We don't "know" that, because that's not a category of thing we can know or not know. It's a matter of semantics of whether we consider birds dinosaurs, just like it's a matter of semantics of whether we consider people a kind of fish.
"Know" in science is used informally to mean "the preponderance of evidence supports this conclusion. Which could turn out to be wrong if enough contrary evidence is later found."
The scientific consensus today is that the evidence supports the idea that not all the dinosaurs died out when the Chicxulub comet struck 66 million years ago. The ones that could fly or quickly learn how to fly survived and even thrived, and their grandkids are in your back yard right now.
You're not understanding my point. The scientific evidence is that birds are descended from dinosaurs. Scientific evidence can't make birds be dinosaurs, because it's up to us to define "bird" and "dinosaur", and we can define "bird" to exclude dinosaur (and I think most people still do). Just like we can define "fish" and "human", and we can define "fish" to exclude human, even though some fish learned how to walk on land and one of their grandkids is typing this comment right now. Biologists might wish that we all adopt cladistic definitions for types of organisms, but that doesn't the world will follow.