Prior to deployment of AIs in various industries, I expect noticeable economic effects from personal productivity boosts -- much larger context windows kicks this off I think, then I expect a quite significant impact from the widespread usage of relatively near-future, personal "executive assistant" systems.
Great podcast! 1:49:16 I think Tom answered the less interesting question here. The parent question assumed that AGI will come to be, so "can you see things continuing as they have for the past hundred years?" might have been interpreted with the same assumption in mind. In the world where we have AGI, could that somehow NOT have a hugely transformative impact? Neither beneficially or detrimentally? I think it's very unlikely. AGI will almost certainly be transformative, but if there are ways of unfolding in which it isn't, it's worth wondering about the features of those futures.
The thought of a company board holding back what would be a very powerful AI system out of concern for the public interest is almost laughable at this point. I appreciate the sentiment of course.
Any time you think about takeoff speeds, remember that AI experts most familiar with exponential progress are orders of magnitude too pessimistic in their wildest predictions.
UBI is not a reliable solution in the long term. When most humans are no longer laborers, companies no longer need money for labor costs, and everything else they need they will get from other companies (or "factions") with their own AI. They will trade in kind, or adopt a new currency. In that case, the markets we are currently familiar with will collapse, the economy will cut most humans out of the loop, and UBI will just be a way of describing how a government or other entity will somehow coerce those who have AI to create stuff for humans who don't. Somehow.
Prior to deployment of AIs in various industries, I expect noticeable economic effects from personal productivity boosts -- much larger context windows kicks this off I think, then I expect a quite significant impact from the widespread usage of relatively near-future, personal "executive assistant" systems.
That lamp is about to take off for sure
Great podcast!
1:49:16 I think Tom answered the less interesting question here. The parent question assumed that AGI will come to be, so "can you see things continuing as they have for the past hundred years?" might have been interpreted with the same assumption in mind. In the world where we have AGI, could that somehow NOT have a hugely transformative impact? Neither beneficially or detrimentally?
I think it's very unlikely. AGI will almost certainly be transformative, but if there are ways of unfolding in which it isn't, it's worth wondering about the features of those futures.
The thought of a company board holding back what would be a very powerful AI system out of concern for the public interest is almost laughable at this point. I appreciate the sentiment of course.
Any time you think about takeoff speeds, remember that AI experts most familiar with exponential progress are orders of magnitude too pessimistic in their wildest predictions.
They always love slow as possible to make as much chatter, media content, and ultimately profit with anything they do don't they? 🤑
All economic tasks are cognitive tasks.
UBI is not a reliable solution in the long term. When most humans are no longer laborers, companies no longer need money for labor costs, and everything else they need they will get from other companies (or "factions") with their own AI. They will trade in kind, or adopt a new currency. In that case, the markets we are currently familiar with will collapse, the economy will cut most humans out of the loop, and UBI will just be a way of describing how a government or other entity will somehow coerce those who have AI to create stuff for humans who don't. Somehow.