A few observations to add, in the form of common mistakes.
A. "The audience as computer" aka "important things need to be written to memory once." In fact, nobody learns anything except the simplest things the first time. Important points should persist across a few slides.
1/
A. "The audience as computer" aka "important things need to be written to memory once." In fact, nobody learns anything except the simplest things the first time. Important points should persist across a few slides.
1/
https://twitter.com/ShengwuLi/status/1443938193674485761
E.g. if there is an important, longish Bellman equation, DON'T just put it up once and move on. Have it hanging around on several slides.
Blackboards talks are better because of persistence. Most speakers using slides ignore that things take time to absorb.
2/
Blackboards talks are better because of persistence. Most speakers using slides ignore that things take time to absorb.
2/
B. "Literature details before content."
It's amazing how often you hear, at minute 7, "but our eigenvalue condition will actually be in terms of the Laplacian matrix, rather than the usual adjacency matrix as in XYZ (2014)"
before anyone can follow this.
3/
It's amazing how often you hear, at minute 7, "but our eigenvalue condition will actually be in terms of the Laplacian matrix, rather than the usual adjacency matrix as in XYZ (2014)"
before anyone can follow this.
3/
Put contrasts and details where a non-specialist member of your audience (e.g., non-networks theorist, or non-theory economist for a job talk) can appreciate them, not where they went, say, in your paper.
4/
4/
C. Alluding to important background information without teaching it.
If you ever find yourself saying something like, "And as we all remember from X," where X is any literally almost any paper, rethink your model of the audience.
5/
If you ever find yourself saying something like, "And as we all remember from X," where X is any literally almost any paper, rethink your model of the audience.
5/
A good talk presents not just your paper but teaches the immediate research context (in a focused and relevant way) when it's needed. If you find yourself talking to an idealized audience of your referees, remember that your actual audience is never that audience.
6/
6/
D. Following conventions in the field, without realizing that the modal talk is unsuccessful.
Think of the best talk you've seen in the last year. Are you trying to imitate ITS best attributes, or enacting a template you're personally bored and confused by?
7/
Think of the best talk you've seen in the last year. Are you trying to imitate ITS best attributes, or enacting a template you're personally bored and confused by?
7/
E. Not treating a talk as an important performance.
A senior theorist told me that after his first big talk, a very senior theorist told him that while a seminar looks like no big deal, it's Carnegie Hall for us - "Why didn't you treat it that way?"
A senior theorist told me that after his first big talk, a very senior theorist told him that while a seminar looks like no big deal, it's Carnegie Hall for us - "Why didn't you treat it that way?"
The best speakers take the craft very seriously. They rework important visuals many times, get all the feedback they can, kill their darlings, pare down tirelessly, etc. A good talk takes many days of work, at a minimum, to get right.
9/9
9/9
• • •
Missing some Tweet in this thread? You can try to
force a refresh
