another factor militating against longer planet durations is hydrogen escape. If a planet is massive enough to hold on to its hydrogen for billions of years, it will rarely have a thin enough atmosphere for light to reach the surface, and without photosynthesis your raw energy available for evolving life is reduced by at least 6 orders of magnitude

@JonathanRay and the probability of evolving intelligent life on a planet is probably not just a power-law on time, but a power law on the product of time*area*energy*otherconditions. Suboptimal planets (e.g., no light reaching the surface) can basically be ignored as having a negligible chance of evolving intelligent life.

The dark forest hypothesis seems like a reasonable explanation: why put the effort into constructing a giant transmitter it it creates a large X-risk for you.

The longer-lived low-mass stars are flare stars incompatible with life because their luminosity varies over too wide of a range (multiple orders of magnitude) So it's a bit misleading for the grabby aliens paper to proceed from the assumption that the average star lives for 5 trillion years. To get out of the flare star range you need at least 0.6 solar masses which leads to a star lifespan of only 35 billion years.

@JonathanRay https://en.wikipedia.org/wiki/Habitability_of_red_dwarf_systems

Presently, we can only perceive ourselves as purely deterministic animals with no unique extent into the quantum multiverse. If there was an unavoidable multiverse ethics that concluded ‘grabby aliens’ was undesirable then by choice all aliens would, in almost every branch, look inward rather than expand outwards.

If building misaligned AI were a common failure mode for alien civs we ought to see some utilitronium shockwaves spreading out at significant fractions of the speed of light. D-T fusion with a Q-factor of infinity has a theoretical maximum specific impulse of about 10% c (any other kind of fusion being much worse) and U-235 fission would have a theoretical maximum specific impulse of 3% c. By the classical tsiolkovsky equation you need a ship that's 99.3% fuel to get to half the speed of light. Seems likely that even a superintelligent AI would not be able to make its utilitronium shockwave go faster than 80% of the speed of light.

Or

The Great Filter:

@Gigacasting most alien civilizations as smart as us would be increasing their intelligence via breeding, biotech, information technology, or other means