Evolution happens to populations, not individuals. Selection pressure is exactly what shapes population genetics. During times of stability, genetic diversity is introduced through gene drift, but it is selective pressure that leads to population-level shifts. (Please don't nitpick -- everything is statistical a neutral change can spread through a population even without pressure, but that isn't the main vector).
The article was about frogs, so my comment is in that context. In the case of bacteria, which are asexual and which exist in much greater numbers and have much shorter generation times, the "eureka" model of evolution you seem to be thinking of is more defensible, where one lucky bacterium out of a trillion gets a mutation that gives it some significant new ability (say, like being able to metabolize plastic or something) that allows its descendants to out-compete the rest of the population and take over.
What leads you to believe this adaptation isn't genetic? Do you have an alternative, non-genetic explanation as to why these certain tadpoles are better able to tolerate otherwise lethal doses of salt?
The latter is still evolution.
For example, imagine a population with a current mutation-rate of zero, but there's a dominant allele that makes individuals instantly die to new highly-contagious disease. Later on, the (surviving) individuals are all safe from the disease because that allele is gone. Did they evolve? Yes, even if on a chemical level it was purely subtractive.
This is hardly "scientific" in the sense we like to think of the term, but it's the best one can do when trying to answer unobservable untestable hypotheses without resorting to the supernatural.
Also, science tests models not history. There’s quite a bit you can predict via evolution that turned out to be verified from new discoveries. Consider you look at the local geology and then make predictions about what fossils will be discovered where and then check the actual rocks, that’s an experiment.
The vanishingly rare trick with speciation by natural selection in sexually reproducing creatures is that you have to have find a compatible partner.
For sexual reproduction simply isolate populations and they drift apart genetically. This is why you can get ring species and populations that can’t get in touch naturally are considered separate species even if when artificially reintroduced they could interbreed. Thus many invasive species technically count if they could naturally get in touch the species would already be in that location.
Anyway that’s based on the current definition used by zoologists though some people find this definition unsatisfying but in both directions. Grizzlies and polar bears can interbreed and therefore count as the same species in that definition even though they are developing different traits not just hair color but body weight is diverging.
In the end biology doesn’t give a fuck about our definitions so things are messy.
I had a bit about horizontal gene transfer being more common on closely related organisms which somewhat maintains the idea of species in a single non sexual world, but it wasn’t worth going into.