Thomas Kuhn objected to Popper's positivist approach to science. Although Popper set a high bar for what he regarded as "scientific," he nevertheless believed that humans can and do make steady progress in learning more about the world. Kuhn was more sceptical; he thought that the best we could do is come up with more and more sophisticated models of reality, without, however, approaching "truth" (knowledge of the Ding-an-sich, or what really lies behind our models).
Kuhn's inspiration was what he called "the Copernican revolution." Prior to Copernicus, Western astronomers since Ptolemy had worked out a very detailed model of how the planets, sun, and stars revolve around the earth. To make their model match observations, they had to add layers of complexity: celestial bodies not only moved in great circular orbits, but also sometimes in smaller orbits around a point in their major orbit (see the explanation and diagram at Wikipedia). Sometimes there were epicycles on epicycles. It was a messy model. Copernicus created his model of a heliocentric solar system partly because it allowed him to dispense with some of the complexity of the older Ptolemaic system. His model was no more accurate, but it appealed to him because it was simpler.
Kuhn described the shift from a Ptolemaic to a Copernican astronomy a "paradigm shift." It was not the result of a gradual improvement in science through falsification or any other such method, but a radical rethinking of the universe on new terms. To him, this proved that Popper's rigorous scientific method did not lead to an ever-closer approximation of the truth, but rather to ever more sophisticated models of reality. He compared these models to human evolution, which has seen homo sapiens evolve from primitive, simple forms to ever more complex forms; and yet humans are not evolving toward any particular end, just as scientific models are not evolving toward any particular truth.
I was with Kuhn up until the analogy with human evolution. For one thing, it is curious for him to point to scientific models as ever more complex, when one of his points with the Copernican revolution is that Copernicus's model was actually simpler than what it replaced. More important, while I see his point that scientific models are only models and not an actual representation of the Ding-an-sich, I find his analogy fundamentally flawed. Humans are not evolving toward any particular end, but science is not the same as evolution. It is true that Copernicus's paradigm of planets orbiting in circular patterns around the sun was not perfect, and would be subject to further revisions by later astronomers, notably Kepler's insight that orbits are elliptical.
On the other hand, there is something fundamentally right about Copernicus's idea. No one is ever going to discover that the earth really is the center of the universe after all, and that the planets and sun are really revolving around it while it remains stationary. They can't, because it is wrong. Neither is anyone going to demonstrate that Kepler was wrong and orbits are really circular rather than elliptical. Unlike evolution, scientific advances cannot travel down certain paths. We may lose knowledge, and people may be deceived for a time, but a scientific advance is not repealable in the logical sense.
I can't quite express my ideas in rigorous terms, because I know that it's possible for scientists to be mistaken; I can't, therefore, assert that science is always moving toward truth. On the other hand, I feel that there is a truth in Copernicus's ideas that go beyond mere modelling to represent what actually happens in the solar system better than the Ptolemaic system. I'm convinced, therefore, that Kuhn is wrong, without being able to come up with a complete theory of my own to replace his.