Writing and Speaking

pg is certainly not the best speaker I've seen but is one of my favourites to listen to. Every word has so much value behind it, and the honesty of his thoughts comes across much better in-person than in-essay.

I notice pg uses footnotes very liberally throughout all of his essays. How do you decide when something should be a footnote, as opposed to another sentence or parenthesized thought or simply redacted?

I'm used to footnotes being links and references, not clarifications or justifications as I often find here.

pg, I may be alone in this, but I think your talks, even when read out verbatim, have an extra dimension that is missing in your essays. When you speak, your curiosity and sense of humor come through strongly.

You like to use writing to explore radical new ideas, and to this end, you refine your essays to have as few qualifications as possible. On the page it sometimes comes off as arrogant. But with your voice, I can hear you proposing these ideas for the sheer delight of a new perspective... the tone says "what if we thought about it this way?"

Also, I'd like to slightly disagree that when one is in an audience, one's critical thinking goes down. It's a matter of knowing how to focus your attention. When I watch someone speak, I'm looking for the unintentional parts as much as the intentional. Where does the person smile and feel relaxed? Where do they seem stressed? What's their body language saying? For a geek metaphor, think of that part in Neal Stephenson's Snow Crash where he describes how certain people have the ability to "condense fact from the vapor of nuance". This gives a whole other channel of information to engage your analytical mind, so watching a speech can become like reading.

False modesty aside, I am a very good public speaker. Doing debate in high school, I had an undefeated regular season, as in not losing a single round. I say this just to point out that I'm not a mediocre public speaker championing the written word.

Paul Graham is right, but it depends more on context than he suggests:

Speaking about a technical subject, you want to communicate the ideas themselves. The emotional content in this case is noise. Paul suggests in the notes that academic talks are more immune to this, but having been to quite a few academic talks and given a few myself, I still find them quite inferior to written papers and one-on-one conversations. True, people can still inject the emotional appeals in papers or conversations, but they tend to get more easily noticed and filtered by the reader or listener without the spellbound effect.

Political debates are perhaps an exception. When you watch a presidential debate, you're not only looking for the president with the best ideas, but a president you believe has the leadership capacity to carry them out. You might personally want the president who has the best ideas, regardless of how charming they appear on camera, but like it or not, a lot of that leadership rests on personal charisma.

It feels to me that PG is simply making excuses for not preparing for his talks. There is no reason technical talks can't be fun, engaging, and full of information. If you only read it out loud once, you're not doing enough prep. Sure you could do a funny talk, which sounds great and doesn't have substance. But it's not a zero sum game.

Don't read your talk out once, read it outloud a dozen times. Don't present it unpracticed infront of the conference hall, present it in front of friends / coworkers first.

Speaking and writing, the two, are a major way that programmers get to be known. It's important that we learn to communicate clearly in an engaging way with our community. If you're having trouble, take a monologue class at your local theater.

Two key passages:

  > Having good ideas is most of writing well.

Disagreed. Written style matters, and whenever it doesn't matter, neither would it matter in spoken form. Your writing style happens to be lean, concise, reduced. But that doesn't just happen -- or are all your essays first drafts? Would they work as well in flowery prose?

  > (...) it was a revelation to me how much less ideas mattered
  > in speaking than writing.

Disagreed. I think I've got a grasp on your basic point: that the effective or required ratio of flashiness to content is invariably higher in talks than it is in essays. In the general case, that is of course not true; flashy but relatively superficial essays evidently exist, and (as you admit) academic talks can exhibit remarkable SNRs. But I'd go further and say that your rule of thumb rarely if ever applies in a meaningful way. Rhetorics are crucial in both media, and communication of ideas isn't the sole purpose of verbal interaction -- be it written or spoken.

What you're suggesting may apply to your personal approach to writing and speaking. As you mention, you feel much more comfortable expressing your thoughts as essays. That's great. There's absolutely no further conclusion we can draw from that.