>be me
Summary
banteg uses a satirical persona to argue that longtermism relies on multiplying astronomical hypothetical populations by extremely uncertain probabilities to justify prioritizing speculative risks over existing human suffering.
Longtermism scales via astronomical numbers.
By assuming trillions of future digital or biological descendants, the author shows how even a one-in-a-million chance of preventing extinction yields an expected value of several galaxies.
extinction destroys 10^50 future lives
Uncertainty is masked by scale.
The author mocks the practice of assigning arbitrary probabilities to speculative interventions, noting that even extreme reductions in probability still dwarf traditional charity by many orders of magnitude.
still beats curing cancer by 38 orders of magnitude
AI safety creates circular logic.
The thread suggests that the urgency of AI safety drives a cycle where researchers must scale models to study alignment, which in turn increases the risk of dangerous capabilities.
we must build the thing quickly because building the thing quickly is dangerous
Governance becomes a tool for control.
The author satirizes the concept of 'value lock-in,' where a small group of current experts attempts to decide the values for all future generations to prevent 'value drift.'
one ideology controls civilization forever
The constituency is unfalsifiable.
The persona concludes that longtermism creates a perfect political constituency because the trillions of hypothetical beneficiaries are silent and cannot challenge the assumptions of their current representatives.
Read the original ↗Save to your Sponge
Summarized with Sponge. Quotes are the author’s own words.