I liked it, but there are a number of things that could have been a lot better:
There were way too many digressions. Though subsequences work well in writing, it’s hard to follow a chain of reasoning that jumps between levels. Though the stories about peoples’ strange opinions at dinner parties were illustrative, some of them go on for way too long. Likewise, recursing into reductionism then recursing into Bayesian Judo and then popping back out into the discussion of Occam’s Razor was a bit confusing because so much time was spent on those topics that I forgot it was a digression.
There was a little too much meta. Talking about the talk, and talking about how Richard Carrier’s talk should have come first, is off-putting and not useful.
More cognitive science examples might have helped. One of the most interesting and engaging parts of the talk was the beginning, in which the audience was given the red/green die test. More “DIY” examples of cognitive biases may have helped stress that skeptics are also prone to these errors. For example, the hindsight bias test from David Meyers might have helped to drive home that part of the talk. (On that note, why is the talk called “Heuristics and Biases”? It was really about Occam’s Razor.)
And for the love of Cthulhu, beta-test the jokes. A lot of them just aren’t funny, the most awkward one being “Good thing I’m not a god...yet.” The “Bayesian Hell” joke also went on way too long, i.e. well past the point where the audience stopped laughing. In addition, there was a little too much arrogance in some of the jokes. For example: “I must have been divinely inspired, because I said, without any forethought whatsoever, I said...”
Additionally, it could have used more structure, e.g., “I’m going to talk about X, Y, and get to Z” before talking about X, Y and Z—it’s possible to do so much of that that it becomes redundant and annoying, but a small amount would have greatly improved the talk. I followed it (and enjoyed it), but I think a fair amount of the audience was pretty lost.
The contrast was particularly jarring because he spoke immediately after David Silverman, who was a very polished speaker.
Eliezer’s talk has been posted.
I liked it, but there are a number of things that could have been a lot better:
There were way too many digressions. Though subsequences work well in writing, it’s hard to follow a chain of reasoning that jumps between levels. Though the stories about peoples’ strange opinions at dinner parties were illustrative, some of them go on for way too long. Likewise, recursing into reductionism then recursing into Bayesian Judo and then popping back out into the discussion of Occam’s Razor was a bit confusing because so much time was spent on those topics that I forgot it was a digression.
There was a little too much meta. Talking about the talk, and talking about how Richard Carrier’s talk should have come first, is off-putting and not useful.
More cognitive science examples might have helped. One of the most interesting and engaging parts of the talk was the beginning, in which the audience was given the red/green die test. More “DIY” examples of cognitive biases may have helped stress that skeptics are also prone to these errors. For example, the hindsight bias test from David Meyers might have helped to drive home that part of the talk. (On that note, why is the talk called “Heuristics and Biases”? It was really about Occam’s Razor.)
And for the love of Cthulhu, beta-test the jokes. A lot of them just aren’t funny, the most awkward one being “Good thing I’m not a god...yet.” The “Bayesian Hell” joke also went on way too long, i.e. well past the point where the audience stopped laughing. In addition, there was a little too much arrogance in some of the jokes. For example: “I must have been divinely inspired, because I said, without any forethought whatsoever, I said...”
I saw it in person and agree with all the above.
Additionally, it could have used more structure, e.g., “I’m going to talk about X, Y, and get to Z” before talking about X, Y and Z—it’s possible to do so much of that that it becomes redundant and annoying, but a small amount would have greatly improved the talk. I followed it (and enjoyed it), but I think a fair amount of the audience was pretty lost.
The contrast was particularly jarring because he spoke immediately after David Silverman, who was a very polished speaker.
Agreed, more structure would have been good, because I had no idea where he was going with each chain of reasoning.