A while back I wrote an article titled What Standup Comedy Has Taught Me About Public Speaking and Life. That article provides a pretty comprehensive coverage of everything I know about public speaking. But even with all of that experience and knowledge, I’ve been learning a lot more in the process of preparing for my first public technology talk. Or perhaps I’ve just been revisiting, deepening, and consolidating existing realizations.
Despite the fact that I’ve been immersed in the technology world since I was a child, and having now been a professional engineer for 23 years, I somehow managed to avoid giving a public talk about technology until now. Next Wednesday, I’ll be presenting at the 2019 GPU Technology Conference (GTC) in San Jose, an artificial intelligence (AI) conference that attracted over 8,400 attendees last year. This conference happens to be organized by my employer: NVIDIA. That makes the experience a little more familiar than it would otherwise be, but also raises the stakes because I’ll be representing my company on its own turf, and I expect that many of my colleagues will be watching.
I anticipate that my talk will be held in a room that can hold at least one-hundred people. I know which room it is, and it looks pretty big on the map, but I have not yet stood in it. I’ve performed five-minute standup comedy sets to audiences of over one-hundred people before, and I’ve also presented at work to over one-hundred people for ten or fifteen minutes. But this talk will last for fifty minutes, consisting of forty minutes of presentation followed ten minutes of questions and answers.
The title of my talk is Determinism in Deep Learning. I’ll be telling the tale of my technical quest over the past year to make AI-related systems operate the same way every time. For various reasons, it’s extremely important—assuming you don’t change any parameters—that your AI “factory” produces the same AI every time. It’s equally important that any given AI will behave the same way every time it is stimulated with the same input.
The most compelling scenario is if an artificially-intelligent robot kills or injures someone (by accident of course). In that case, it must be possible to exactly reproduce the process that generated the machine and its behavior. That way, the machine itself, and the process that created it, can be debugged and possible future harm can be mitigated.
When I started preparing for this forty-minute talk, I didn’t think I had enough material to present. However, as I’ve fleshed-out my story, I have ended up with forty-nine carefully-constructed slides and a rehearsed presentation that takes between forty and forty-three minutes to deliver. Coincidentally, to present forty-nine slides in exactly forty minutes, it’s necessary to spend, on average, forty-nine seconds on each slide. That’s a very short amount of time, which requires the presentation to be very crisp and focused.
The integration of feedback from various reviewers of this presentation and previous similar presentations has converged on this particular set of slides. The slides tell a comprehensive narrative and they include enough information to be useful on their own, without my presence. But it turns out that with technology presentations there is a challenging tension between completeness and distraction. The slides need to speak for themselves, but if they contain too much information, then they can distract from the talk.
A couple of weeks ago, I presented my talk to a small group of colleagues, after which I collected and integrated their feedback. One of the reviewers told me that I should “um” and “ah” less. Then, yesterday, I did a run-through with my wife, Cindy Riach, who is an executive performance coach specializing in authentic communication. She gave me a lot of feedback, the most important part of which related to those “um”s and “ah”s. Even after it had been pointed out by my colleague, I still didn’t realize that I continued to produce these noises. I also had no idea why I was doing it.
Cindy was able to use her psychological insight to facilitate the debugging of the problem. She noticed that I would be doing fine—looking out at the audience, standing confidently, and behaving engagingly—when suddenly I would say “um.” She noticed that the “um” and the words that followed the “um” would sound robotic and monotonous. I would look down, cave-in my chest, and seem to be talking to myself. She imagined that this would cause the audience to switch off and stop listening.
We zoomed-in on that and discovered that I say “um” when I feel under pressure to say the “right” thing and when that “right” thing doesn’t seem to be appearing in my mind quickly enough. I say “um” both as a filler and to hide the shame I feel when I notice that my mind is blank. At the same time, because of the shame, I also shut down emotionally; my “heart closes,” and then my voice loses its emotional tone, its vibrant aliveness. My voice becomes dead and lifeless for the next few words until my heart begins to naturally open once again. This is a very accurate phenomenological description of what seems to happen in my experience.
The work with Cindy lead to the realization, a realization which I’ve had before, that I must primarily focus on keeping my heart open. I really like the phrase “keeping my heart open” because, even though it could sounds like new-age bullshit, it’s a great label for what I’m about to describe.
I can be a compelling speaker. Cindy has watched how I become engaged and passionate in a conversation, and that I can talk endlessly without ever saying “um.” Partly because of all the therapy and meditation I’ve done, I’m able to be transparent and authentic, to appear “real” in my behavior. The problem is that when I feel threatened, or scared, or stressed, I have a tendency to shut down my emotions and “power through” with execution. This is actually a really good description of the dysfunctional behavior of my personality type: enneagram type three, the achiever. Taken to the extreme, and when made chronic, this looks like workaholism, a never-ending treadmill of doing at the expense of being. This is all driven by the unconscious hope that by doing enough I will finally feel that I am enough, necessarily missing the fact that I already am, and always have been, enough.
My mind goes blank because I think there’s a right thing to say next, and because I’m looking for that, looking for something that’s not there. Looking for what to say is the cause of the blankness. There’s no need to look, the words appear naturally when there is no effort, when there is a playful engagement with the audience. But when there is a blank mind, it is judged as being wrong and bad, which leads to the arising of fear and shame. The body responds to all of this by shifting into a protective posture, and the emotional-self closes in defense. As Cindy says, I then come across as some kind of an asshole.
I’ve worked on this project for a year. I think it’s important. I know enough about it to talk for hours on the topic with no slides. I often do deliver what seem to be compelling ad-hoc monologues on the subject. The trick for me is to ensure that I feel safe on stage, to remember that I can do this with no slides at all, and to focus on engaging with the audience and acting “as if” I am calm, confident, and playful. If I notice that my mind is blank, it’s a sign that I’m looking for some kind of “right” words, it’s a sign that I can be silent for a moment and then shift back into a playful stance.
With all of that in mind, I ran through the presentation again today while Cindy watched. I also recorded a video of it. This time, instead of talking through the slides, I said to myself, fuck the slides. The slides came with me on my talk. They supported what I was saying. I didn’t talk to them. My dynamic, alive, fresh presentation, filled with off-the-cuff anecdotes, jokes, silliness, and excitement drove the presentation. I felt free to briefly address slides that spoke for themselves, to summarize things in simpler terms for the audience, and to make the presentation into an act of play. Anyone can read the slides while I’m not there. It’s not supposed to be a somber and perfect delivery of a particular set of words.
After today’s rehearsal, I re-asserted a goal that I set early-on for this presentation. At the time, it wasn’t what I had imagined “should” have been the goal. The first-order, logical goal might have been for people to understand the technical details. But that’s not what I want to focus on. I want for people to get a feeling for how much fun it can be to work with me. The goal is for attendees to come away from the talk wanting to join NVIDIA, or to at least be motivated to reach out and connect and to follow the open-source project that I’ll be announcing. If I can achieve only that, then I will have achieved more than enough. And the “business” reasoning for why that’s actually more important is that if I have that—if have rapport, trust, friendship, and enjoyment—then the technical details will be delivered automatically and efficiently from one brain to another anyway.
I’m blessed to be able to talk coherently about this topic in my sleep. I don’t have to prove anything. I don’t have to “get it right.” All I have to do is have fun.
The talk was attended by 137 people. I enjoyed it immensely and I learned a ton. The average ratings by the attendees were:
- Content: 4.68/5
- Speaker: 4.74/5.