Not so fast!
I always enjoy Olivia Judson's columns in the New York Times, but today's post on evolution "failing" left out an important point. She referred to a paper published last year from Richard Lenski's long-term evolution experiment, showing that a bacterial population took 31,000 generations to evolve the ability to use citrate. Furthermore, although she didn't mention this, this trait has only evolved, so far, in one of their twelve replicate populations. If evolution is too slow to keep up with the changes we humans are making in the environment, then species that might evolve and survive if changes were slower will instead go extinct.
I agree that this is a significant problem, but I wouldn't assume that it would take polar bears, for example, 31,000 generations to evolve adaptations to warmer temperatures. The bacteria that Lenski's group studies don't have sex. So if one cell has a mutation that would allow it to use citrate, but only in combination with a second mutation found in another cell, they don't have any way to combine the two mutations in one citrate-using individual. If cells with only one mutation or the other have no advantage over cells with neither, then lineages with the first mutation will usually die out before acquiring the second mutation. A lineage could die out, for example, because the next mutation is gets is one of the many lethal ones, rather than one of the few beneficial ones.
Bacterial populations can sometimes evolve rapidly (with significant changes in only a few days) because their generation times are so short and because their large population sizes include many mutants. Evolution requiring a series of steps isn't a problem so long as each step is an improvement. But when a mutation is neutral or negative, except in the context of a second mutation, sexual species can evolve faster. Not necessarily fast enough to save the polar bears, though.