It didn't come from either, it came from me, as a benchmark for what seems conservatively possible to achieve in the near-term, and for the size of impact necessary to be plausibly cost-competitive with other causes like global poverty. (If the same amount of funding yielded only one year's delay in the DALY burden of age-related diseases, I think that would make global poverty likely to be a "better buy.")

I believe there are larger effect sizes out there than metformin; metformin has a relatively small effect size on mice compared to other lifespan-modifying interventions, and the TAME trial chose metformin (as Barzilai admits) because it's extremely safe and well-studied, not because it's expected to be the best.

I agree with you; I don't think aging research would be cost-effective at a trillion dollars of total funding. I expect that's hugely more money than necessary.

Fixed, thanks.

Yes, I'm aware, and thanks for explicitly flagging that.

DALY estimates are obviously fraught, and I understand that they're not exact for any particular charity, but they seem seem relevant to back-of-the-envelope cost-benefit thinking about cause areas.

This is a correct point.

However, I think it's worth noting that the most striking advance in aging research in recent years -- the discovery of senolytics -- came from biogerontologists, including those at the Buck Institute, which is dedicated to aging research. Nearly all studies on lifespan are conducted by researchers who specialize in aging. In that sense, I think it's fair to say that a good deal of aging research comes from specialized aging researchers.

Cancer research, I think, should not be considered aging research, because the vast majority of cancer therapies are not also aging-modifying therapies. If aging research pans out, of course, it will produce ways to prevent or treat cancer, but this will probably not depend very closely on the vast corpus of general cancer research.

Stem cell biology and genetic engineering seem to fall into a different category, as important inputs to aging research as well as to many other types of biomedical research. I don't know if EAs have a general framework for evaluating the cost-benefit tradeoffs of more "upstream" or "basic" or "tool" research compared to "downstream" or "translational" or "applied" research -- obviously the benefits of basic research can be much larger than the benefits of translational research, but the variance is also larger.

At any rate, my (non-quantitative, tentative) belief is that aging-focused translational research is more underfunded than research into multipurpose biological technologies (like genetic engineering or stem cell reprogramming), but this may be changing in the near future.

I might do more cost-effectiveness estimates, but it's not a top priority -- I'm currently running the Longevity Research Institute, a nonprofit devoted to experimentally testing anti-aging interventions, and I have a lot of object-level work to do there. Definitely happy to consult, make intros, and share my own existing notes whenever you have questions.

Yes, my estimates are a large underestimate of the potential benefit of life-extending therapies if you assume that they extend life at all, rather than just delay the onset of disease-related disability. I wanted to indicate that the impact is large even with rather pessimistic assumptions.