I do really like the BLAKE2b hash, but I've been concerned about actually using it in practice (although recently I had an application which it would have suite

I do really like the BLAKE2b hash, but I've been concerned about actually using it in practice (although recently I had an application which it would have suited very well). I'm worried that, having failed to win the SHA-3 contest it will end up relegated into obscurity, and using obscure hashing functions isn't usually a great idea. Is this a valid concern, or am I placing too much weight in the NIST process?