> all Scientific Knowledge has an 80% chance of being proved wrong
> in twenty years time

Apologies, but I can't let this go unchallenged. All scientific theories are approximations to what people think is The Truth. Theories aren't usually proven "wrong" so much as "not sufficiently right" in the light of new evidence. The classic case is Newton's laws: they're still accurate enough for NASA to use it as-is to launch rockets, but not *quite* accurate enough for the clocks on Sat-Nav satellites (which require a General-Relativity correction to keep in step with groundside clocks).

Oh, and QI is entertainment, lightly leavened with enlightenment. Go not to the comedians for scientific rigour, for they tell both truth and jest.

.... Again apologies: I'm short on minutes (we're going to Eat Out soon), and writing diplomatically is a slow artform.