A number of New Atheist scientists have begun calling the legitimacy and usefulness of philosophy into question. Most of these people have little clue of what philosophy is even for of course--in addition to thinking that science can be used meaningfully to answer non-scientific questions. Here are three things philosophy can do that science cant't:
1. It can explain what philosophy can do that science can't. Which is another way of saying it can explain how philosophy and science are different. Which is another way of saying that it can tell us what science is and isn't. I have said before that where science fits into the total scheme of things is not a question for an expert in science (a scientist), but a question for the expert in the total scheme of things (a philosopher). Those who professionally deal with what is called "demarcation issue" (where to draw the line between science and non-science) are called philosophers of science. This is a whole specialty within academic philosophy. The question "What is science?" is not itself a scientific question. If there are those who think it is, then they need to explain how you answer it scientifically. Do you take science into a lab and heat it up to see what color it turns? Oh, and who do most scientists invoke when they see a non-scientist intruder in their midst? Karl Popper, a philosopher. In other words, the very question of the relative value of science and philosophy is a philosophical question.
2. It can provide a justification for the assumptions and methodologies science takes for granted. Science employs assumptions about reality and methods of rational procedure that are not themselves amenable to scientific analysis. Scientists (except perhaps quantum physicists of the Copenhagen interpretation) assume cause and effect. Why? Do they have a scientific reason for assuming it? Of course not. Causation is, in fact, a metaphysical assumption. Questions about it are answered in the discipline of metaphysics, which is not the province of science (in fact the people who are now so enthusiastic about running down philosophy specifically repudiate metaphysics), but of philosophy. Induction, which (somewhat misleadingly) is identified most closely with science is a logical procedure which is actually a branch of logic, which, in turn, is a branch of philosophy. The analysis of scientific analysis is not scientific analysis, but philosophical analysis. If someone wrote a book on the nature and process of cause and effect or induction, it would not be a science book; it would be a philosophy book.
3. It can explain how the results of science should and shouldn't be used. In addition to being very useful in discovering things about the natural world, science is also very helpful in the development of technology. But when it comes to questions of how, when, and whether to use this technology, science is of little help, other than to help make these non-scientific decisions better informed. It takes science to make an atomic bomb. It takes philosophy to figure out whether to use it.