So recently I’ve found myself not finishing, or even skipping, some of the AI episodes. I see the guests, think I can guess the general perspectives they and Rob/Luisa will take on AI, and don’t think it’ll add too much to my knowledge of the topic. If there are particular episodes that you think this is particularly incorrect about, then please let me know!
I’d be very surprised (and very impressed) if the Carl Shulman episodes did not add much to your knowledge of the topic (relative to how much you learned from the listed episodes).
For the avoidance of doubt, not gaining knowledge from the Carl Shulman episodes is at least as much my fault as it is Rob and Carl’s![1] I think similar to his appearance on the Dwarkesh Podcast, it was interesting and full of information, but I’m not sure my mind has found a good way to integrate it into my existing perspective yet. It feels unresolved to me, and something I personally want to explore more, so a version of the post written later in time might include those episodes high up. But writing this post from where I am now, I at least wanted to own my perspective/bias leaning against the AI episodes rather than leave it implicit in the episode selection. But yeah, it was very much my list, and therefore inherits all of my assumptions and flaws.
I do think working in AI/ML means that the relative gain of knowledge may still be lower in this case compared to learning about the abolition of slavery (Brown #145) or the details of fighting Malaria (Tibenderana #129), so I think that’s a bit more arguable, but probably an unimportant distinction.
(I’m pretty sure I didn’t listen to part 2, and can’t remember how much I listened to of part 1 over reading some of the transcript on the 80k website, so these episodes may be a victim of the ‘not listened to fully yet’ criteria)
I’d be very surprised (and very impressed) if the Carl Shulman episodes did not add much to your knowledge of the topic (relative to how much you learned from the listed episodes).
For the avoidance of doubt, not gaining knowledge from the Carl Shulman episodes is at least as much my fault as it is Rob and Carl’s![1] I think similar to his appearance on the Dwarkesh Podcast, it was interesting and full of information, but I’m not sure my mind has found a good way to integrate it into my existing perspective yet. It feels unresolved to me, and something I personally want to explore more, so a version of the post written later in time might include those episodes high up. But writing this post from where I am now, I at least wanted to own my perspective/bias leaning against the AI episodes rather than leave it implicit in the episode selection. But yeah, it was very much my list, and therefore inherits all of my assumptions and flaws.
I do think working in AI/ML means that the relative gain of knowledge may still be lower in this case compared to learning about the abolition of slavery (Brown #145) or the details of fighting Malaria (Tibenderana #129), so I think that’s a bit more arguable, but probably an unimportant distinction.
(I’m pretty sure I didn’t listen to part 2, and can’t remember how much I listened to of part 1 over reading some of the transcript on the 80k website, so these episodes may be a victim of the ‘not listened to fully yet’ criteria)