Living to 1000: an interview with Aubrey de Grey
In his recent talk, Aubrey de Grey argued that the SENS foundation’s life extension research should be the most important project for any effective altruist. I recommend listening to the talk before reading the interview below, but it’s not essential.
Similarly, more information on SENS’s anti-ageing research is available for them that want it.
On with the questions, then - as in the interview with Luke Muehlhauser I’ve separated them into a section for people interested in working in gerontology and a section for people interest in supporting it in other ways.
Questions for potential supporters
ZR: It seems like you gain interest from the same sorts of people who might be interested in SIAI, especially given that you’re an adviser to the group. But the mercenary question remains - given finite resources, which organisation should we prioritise? Here’s the key argument Luke Muehlhauser gave for SIAI:
In the next two centuries we will have a multitude of chances to destroy ourselves with powerful new technologies. AI looks like it may be one of the first existential threats to be created. But unlike the others, doing AI right can actually prevent the other existential risks from happening. There is no more important thing humanity can do.
Do you think friendly AI will not be as good at fixing the world’s problems as longer-lived humans, do you just rate the latter as much more likely to happen soon, or is there some other reason you’d advocate the SENS Foundation over them?
AdG: I’ve never been convinced that the elimination of humans all in one go is all that much worse than the elimination of humans on a steady schedule of 150,000 per day. The only difference seems to be that in the latter case there get to be more people in the long run - but so what? One thing that does make defeating aging particularly important is that aging doesn’t just kill people, it generally puts them through an extended period of disease and disability and general misery beforehand.
However, I strongly support the goal of AI, because success in that area - if, as Luke says, it’s done right - will indeed solve all other technological problems, including the development of medicine to defeat aging. But yes, I do think that there will probably be be quite a lot of progress in combating aging before we get to proper AI, so I think the approach which will save the most lives in the long run is to pursue both approaches as aggressively as possible.
ZR: How confident are you about the success of your program? I’ll paraphrase a couple of off-the-cuff remarks on the subject you made in your 80,000 Hours talk: ‘We’ll get to robust human rejuvenation - within the next 25 years with 50% probability.’ ‘I’d give a 10% chance we won’t get there for 100 years.’ Would you modify these now you’re not on the spot? What exactly do you mean by robust human rejuvenation? Is that the same thing as reaching what you call the ageing escape velocity? (for the benefit of readers: the point at which progress in our ability to extend our lifespans surpasses the rate at which we age, effectively making us immune to age-related death)
AdG: I still stand by both those statements, but please note that I always add the caveat that the former depends on adequate funding, especially in the coming decade. I define “robust human rejuvenation” as the addition of 30 years of extra healthy life to those who are already 60 when the therapies are first given. Longevity escape velocity is different - it’s the postponement of aging faster than time is passing, which results from continued progress in improving the comprehensiveness of the therapies. The moment at which we reach LEV, which we call the Methuselarity (and we’re pretty sure there will be inly one such moment, i.e. that once we exceed LEV we will never fall below it again), will probably occur at around the same time when we achieve RHR - maybe a little sooner, maybe a little later.
ZR: How much of your optimism is focused around the success of the research program you described to us? Suppose that were to fail entirely, would you still predict a reasonable chance of the same positive outcomes?
AdG: I’m pretty certain that the SENS approach, i.e. regenerative medicine against aging, is far more promising than any other. However, it’s certainly quite possible that details of what particular therapies will be involved will change as the research moves forward.
ZR: How (much) has your confidence changed in say the last 10 years of your research, both in the probability of your own success and any alternative programmes?
AdG: Very little. Things have gone pretty much as fast as I’d have predicted given the level of available funding.
ZR: Come to that, can I ask you for your views on other gerontological research programmes? Are there any you’d be willing to single out as being particularly (un)worthy? In particular, if yours were to fail (or succeed) tomorrow such that no further progress were possible, is there somewhere else you’d advocate putting money?
AdG: Most biogerontology research is useful one way or another, because it improves our understanding of aging. That can lead to the identification of new avenues for intervention, and also to refinements of the SENS approach.
ZR: Let me put you in the crosshairs for a bit: Gwern from Less Wrong has written a rather gloomy prognosis of our near technological future. If his view is right, does it imply that the chance of reaching what the ageing escape velocity soon would be lower than you predict? Or perhaps even that we could temporarily reach it, but that we might get a bunch of quick gains from technology and then start dying of old age somewhat later than we currently do?
AdG: I don’t think current trends in rates of increase in longevity really tell us anything about how soon LEV will arrive, no.
ZR: You’ve drawn some aggressive criticism for the methods you propose for curing aging. The most direct engagement I found with the science was the MIT Technology Review’s Life Extension Pseudoscience and the SENS Plan. Have you written anything extensive in response to these (or comments like them) that we can read online?
AdG: Yes, of course I have! The piece was not written by MIT TR, but by a few gerontologists in submission to a prize competition that they organised with my cooperation, and the rules were that I would write a rebuttal of each submission and the two would be evaluated by a panel of five neutral experts (which included Craig Venter). Naturally, no entry (including this one) succeeded in persuading the panel. My rebuttal is here:
and the report of the whole prize process is here:
which also has links to the other entries and my corresponding rebuttals.
ZR: I’ll let you down from the gallows now. Do your calculations of economic benefit take into account the cost of turning a baby into a productive adult, and the fact that longer life improves the ratio?
AdG: They can include that, yes, but the main factors to consider are the results of increased health. Having no sick people to look after will make society astronomically wealthier.
ZR: Back to matters financial… How much funding would your programme actually need to complete its goals? Paraphrasing your talk again, you guesstimated ‘$50,000,000 allocated appropriately would be “enough for the next while”’. Can you be a bit specific about what ‘the next while’ means?
AdG: I think I said $50M per year. 20M/year would be a good start, and I think above $100M/year we’d be reaching diminishing returns. This is tiny compared to (for instance) the NIH budget, but that’s because it’s only the budget for the next several years, when the key work will be on mice rather than humans.
ZR: Perhaps more importantly, do you have a sense of how valuable funding on the margins is – eg if someone were thinking of giving $25 in the near future, can you guesstimate exactly what that donation would gain? I’ll give you the same impossible challenge I gave Luke Muehlhauser - can you try to compare it to the (already very rough) one DALY’s worth of expected benefit from giving it to Against Malaria Foundation? And can you give an idea of what that $25 would be used for?
AdG: It’s pretty clear that a dollar makes more difference when spent on early-stage biomedical research (which is what we’d spend it on) than on the delivery of existing therapies.
ZR: Can I press you for a guesstimate of expected value? Such data might be very relevant if one is considering eg soliciting donations from aquaintances who’re significantly less likely to donate to something that sounds so much less tangible than giving money to the poor.
AdG: Well, the thing is that there’s no new life expectancy to replace the old one with. If you save a kid from malaria you change their life expectancy by a finite amount, but here it’s an indefinite amount. So I guess the answer has to be infinite.
ZR: You addressed in the talk personal identity objections to indefinite life (they seem irrelevant to me, but some people think they’re important…), and your response was that we generally get along from day to do without worrying about it - that if we think we’ll be the same person tomorrow, we should assume that we’ll be the person in two days time, etc. But this doesn’t strictly follow - one could see identity as a question of degree, such that after a while you really are talking about a different person. If so then it might still make sense to say that some minimum amount of life is rationally desirable, but then it doesn’t make sense to desire it any longer. Can you offer an argument for people who believe that that minimum amount of desirable life is greater than a day, but less than our typical lifespans?
AdG: Sure. The practical implication of the day-at-a-time argument is not that we should care as much about our distant-future self as about tomorrow-self, but rather that however long we live we will always care about our next-day self. So developing medicine to stay healthy for as long as we live is just forward planning, in terms of our self-regard as well as in terms of our wish to stay healthy.
ZR: You’re obviously a techno-optimist, and your concept of the longevity escape velocity explicitly requires constantly improving technology so that we can stay on the right side of senescence. But given the imminent peaking of various resources, isn’t it quite likely that our functional technology will decline at various points in the 21st century? If so, how much will that threaten your estimates of expected gain?
AdG: If you look at the numbers, it’s vanishingly unlikely that that will make any difference. This is especially so when you notice that LEV actually slows down as time goes by - i.e., the further we suppress the damage of aging, the less rapidly we need to suppress it further in order to maintain LEV.
ZR: How well do you expect the treatment, assuming it success elsewhere, to work on degradation of the brain? Surely that would be much harder to maintain than the rest of the body?
AdG: No, not much harder (if at all). The brain is far more complex, so its operation is far harder to understand, but SENS doesn’t require understanding of operation - it only requires characterisation of the changes to molecular and cellular composition. If we restore that composition to that of a young adult, we’re done.
ZR: Given the relative modesty of the money you described above, it seems amazing you don’t get more funding from - how to put this - the less benevolent members of the super-rich community (I gather you’ve had some generous interest from Peter Thiel). If your research is going to bear fruit, it would be insane of the wealthier member of societies not to funnel large amounts of their income into it, and encourage their peers to do the same. The same people are far less likely to support purely altruistic causes like AMF, so it seems like arguments for 80,000 Hours members supporting SENS suffer from replaceability effect which underlies much of the 80,000 Hours philosophy. I suppose there are two questions here - why you haven’t got more funding from such sources - and whether a professional philanthropist or similar who agreed with your organisation would do better to pass the word around to his less altruistic colleagues but then get on with funding more selfless goals.
AdG: There are many reasons people have for not supporting this work financially, and by and large they are not rational reasons - they are the same sort of reasons why people don’t support it verbally. Some people support it verbally but not financially, and that’s typically because their wives are against it, or something like that. Grim but true.
ZR: If people do want to get involved, are there any non-financial ways in which people can contribute, particularly at low opportunity cost? For example computer cycles to a relevant research program, crowd-sourcing opportunities they can distract their less productive friends with, etc?
Careers in gerontology
ZR: Easy question to kick off: please talk about yourself :) More specifically, can you describe (in as much detail as you like) a typical working week? How many hours would you put in, what proportion of them would be spent on work you find engaging and what on admin/other chores? More importantly, what, physically do you do in what proportion from day to day? Can you give a sense of the highs and lows of the job? What about others from your organisation and others like it? (I basically want to give people thinking of actually working in gerontology as vivid a possible a description of what their life would be like)
AdG: Well, my life is very far from typical for a career in gerontology! As the global spearhead (and figurehead, I suppose) of the crusade to bring aging under complete medical control, I spend an large amount of my time doing interviews, which most people do none of, and lectures, which most people do not many of and only to other scientists rather than to general audiences. Also, when I’m not doing outreach I mainly spend my time advising and connecting other scientists - I don’t do any bench science myself. Ifind all of these activities engaging and fulfilling - it’s an immense privilege to have the chance to spend all my time doing the thing I most want to do it the whole world, namely to hasten the solution of humanity’s single biggest problem.
The work of a typical gerontologist - and here I refer to biogerontologists, who study the biology of aging, rather than clinical gerontologists (aka geriatricians) or social gerontologists - is actually not materially different from that of any other bench scientist. You do experiments, you do them again and again until they work, you talk about them, you write them up for publication, you ask the government for money for the next experiment - you get the idea. But of course you also get at least a little of what I get: the satisfaction of knowing that you’re contributing to solving humanity’s worst problem.
ZR: What about a slightly atypical week? What sort of events of note happen rarely, but reliably?
AdG: Days off!
ZR: Now let’s suppose you’ve won some readers over. What attributes would someone thinking about going into gerontological research ideally have? How much competition is there?
AdG: Competition is certainly not the issue. Excessive competitiveness is actually a major handicap in science, because it encourages one to study things that other people are already studying, and thus it minimises one’s chance of actually making much of a difference. I’d say the main attributes are determination and patience, which are vital for any scientist, and humanitarian drive, which gives the translational focus that anyone working on really hard technlogical (including medical) research needs.
ZR: I was thinking more of the type of competition that would make it hard to get work in the field at all - if someone studied science with this kind of career in mind, roughly how outstanding would they need to be to expect to find work in it? How much would just studying the most relevant fields compensate for a lack of natural ability if they were prepared to work hard - and what would those fields be?
AdG: There’s certainly not that much competition. Aging is so complicated that the best way to get to be good at studying it is to get a very general biology education. However, for the same reason, people with specialist expertise in pretty much any subfield of biology can also do research that’s relevant to combating aging.
ZR: I’ll end with a quick plug for your book, _Ending Aging. Given the rate of progress in the field, though, can you predict how much longer it will be worth reading for? Any plans for an updated edition?_
AdG: It’s still worth reading! No major changes have occurred to SENS - mostly what has happened is just progress in the very directions that we discussed and predicted in the book.
ZR: Lastly can you recommend any further reading for anyone wanting to really immerse themselves in the scholarship on (bio)gerontology?
AdG: Once you’ve read Ending Aging, the only real place to go is the sens.org website, where we have a regular blog on the key advances in the relevant fields.
Aubrey has agreed to field a few follow-up questions if we have any. Let me know, and I’ll pass them along to him.