The Great Filter

Robin Hanson (
Sat, 24 Aug 96 14:16:09 PDT

John K Clark writes:
>Earth formed about 4.6 billion ago, but life didn't have a chance for it's
>first billion years or so, it was far too hot and the Earth suffered impacts
>so large it would make the one that killed the Dinosaurs look like a joke.

A billion years seems too long. Do you have a source re the actual
temp vs. time curve? And bacteria are lots hardier than Dinos.

>The oldest undisputed fossils are a little over 3.5 billion years old ...
>Best estimate I've seen is that Eukaryotic cells evolved between 1.2 an 1.4
>billion years ago, well over 2 billion years after the first life. Well OK, I
>remember reading a report somewhere about 1.6 billion, but I don't think
>that's generally accepted.

My sources are pretty authoritive (see paper cites), and clearly say
that Eukaryotic cells appeared about 1.8-2.0 bya (billion years ago),
when the atmosphere became oxygen-dominated after all the ocean's iron
was finally oxidized. (Thus this was probably not a hard step.) Such
cells became lots more common though about 1.1 bya, apparently with
the invention of sex (a good hard step candidate).

>>For Great Filter issues, "hard" would be expected times of a trillion
>>years or more. An expected times 700my is far from hard enough to matter.
>Obviously we have zero chance of finding anything a trillion years old, but I
>don't think 700 million is a trivial percentage of life's 3.5 billion year

I'm speaking of "trial and error" type steps, search across a flat
fitness landscape. If the expected time is a trillion years, the
chance of it happening in a billion is one in a thousand.

>>The really hard stages, I think, do not cluster near us in time.

Here I meant such trial and error steps, which can only go up. Steps
which can go up or down, such as mass descruction, can cluster in time.

Robin Hanson