When I lecture at Cornell about machine learning evaluation or managing software engineering, a lot of the content is shaped by my opinion of what is important to cover based on what previous students have reacted positively to. When I say, “positively”, I mean in a manner which helps motivate them to make progress in their own work.
When I speak to students and faculty at other institutions, I learn slowly how to adapt my speaking style to have the same effect. It is more difficult than you might think, especially in academia where people believe in their style and system of learning evaluation.
When I watch MIT’s Entrepreneurship Center lose track of a reasonable way to treat an invited speaker, it gives me pause to reflect on some of the “bad” talks that we’ve seen at Cornell and Harvard. A “bad” talk is one that doesn’t connect with the audience. It happens frequently because the audience wants to see evidence that the speaker isn’t presenting. But the purpose of a talk is to engage in discussion that may provoke one side to do something better. “Something better” isn’t defined as launching public personal attacks against the invited speaker on the Internet.