Feynman’s ornithology

According to Richard Feynman “philosophy of science is about as useful to scientists as ornithology is to birds.” It is undoubtedly a humorous remark, but it is similarly problematic.
First of all, it is a statement about the role of philosophy in science. It is a defendable claim that examining the usefulness of philosophy in connection with science is belongs to the philosophy of science. So it is slightly self-contradictonary to use philosophical arguments against the usefulness of philosophy (in our case against philosophy of science).
Feynman’s concept echoes the so-called Baconian tradition which proposed “theory-free” experiments and observations [John Henry: A short History of Scientific Thought (2012), p. 85]. But it is well known today that there is no an observation-independent observation; there is no way to interpret an experiment without a framework of interpretation, etc. Simply speaking, there is no a theory-free theory – and every scientist uses theories in one or another form.
What is more, Feynman’s bon mot states that the usefulness of ornithology for birds is equal to the usefulness of philosophy of science for scientists. That is, if a bird can sing without interpreting its activities, then a physicist can work without any reflection for the nature of science.
It is an analogy, but we always have to make a distinction between its two functions. Sometimes we use illustrations as Plato did in his cave allegory and sometimes analogies are integral parts of the arguments. Obviously, it is critical whether an analogy belongs to the first or second category. According to Hume “all our reasonings concerning matters of fact are founded on a species of Analogy”, and even the idea of natural law is based on some essential analogies between past and future [Julian Baggini – Peter S. Fosl: The Philosopher’s Toolkit, 2010, p. 53 - 54]. Darwin used the analogy between natural selection and the pigeon breeders’ wok. And don’t forget Galileo’s analogy: “the book [of Nature] is written in mathematical language”. The mathematical equations which are used to describe the reality based on analogies between the abstract mathematical formulations and the nature of physical objects and so on. It would be funny to rewrite the whole history of science as the history of usage and abusage of analogies from Plato to Feynman. And I am curious whether it would be possible to find another solution to depict everything