I have stumbled upon a talk that David gave at the International Congress of Mathematics in 2001 titled
“Trends in the Profession of Mathematics”. The first quote reflects David’s style a lot, as I remember him giving lectures and talks:

“I think mathematicians have a special problem in making new ideas accessible to their colleagues, a problem that is tough but not unsolvable if they only recognize it more honestly. It is our obsession with seeking to express each new result in its greatest generality! […] But do we want to live in the house that Bourbaki built? I want to express a radical alternative that I learned from Sir Michael Atiyah. His view was that the most significant aspects of a new idea are often not contained in the deepest or most general theorem which they lead to. Instead they are often embodied in the simplest examples, the simplest definitions and their first consequences. […] The most important message is often contained in a few simple, but profound observations which underlie the whole rest of the theory.”

This is something that I aspire to do and hopefully accomplish some day. As my friend told me once after the first encounter with David at the conference “his comments are simple, but deep at the same time.”

And the second quote tells a bit about applied mathematics:

…there is a teeming cauldron of phenomena present in the world asking for clarification and analysis. One tries to snatch out of this cauldron some specific things which lend themselves to a precise analysis. This can only be done by radical simplification but it _must_ preserve the essence of some aspect of the complexity of the full rich situation. I think mathematics can benefit by acknowledging that the creation of good models is just is significant as proving deep theorems. Of course, for a model to be good, you must show it leads somewhere: this may be done by mathematical ‘experiment’, i.e. by computations or by the first steps in its analysis. PhD’s, lectures and _jobs_ should be awarded for finding a good model as well as proving a difficult theorem.

I suggest to all of you to read the full talk.