The escape velocity claim, and the one paper that argues back

Longevity escape velocity is a real idea with a precise definition, and it is worth stating before the argument starts. Aubrey de Grey, who coined the term, defines it as the point at which medical progress adds more than a year of remaining life expectancy for each year that passes. Cross that line and, in principle, you stop losing ground.

That is a testable claim about a rate. It is not a claim about immortality, and the people who make it carefully are not making one.

Who has put a date on it

Ray Kurzweil is the most aggressive. In The Singularity Is Nearer and in interviews since, he places escape velocity around 2030 for people who are diligent and well informed, on the argument that AI is doing to biology what it did to computing.

Aubrey de Grey is the most conservative of the optimists, which is notable given that it is his term. He puts it at roughly a 50% likelihood in the mid 2030s, and his foundation’s current work is Robust Mouse Rejuvenation: doubling the remaining lifespan of a middle-aged mouse. He is explicit that human results follow mouse results, not the other way round.

Peter Diamandis puts the window at 10 to 15 years, and frames his own regimen as staying healthy enough to be there when it arrives.

David Sinclair argues from the information theory of aging: that the cell retains a usable backup of its youthful state, so aging is a software problem rather than a hardware one. He has said control and reversal of biological age is roughly a decade out.

The one who declines the frame

Peter Attia is frequently grouped with the above and does not belong there. In Outlive his subject is healthspan and what he calls the marginal decade, the years from roughly 75 to 85 where function is either retained or lost. He uses “escape velocity” too, but to mean escaping functional decline, not escaping aging. The distinction is not pedantry. One is a claim about when medicine outruns mortality; the other is a claim about how well you will be able to move at 80. Only the second has a body of evidence behind it today.

The paper that argues back

In October 2024, Olshansky and colleagues published an analysis in Nature Aging covering three decades of mortality in the nine longest-lived populations plus the United States. Their finding: improvement in life expectancy has been decelerating since 1990, and without interventions that slow aging itself, radical life extension in this century is implausible.

You can see it in the American record without taking anyone’s word for it. US life expectancy at birth ran from 47.3 years in 1900 to 68.2 by 1950, a gain of nearly 21 years in five decades. From 2010 to 2019 it moved from 78.7 to 78.8. A decade, one tenth of one year.

Olshansky’s own framing is worth keeping, because it is not defeatist: a glass ceiling, not a brick wall. There is substantial room left in reducing risk factors. What is implausible is the escape.

The asymmetry

Here is what we think matters most, and it is structural rather than scientific.

Of everything above, the only peer-reviewed entry is the one that says no. Kurzweil’s claim is in a trade book. De Grey’s is a stated probability from a foundation he runs. Diamandis writes on Medium. Sinclair’s is a book thesis plus interviews. These are not disreputable venues, and none of these people is a crank. But a prediction in a book and a refereed analysis of mortality data across ten populations are different kinds of object, and a reader deserves to be told which is which.

The asymmetry cuts both ways, in fairness. Absence of a peer-reviewed case for escape velocity is not evidence against it; you cannot referee a forecast. The honest summary is that the optimistic case rests on extrapolating a technology curve, the pessimistic case rests on extrapolating a mortality curve, and only the second one has a dataset underneath it.

One asterisk worth stating plainly

Enthusiasm in this field has occasionally outrun the evidence, including among its most credentialed voices. In March 2024 Sinclair resigned as president of the Academy for Health and Lifespan Research after criticism from other researchers over an age-reversal claim made in a press release for a dog supplement from a company he co-founded. The quote was subsequently corrected.

We raise it for one reason. The field’s failure mode is not fraud. It is a credible scientist describing a real result in language that outruns it, and the correction never travelling as far as the claim. That is the failure mode this publication is built to avoid, and pretending it only afflicts other people would be its own version of the problem.

What we do with it

We do not have a date, and we are suspicious of anyone whose confidence exceeds their dataset in either direction.

What is not in dispute is that the interventions with measured mortality effects are unglamorous and available now: cardiorespiratory fitness, not smoking, blood pressure, lipids. None of them require a breakthrough. All of them are being neglected by people waiting for one.

If escape velocity arrives, it will find you in whatever condition you kept yourself. That is the version of the argument we are willing to put our name to.