The Methuselah Foundation. Science for the End of Aging.

Questions and Answers with Aubrey

Note: even though there are links to separate questions below, I recommend reading this page straight through from start to finish.

1. Get to the point. How long do you think I could live?

   The answer to that question is of course very speculative, because it depends enormously on (a) how fast the research goes and (b) how long one is naturally predisposed to live even without these advances. I've said in the past that the first person to live to 1000 was probably born by 1945, and the first person to live to 150 probably by 1935, but those are people who would naturally live to 110. Conversely, lots of people are not predisposed to live beyond 70, and they have very little chance of benefiting from these therapies if they are already 50 or older.

   Because of this uncertainty, I find it preferable (both for myself and when discussing this with others) to think not so much in terms of how long I can expect to live, but instead to focus on three other things:

   1. Whatever one's probability may be of living to a given age, one will increase that probability by acting to hasten the research and by looking after one's own health.
   2. Younger people obviously have a better chance of benefiting, since they'll be more likely to be still alive when the therapies arrive, so if you don't think your chances are very good, focus on the fact that your kids' chances will be improved if we accelerate this work.
   3. Globally, 100,000 people die of aging every day. Thus, if we can bring the defeat of aging even one day closer by our actions today, we'll save 100,000 lives. If we advance it by a year, we save 35 million lives. These are pretty staggering numbers, and pretty good reasons to get our act together without delay.
2. Why do you engender possibly unwarranted optimism about timescales?

   That's more of a criticism of how I'm going about making SENS a reality, so my answer is on that page.

3. When will we prove that human aging can really be repaired?

   This is the first major SENS milestone, and I believe it will be achieved with laboratory mice. I also consider that it will be the point at which society becomes convinced that curing aging is very urgent, and that it will kick-start a genuine "War on Aging".

   My estimate for the time until this milestone is reached, if there is adequate funding, is ten years from now; almost certainly not as soon as seven years, but very likely to be less than 20 years. If funding is sluggish this could be doubled.

   The degree of control that I consider sufficient is the ability to take a cohort of mice of a strain whose normal life expectancy is three years, do nothing to them until they are two years old, and get them to live an average of three more years, i.e. tripling their remaining life expectancy. I often call this "Robust Mouse Rejuvenation" or RMR.

   Detailed justification for this prediction can be found in my publications, which are here.

   The main reason this milestone matters so much is that it will be the trigger for enormous social change. Changes of life choices will not await the arrival of human life extension; they will occur as soon as that eventuality becomes widely anticipated.

4. When will we have the first human rejuvenation therapies?

   This is the second major SENS milestone, and it can reasonably be defined as the arrival of therapies that confer a postponement and repair of human aging proportional to that described for mice in milestone 1, i.e. a tripling of our remaining life expectancy with therapies initiated on our late fifties or so. Inevitably I call this "Robust Human Rejuvenation" or RHR.

   My estimate for the time until this milestone is reached, starting from the time that the mouse target is achieved, is 15 years; almost certainly not as soon as five years, and could be as much as 100 years. Note that this time I make no caveats about funding, because I think it is inconceivable that shortage of funds will be allowed to slow down this work once milestone 1 is achieved.

   The best I can do by way of justifying this is to consider the scenario in which we reach the mouse milestone without having made any progress whatsoever on the techniques that would be needed to translate it to humans. The main requirement at that stage will be extremely safe and effective gene therapy, both of the insertional variety (putting new genes into chromosomes) and the replacement ("gene targeting") variety (changing an existing sequence). Several approaches to these problems are already the focus of considerable ongoing research, so I feel that a 15-year timeframe is a reasonable one, at least once funding is increased by the factor that can be expected when a "war on aging" has been declared. However, we could of course be as wrong about aging as Nixon was in 1971 about cancer, which is why my upper confidence limit is as large as 100 years.

5. How long can I/my children expect to live?

   Clearly that depends how old you/they are, but I feel able to give a pretty definite prediction relative to the previous milestone. I claim that the average age of death of those born in wealthy nations no sooner than 40 years before the achievement of milestone 2 will exceed 5000 years.

   Of all my predictions, this is the one that most thoroughly stuns most people -- and, I claim, for the least justified reasons. My logic here is pretty simple:

   1. When we reach milestone 2, those with access to the relevant therapies will have an absolutely non-increasing risk of death per unit time -- they will not age. This is because we will be identifying, characterising and solving aspects of aging that appear at progressively later ages, faster than they progress to a life-threatening state. We have no idea at present what we will need to do to keep 200-year-olds hale and hearty, but that's OK, because we won't need that information for at least another 100 years. If we just pay attention to things that begin to appear in 180-year-olds as soon as we have any, as well as in 80-year-old chimpanzees as soon as we have them, and given the amount of effort we'll be putting in, our chances of perpetually keeping one step ahead of the problem are very good.
   2. At present, the risk of death per unit time that Westerners experience in their early teens is such that if it were maintained indefinitely we would live to around 1000 years on average. (This calculation has been done many times with different data and some people get 700, some 1200; 1000 is a fair consensus.)
   3. My expectation is that our risk-aversion will rise sharply if we perceive our lifespan as indefinite, so that this 1000-year life expectancy will be extended by a modest factor; a factor of five is my conservative guess. This of course relies on our risk-aversion being ubiquitous -- applying to the willingness to go to war, the effort to subvert new infectious diseases, etc, as well as our attitude to mere accidents -- but I see no reason why that should not be the case.

   A lot of people, when they read this, just stop listening because it's beyond their imagination -- they get the conceptual "bends". Well, if that's you, please try again -- and if you'd prefer a more mathematical treatment, focusing not so much on how we will achieve these very low mortality rates but on what it will mean statistically, read on.

6. Will we ever stop dying involuntarily of old age (at whatever age)?

   You've doubtless guessed my answer by now: not exactly stop, but it'll become a very rare occurrence. I've given a short name to this concept recently: I say that once we get to be improving our rejuvenation therapies faster than the improvements are needed, as described in the previous section, we'll have exceeded "life extension escape velocity". Here's what I mean, in detail.

   1. Actuarial, or Population, escape velocity

      The first thing to consider is the opposing influences of aging and biomedical progress on one's risk of death in the near future (say the next year) as one gets older. In every human population, among people aged between about 40 and 85 (the age range in which most people die in developed nations), the proportion of people of a given age who die in a given year is about 10% greater than the proportion of people a year younger who die that same year and 10% less than that of people a year older. This exponential relationship between age and mortality rate was noticed about 180 years ago and is phenomenally constant across all human populations ever examined -- both in the sense that there is an exponential relationship, and in the exponent of that relationship.

      What's not constant, however, is the "intercept" of this curve -- the absolute proportion of people of age N who die in year X. In particular, there is a good deal of progress over time: people in most developed countries have a much lower mortality rate than people in the same country, at the same age, 50 years ago. This is mainly as a result of medical advances.

      Where this gets interesting is when we consider how this affects real people. Individuals aged N at the beginning of year X are not also aged N+1, N+2 etc. at the beginning of that same year: they're aged N+1 at the beginning of year X+1, and so on. So this means that even though the ratio of N-year-old to N+1-year-old mortality in year X is 1.1, and it is still 1.1 in year X+1, the ratio of N+1-year-old mortality in year X+1 to N-year-old mortality in year X is less than 1.1. And that ratio is what matters to real people.

      You can probably see by now where I'm going with this. How fast does biomedical progress need to be in order to stop people from having a progressively increasing risk of death in the coming year as they get older? Obviously, it has to be fast enough that the risk of death of people aged N in year X+1 is 10% less than that of people aged N in year X. Now, here's the crux: that's not ridiculously fast. If we look back at the past century or so, we find that mortality rates at certain ages did indeed decrease rapidly -- not quite 10% per year, but 3% was seen and 2% was seen often. So, the 10% per year rate of progress is what I've started to call "escape velocity". It's defined on the population, looking at mortality rates, so when being precise I call it "population escape velocity" or "actuarial escape velocity". Of course 10% isn't a ceiling, and we can just as easily anticipate an actual rate of mortality rate decline of 15% or 20%, which would mean that someone's risk of death in the coming year was actually falling as they were getting older. The only limit to this is when aging ceases to be the main cause of death for most people at the age in question.

      Now, there is a potential problem with this: inevitably (though very probably not forever -- see below) there will be new things going wrong with us as we attain unprecedented ages, and we'll only have a certain amount of time to characterise and work out how to repair these new things before they start killing people. This might be a real problem -- if it weren't for monkeys. Monkeys save the day, because they are (a) fabulously similar to us, (b) unable to speak, which means that given sufficient biomedical imperative we don't mind putting their lives at risk, and (c) prone to age at least twice as fast as us. So we don't yet know what 200-year-old humans will die of, but we don't need to until we have some, and by then we will for some time have had 100-year-old monkeys that we've been treating in just the same way that we treat ourselves (bad diet, no exercise, but all the life extension technology that we use on ourselves). And because of (a), those monkeys will have exhibited the symptoms that 200-year-old humans exhibit, so we'll have been working for a long time on fixing them (in monkeys), so by the time they occur in humans we will know how to fix them well. By the time we reach 300, the same will be true by an even greater lead-time. This all relies on at least some of the monkeys forever getting the same problems that we get but at under half the age, but that's a pretty safe assumption.

   2. Individual escape velocity

      Jay Fox, one of 2004's more conspicuously bright and energetic arrivals on the life extension scene, has written a very thoughtful analysis of the escape velocity concept. In it he pays particular attention to the important point that escape velocity is defined on populations in terms of death rates, and thus that it says nothing about how frail people are. This is always an important thing to reassure people on, so here goes. The problem is that biomedical progress in the future, just as in the past, may reduce people's rate of decline from health into frailty and also their rate of decline from frailty to death. So in principle it is possible that escape velocity as defined above -- i.e., population escape velocity -- could be achieved by rapid and sustained improvements in our ability to keep frailer and frailer people alive. That isn't the point at all, of course, and luckily there is no reason to fear such a scenario, because it is a great deal harder to keep someone alive in a frail state than to stop them becoming frail in the first place or even to restore a frail person to robust vitality. This is demonstrated by, for example the rate of progress in longevity in the past 20 years: the life expectancy of most nations has risen by a couple of years, but so has the "healthy life expectancy". In other words, there has been very little progress in keeping frail people alive. There's no reason to expect the future to be any different in that respect. But for the sake of precision, it makes sense to define a second type of escape velocity, individual escape velocity. This is the point at which we are beating back aging fast enough that people's vigour, not just their likelihood of death, is being maintained as they get older. Vigour is harder to define and to measure precisely, of course, but we now have pretty established objective measures of it.

7. Will we ever have therapies that stop us dying of old age at whatever age? (not the same question!)

   One can define other, stronger versions of escape velocity. Even though we're extremely complex machines, our complexity is finite, so we can realistically anticipate that there will come a time (albeit only in the very distant future, many centuries from now) when we can just stop improving our rejuvenation therapies, because we're already fixing everything that naturally goes wrong with time. This is the situation we're in with vintage cars: it's a lot of work to keep a car on the road past its 70th birthday, but if you're willing to put in that work indefinitely then you'll be able to keep it going indefinitely -- the number of things you need to fix and the frequency with which you need to fix them don't rise any further. That's a strengthened version of individual escape velocity.

   If you've read this page from the top, you will now understand my frustration when people persist in ridiculing me for saying that we will in a few decades have therapies that will let us live to 1000, when I don't say that at all.

8. Will we ever make ourselves truly non-aging?

   An even more strengthened one would be to reach a point where not only didn't our therapies need to be further improved as we get older, they didn't even need to be repeated, because we'd eliminated aging from our bodies entirely. It's conceivable that this could be achieved by some sort of transformation of our bodies into purely electronic form, with no actual moving parts, but it's perfectly certain that we would never achieve it with bodies made out of meat. Unfortunately it could be said that this most extreme, least realistic version of individual escape velocity is actually the one with the best analogy to actual escape velocity -- but hey, analogies are never perfect.

9. Will we all die some time?

   An interesting type of escape velocity concerns the risk of death from all causes, aging-related or otherwise, and is a strengthened form of actuarial escape velocity. If we achieve actuarial escape velocity as defined earlier and exceed it by some way for a long time, we will get to a point where essentially no one is dying of old age and indeed no one is even frail. But of course there will still be death from all the familiar age-independent causes. If we suppose that this risk is a constant -- not only the same in year X for people of any age, but also the same in all years from X onward for people of any age -- then clearly we all have a "half-life", like radioactive materials, which means we all die eventually, just as all the atoms in such materials decay eventually. If we all had the mortality rate of present-day Western 11-year-olds, our half-life would be about 1000 years.

   But it is not actually all that realistic to suppose that our risk of death from age-independent causes will be constant indefinitely, because we will probably not be very keen on that sort of death either and we'll be forever working to improve our technological ability to avoid it. It is dangerous to presume that there are any real limits to what we may be able to achieve in this regard in the extremely distant future. Thus, it's worth considering what our life expectancy would be if we had a continuous -- let's assume constant, for sake of simplicity -- and indefinite improvement in our ability to avoid death from any cause. Suppose your chance of dying in the next 1000 years is 0.5, but if you survive that 1000 years then your chance of dying in the next 1000 years is only 0.25, and if you make it through that 1000 years then the chance of death in the next 1000 is only 0.125, and so on. It turns out that this sequence does not have a zero asymptote -- you have a roughly 28% chance of living literally forever, genuinely never dying at all, even though your chance of dying in any given millennium is always non-zero. If the chance of death diminished by a smaller factor than 2 in each initial half-life -- say it went 0.5, 0.3, 0.18 etc. -- then of course the chance of never dying would not be as high as 28%, and if it diminished by a greater factor than 2 the chance of never dying would be higher than 28%. But if it's greater than zero, i.e. you have some chance of never dying, then it turns out that that means your life expectancy (the time you have a 50% chance of living) is actually infinite, because the people whose lifespan really is infinite outweigh the ones whose lifespan is finite. That's a pretty strong version of escape velocity!

   You might be thinking the above scenario seems unrealistic, because surely some halvings of risk are going to be harder than others and the scheme only works if we halve the risk every single millennium. Well, then let's look at a more pessimistic scenario, where the halvings of risk take progressively longer to achieve on average. As an example, suppose the halving time doubles each time. In other words, the risk of dying in the next 1000 years is still 1/2, but the risk of dying between 1000 and 3000 years from now is 1/4, the risk of dying between 3000 and 7000 years from now is 1/8, etc. What proportion of people live forever then? The answer is: just the same proportion, 28%, as in the previous example! This is obvious if you think about it, because the proportion of people who live forever is just the limit of the series (1/2) x (3/4) x (7/8) x .... and thus is independent of the lengths of time involved. The survival curves of these two examples do have different shapes -- for the first 5000 years or so one's risk of death in any particular millennium is lower in this second example than in the previous one, and thereafter it's higher -- but the two curves have exactly the same asymptote.

10. Will we ever make ourselves immortal?

    No. None of the above forms of postponement of aging and death -- not even the last one -- can correctly be described as "immortality". Immortality means inability to die, i.e. a certainty of never dying. Even in the last case above, there is always a non-zero probability of dying some time -- and indeed a non-zero probability of dying in any given year. So this last question has an easy answer: no, we will never make ourselves immortal.