I hate the term "dumbing down". It implies that only smart people can understand science to the max, and in order to communicate with "dumb" people we have to grossly simplify our explanations and leave out key details. In the process, we might feel there's a risk of communicating the wrong thing, and the audience will end up misinformed about the topic at hand.
First off, the very act of using natural language to describe scientific concepts can be considered "dumbing down". Most sciences, and especially the physical sciences, are rooted in math. Whether for theory, simulation, or observational analysis, it's all math. And when results are summarized for communication with other scientists, the paragraphs are loaded with so much jargon and technical terms that it might as well be a foreign language. So if you're not using the math and jargon, you're already several steps removed from the original concept.
Even the word "simplify" gives me heartburn. People usually don't understand the intricate details of science not because they're dumb or because they can't handle complexity, but because they're not fluent in the language of the discipline. It's a matter of training and experience, not raw intelligence.
The game of science communication is translation. It's about taking concepts naturally expressed in a particular (mathematical and jargon-heavy) language, and translating it to the language of the audience. When scientists communicate with the public, it's not as their teachers or professors, but as their interpreters.