It's better to be mostly right, and getting righter, than to be eternally and unchangingly wrong.
Your friend is just scared to death by the notion that our knowledge is provisional and subject to revision. He has an unreasonable demand of certainty in all forms of knowledge, but if you really think through it, there's no reason why we have to be perfectly and absolutely certain about anything.
For almost any application, provisional certainty subject to revision in the light of better knowledge is good enough. Better, because it doesn't put you in the position of being committed to a dogma that later turns out to be flawed.
Perfect certainty is an illusion. The problem your friend has isn't with science, it's with certainty. He has an unreasonable need for it.
Edited by crashfrog, : No reason given.