As a follow-up to this comment: I gave my 10-minute talk on effective altruism at Scribd. The talk went better than I expected: several of my coworkers told me afterwards that it was really good. So I thought I would summarize the contents of the talk so it can be used as a data point for presenting on effective altruism.
You can see the slides for my talk in keynote, pptx, and html. Here are some notes on the slides:
The thought experiment on the second slide was Peter Singer’s drowning child thought experiment. After giving everyone a few seconds to think about the thought experiment, I asked everyone who thought there was 50% probability or higher that they would save the drowning child to raise their hand (inspired by this essay). Almost everyone raised their hands.
I threw in a few ideas that haven’t seen wide discussion of in the effective altruist community. For example, in the last chapter of Martin Seligman’s book Learned Optimism he explains how he thinks that Western culture’s focus on consumerism and our lack of purpose and connection has contributed to our depression epidemic, which I covered on slides 6-7.
On slide 9, I tried to make things concrete and interesting by suggesting that Scribd could save money by giving everyone Chromebooks to work with, but this would probably end up being bad for the company’s bottom line in the long run because we would work less efficiently.
Another thing I haven’t seen wide discussion of in the EA community is the analogy between groups like Givewell and scientists (see this LW comment for more on this idea). I discussed this on slides 11-12.
On slide 14, I discussed in depth the idea that doctors don’t do all that much good due to replaceability effects. (The 4 principles were stolen from Ben Kuhn’s writeup.)
Overall, I found this experience really encouraging. Initially I was afraid that the drowning child thought experiment would make people hostile, but that didn’t seem to happen at all… there wasn’t any criticism of the idea even during the Q&A period at the end. I was also afraid that the talk tried to cram too many ideas in to just 10 minutes, which may have occurred but all the evidence I observed afterwards suggested to me that the concepts I tried to communicate were well-understood. The people at Scribd are pretty smart though: the talk before mine was about the physics of motorcycle riding, and the talk after mine was by a champion Go player. So a different presentation might be optimal for a different crowd.
Although several people told me they thought the talk was good, I didn’t hear much discussion about the concepts I presented afterwards. And of course it’s hard to measure whether people actually became significantly inclined towards effective altruism or not. So in the long run we should still probably do rigorous message A/B testing.
As a follow-up to this comment: I gave my 10-minute talk on effective altruism at Scribd. The talk went better than I expected: several of my coworkers told me afterwards that it was really good. So I thought I would summarize the contents of the talk so it can be used as a data point for presenting on effective altruism.
You can see the slides for my talk in keynote, pptx, and html. Here are some notes on the slides:
The thought experiment on the second slide was Peter Singer’s drowning child thought experiment. After giving everyone a few seconds to think about the thought experiment, I asked everyone who thought there was 50% probability or higher that they would save the drowning child to raise their hand (inspired by this essay). Almost everyone raised their hands.
I threw in a few ideas that haven’t seen wide discussion of in the effective altruist community. For example, in the last chapter of Martin Seligman’s book Learned Optimism he explains how he thinks that Western culture’s focus on consumerism and our lack of purpose and connection has contributed to our depression epidemic, which I covered on slides 6-7.
On slide 9, I tried to make things concrete and interesting by suggesting that Scribd could save money by giving everyone Chromebooks to work with, but this would probably end up being bad for the company’s bottom line in the long run because we would work less efficiently.
Another thing I haven’t seen wide discussion of in the EA community is the analogy between groups like Givewell and scientists (see this LW comment for more on this idea). I discussed this on slides 11-12.
On slide 14, I discussed in depth the idea that doctors don’t do all that much good due to replaceability effects. (The 4 principles were stolen from Ben Kuhn’s writeup.)
Overall, I found this experience really encouraging. Initially I was afraid that the drowning child thought experiment would make people hostile, but that didn’t seem to happen at all… there wasn’t any criticism of the idea even during the Q&A period at the end. I was also afraid that the talk tried to cram too many ideas in to just 10 minutes, which may have occurred but all the evidence I observed afterwards suggested to me that the concepts I tried to communicate were well-understood. The people at Scribd are pretty smart though: the talk before mine was about the physics of motorcycle riding, and the talk after mine was by a champion Go player. So a different presentation might be optimal for a different crowd.
Although several people told me they thought the talk was good, I didn’t hear much discussion about the concepts I presented afterwards. And of course it’s hard to measure whether people actually became significantly inclined towards effective altruism or not. So in the long run we should still probably do rigorous message A/B testing.