Apparently, it's not as true as we thought it was.
Evolution is not necessarily a process that amounts to the fittest organisms in an ecosystem surviving. Over the short-term, parasitic organisms can thrive due to chaotic diversity which sabotages long-term fitness. In order for the paradigm to continue to make sense, multiple ecosystems must be compared against each other where parasites exist or not in order for parasites to be squashed over the long-run.
If multiple ecosystems are not exposed to each other though, then parasites simply kill themselves off which creates self-destruction within the ecosystem. The fittest organisms disappear, but those who create their disappearance disappear themselves shortly thereafter. Left unto themselves in a controlled ecosystem, the fittest organisms would survive and thrive.
Parasite host interactions are well understood and modeled, as well as disruptive environmental factors (which is just a special case of environmental interactions). Parasite host interactions are actually very similar to predator prey interactions when you model them mathematically.
I think fundamentally this is a confusion about what fitness means in biology. It's about survival and reproduction, and can involve trade offs between survival and reproduction of the same organism. You see the same thing in many viruses where the virulence of the virus causes death but also increases transmission rates. The longer the host lives the better transmission rates are and there is an equilibrium between virulence and time the host stays alive that maximizes overall transmission rate.
Also, relevant to this research, it's actually well understood that organisms will segregate into different feeding strategies in order to minimize competition. This is essentially what happened with darwins finches. When finches first reached the islands their population exploded until they were competing with each other. Starvation ensued which created a selection pressure. Finches that could more easily eat seeds managed to get most of the seeds, finches that could easily get bugs got most of the bugs, etc. based on random variation in beak size and shape. Over the generations the selection pressure reinforced these traits and you get a spontaneous segregation of finches using different feeding strategies.
This is not an example of a single finch type becoming the perfect finch but several finch types evolving using different feeding strategies. Diversity resulted from a single species.
Also this study displays another well understood concept, trade offs. A trait will always have a trade off with something else. In the case of the study it was a trade off between efficient energy usage and mutation rates. One species could expoit the medium better but had a higher mutation rate which resulted in less viable cells being produced (lower reproductive rate). These cells had higher survival but lower reproduction while the mutation resilient cells had lower survival but higher reproduction. Both are reproductive strategies we see all the time at all levels in life.
Ultimately I found this study hardly surprising at all.