Created: 2021-07-15 | Updated: 2021-07-16 | Acknowledgement for feedback: Emanuele Ascani, Haydn Thomas-Rose | x-post: LessWrong

By short-termism, I basically mean a method of analysis where one assumes the world stays the same. With longtermism, indirect effects and uncertain futures are taken into account. The longtermist approach focuses on increasing the probability of different stable outcomes, whereas the short-termist approach focuses on the short-term direct impact.

When it comes to personal life extension, short-termist strategies include things like:

  • Increasing the number of dreams, and notably lucid dreams experienced and remembered
  • Reducing the amount of time sleeping
  • Avoiding drugs that erases memory
  • Having habits that are conductive to faster thinking

At a first approximation, these interventions might all look great. They can all increase subjective life expectancy more than they decrease objective life expectancy, and maybe even robustly so.

However, when the stakes are astronomical, this approach is completely backward. The long termist approach for personal life extension focuses on dying later, not (directly) on living more. Using the following as a toy example, that could mean things like sleeping 10% more only to live 1% more, which doesn’t seem to make sense given this would reduce one’s total subjective lifespan at a first approximation, except that this 1% increase in objective lifespan could actually transform into astronomically more because it could allow one to live long enough to live maximally long; i.e. reach a point where cryonics, anti-aging, or other anti-death technologies reach longevity escape velocity. Some longtermist interventions will come at a negligible or zero short-term subjective life expectancy cost, such as lifelogging as life extension, while others will greatly diminish your short-term subjective life expectancy, such as pre-emptive biostasis.

The longtermist approach still has overlapping recommendations with the short-termist one, but when it does, they will sometimes arise for different reasons, and to a different degree. For example, sleeping less might be good to earn more money and therefore be able to afford better biostasis instead of being good because it increases the length of your short-term subjective experience. On the other hand, other interventions are good from both the short-termist and longtermist perspectives because they increase both subjective and objective lifespan. Such interventions include a healthy diet, a healthy amount of safe exercise, breathing clean air, etc. This category tends to be the one people naturally think about when thinking about longevity.

Some interventions are not obvious to categorize; they will depend on the context and the specific reasoning behind them. For example, avoiding a new experience because it cannot be lifelogged would help avoid a micro-death at the cost of the experience. On the other hand, this experience might help preserve your self as a whole (ex.: attending a cryonics conference that is not lifelogging-friendly), and so might be worth sacrificing.

For an agent with no diminishing returns on additional subjective experiences, it’s still not literally impossible for the short-termist approach to be superior (ex.: maximum lifespan could be small or unlikely), but in practice, it seems to me like the longtermist approach is basically always correct.

There will still be various interventions within the longtermist approach that will trade-off against each other. For example, there will be longtermist arguments to get cryopreserved earlier (ex.: avoid further identity degradation) and there will also be longtermist arguments to get cryopreserved later (ex.: to wait until cryonics technologies get better).

For more in this vein, see the Facebook group on death-focused ethics