has produced a genuine miracle for human life expectancy; and meanwhile there have been lots of weird fads and hype. It is very understandable that a distant general public (who has a harder time sifting fads from more promising science) perceives this all as snake oil, or at best overly-optimistic empty promises in the vein of “commercial fusion power is just 10 years away”.
I’m not sure what exactly how transhumanist a longevity cascade would end up looking like. Would it mostly be focused on enthusiasm for curing normal diseases, like cancer or Alzheimer, without touching the still-taboo idea of anti-aging? Could it possibly get crazier than anti-aging, and into other transhumanist goods, like BCIs or etc? Perhaps most importantly, how much would people connect the dots between medical progress and AI policy? Maybe everyone would get excited about medical breakthroughs but this would just translate happily towards pressure for more research funding, FDA reform, etc, without politically connecting to any AI issues? Or maybe people would go full on "worship the coming machine god" at the first whiff of some new AI-developed antibiotics for treating hospital staph infections, or whatever. Who can say!