Don't jump all over him quite so fast. He has a point. Several times we've made the determination (about creatures that are still alive) based on the physical evidence, then retrospectively gone back to examine the DNA and decided they're really not related at all.
On the other hand, that's arguably evolution at work too...parallel evolution.
Convergent evolution. And it's one of the reasons why paleontologists are now so very, very found of advanced/primitive features - such as certain bones and stuff - because they imply actual, not just purely morphological, relationship.
A good example of this would be New World Vultures - condors. Classically they were classified under order Falconiformes. At some point at 1980s people started to pay attention to certain weird features in NWVs' legs that were quite unlike those of Old World vultures, buzzards, accipiters, falcons etc. However, they were quite similar to storks'.
Sibley-Ahlquist phylogenetic tree, based on DNA-DNA hybridization (a relatively new technique, mind you!) essentially confirmed this. Your Turkey Vultures are more closely related to storks than they are to red-tailed hawks or american kestrels (I think you were American. It doesn't really matter though
). They all belong in really ****ed-up superfamily Ciconiiformes, which has stuff like grebes, pelicans, birds of prey (both NWVs and classical Falconiiformes), penguins and stuff.
What's the point? Of course we cannot know everything. The scientific method itself assumes that as we get more information, old theories become unsustainable. The breakthrough of 1950s - explaining DNA - is only slowly starting to pay off. Paleontology, cladistics, phylogenetics, zoology, population ecology and several other things are where they are because of genetics.
OK, so many scientific outbreaks of earlier, more innocent times are now obsolete. Ours could very well be in few centuries or even decades. However, as we cannot predict future we can only take what's observable, provable, falsifiable, empirically testable and use those to build rigorous, mathematic scientific method. Which is self-correcting, mind you. Not anything passes - you can spout inane bull**** and maybe even get it published, but it will get shot down and forgotten. If you actually have something resembling something viable it will come under extremely harsh probing and beating and kicking. If it still lasts, then someone will make a prediction and try to test it. And the bully crowd is still there, just dying to find a proverbial crack in the wall.