Live·Open questions in longevity research
All news
AI in medicineLongevity researchBiotech funding

Psychologist Geoffrey Miller would rather die of aging than impose a ten-percent risk of extinction from AI on his children and humanity

19 November 2026· 261119011

Psychologist Geoffrey Miller would rather die of aging than impose a ten-percent risk of extinction from AI on his children and humanity

Programmer Eric Raymond admitted that he supports accelerating AI development partly because he does not want to die. Aging researcher Kamil Pabis and Vladimir Shakirov, author of a wiki on overcoming aging, challenged Miller: AI offers a chance to save both children and parents from aging.

On November 15, Miller replied to Raymond:

“I would rather die of aging than impose a ten-percent chance of extinction on my children and humanity.”

Miller sees another path: accelerate longevity research without accelerating the development of superintelligence, AI that is far more intelligent than humans and, in Miller's view, cannot be aligned with human goals. He estimates that less than 5 billion dollars a year goes into longevity research, while 200–300 billion goes into such AI.

Pabis objected: “But your children will die of aging too. It is a terrible death that superintelligence could prevent if we get this right.” He then pointed to the work of philosopher Nick Bostrom: given an appreciable chance of achieving aligned superintelligence, “simple math” shows the benefit of creating it.

In a 2026 paper, Bostrom starts with these assumptions: without superintelligence, people alive today have about 40 years left to live; if superintelligence succeeds and keeps mortality at the level of a healthy twenty-year-old, they have about 1400. Even a 3% chance of success yields an average of 42 years rather than 40, so in Bostrom's model, superintelligence increases life expectancy as long as the risk of it killing humanity remains below 97%. Miller's figure of 10% is far below that threshold.

Miller replied to Pabis that everyone will die of aging if money continues to be wasted on AI instead of longevity research. He called the idea that “AI will solve aging” “an absurd, ignorant psychological operation by people who know less than nothing about biomedical research.” On November 2, Miller had already written that language models were trained on the entire internet, while human biology has no comparable body of data that would allow superintelligence to dramatically accelerate longevity research.

Pabis replied that he believes AI will make an enormous contribution, though he is unsure about the phrase “solve aging,” and added:

“The AI industry's money will never go into aging research. It comes from an entirely different pool of funding.”

On November 16, Miller explained how money from the AI industry could reach aging research. Successful initial public offerings (IPO) could bring enormous financial gains to people in the industry, many of whom have pledged to give a substantial share of their wealth to charity. Longevity could be one of the causes they support if they recognize that AI will not solve aging on its own: what is needed is “direct support for biomedical longevity research.”

For example, according to the AI company Anthropic's IPO filing, each of its seven cofounders has pledged to donate 80% of their personal shares to charity. The filing describes the pledges as voluntary and revocable, with the timing, amounts, and recipients unspecified. On November 18, the company set its offering price at 83 dollars per share, bringing the cofounders' combined wealth to more than 140 billion dollars, according to Bloomberg's estimate.

On November 16, Shakirov expressed doubt that funding for longevity research could be increased tenfold or even doubled: “Many have fought for this and lost. Now, with AI, we at least have a good chance.” He then asked whether Miller's children would consider it fair for their parents to have to die to reduce their own chance of dying by 10%. Some parents, Shakirov added, will probably die unless AI offers a chance to save them.

Originally published on Telegram by Ukhvat NewsView on Telegram
Sources
#ai-extinction-risk#superintelligence#aging#longevity-funding#anthropic-ipo#ai-alignment