I think that Peter Stone is speaking to a different question than the one I am trying to address, but it is a reasonable confusion to make since he is speaking to the AI question around which this community is largely organized. I am talking about automation in the colloquial way it gets talked about in media—robots or software doing work humans used to do. The conversation as I witness it usually plays out like this: someone asserts automation is a problem, and then someone else responds (correctly) that big automation waves in the past resulted in bigger, more productive economies, and most people found some other way to contribute.
The question is, if it turned out alright before, why wouldn’t it turn out alright this time? The thing that I feel is lacking, including from the usual economics arguments, is any sense of why it turned out alright the first time. This is important because that is where we can tell whether to worry about it this time.
One of the intuitions that we are working with here is that whether automation is a problem depends on the margins. Rather than a question of whether all people are affected, the question is only whether enough are affected. I think the answer is yes, and I think that because mostly the same processes are at work this time as last time, but in a way that won’t let us roll with the punches like before.
It’s also worth considering that this is for dropping into an already-in-progress conversation, as distinct from starting a conversation from first principles.
watch this video if you don’t believe me
https://www.youtube.com/watch?v=kDqLIwgHf2U
Thank you for linking the video.
I think that Peter Stone is speaking to a different question than the one I am trying to address, but it is a reasonable confusion to make since he is speaking to the AI question around which this community is largely organized. I am talking about automation in the colloquial way it gets talked about in media—robots or software doing work humans used to do. The conversation as I witness it usually plays out like this: someone asserts automation is a problem, and then someone else responds (correctly) that big automation waves in the past resulted in bigger, more productive economies, and most people found some other way to contribute.
The question is, if it turned out alright before, why wouldn’t it turn out alright this time? The thing that I feel is lacking, including from the usual economics arguments, is any sense of why it turned out alright the first time. This is important because that is where we can tell whether to worry about it this time.
One of the intuitions that we are working with here is that whether automation is a problem depends on the margins. Rather than a question of whether all people are affected, the question is only whether enough are affected. I think the answer is yes, and I think that because mostly the same processes are at work this time as last time, but in a way that won’t let us roll with the punches like before.
It’s also worth considering that this is for dropping into an already-in-progress conversation, as distinct from starting a conversation from first principles.