3 ms·

The number of jobs that horses can do is a rounding error compared to the number of jobs that humans can do. If we're living in an imaginary future where machin

The number of jobs that horses can do is a rounding error compared to the number of jobs that humans can do. If we're living in an imaginary future where machines can be trained to 90% of jobs, sure. But google a list of "most common jobs" and see how many a computer could do fully. Even if you think a miracle ai revolution is on the horizon, humans and horses are not comparable in terms of diversity of labor

ben_w 6y ago
> Even if you think a miracle ai revolution is on the horizon, humans and horses are not comparable in terms of diversity of labor As someone who does expect an AI revolution, I think you’re being overly literal with an analogy. For arguments’ sake, a toy model of AGI: copy a human brain onto a substrate that costs less to run than a real human costs to feed [0]. For every intellectual task this human could have performed, it is now cheaper to have the AI do it. As this AI could be based on any human by the nature of the hypothetical, any human intellectual task is now done cheaper by the corresponding AGI. Current robot bodies (for the hypothetical uploaded minds) are currently a bit meh, but (a) not so much as to make the idea ridiculous, and (b) we’ve already got BCIs demonstrated in animals, so an AGI puppeting a fast growing cheap animal is also entirely within the realm of the plausible if you don’t want robots and don’t care much about animal rights (though not much worse in that regard than current meat production, IMO). As an aside: tasks is a better metric than jobs, as almost all jobs have many tasks some of which are easier to automate than others. [0] https://worldbuilding.stackexchange.com/questions/51746/when-will-uploaded-minds-be-a-reality/67531#67531 https://worldbuilding.stackexchange.com/questions/51746/when...

3 ms·

The number of jobs that horses can do is a rounding error compared to the number of jobs that humans can do. If we're living in an imaginary future where machin

by hooande 6y ago

ben_w 6y ago
> Even if you think a miracle ai revolution is on the horizon, humans and horses are not comparable in terms of diversity of labor As someone who does expect an AI revolution, I think you’re being overly literal with an analogy. For arguments’ sake, a toy model of AGI: copy a human brain onto a substrate that costs less to run than a real human costs to feed [0]. For every intellectual task this human could have performed, it is now cheaper to have the AI do it. As this AI could be based on any human by the nature of the hypothetical, any human intellectual task is now done cheaper by the corresponding AGI. Current robot bodies (for the hypothetical uploaded minds) are currently a bit meh, but (a) not so much as to make the idea ridiculous, and (b) we’ve already got BCIs demonstrated in animals, so an AGI puppeting a fast growing cheap animal is also entirely within the realm of the plausible if you don’t want robots and don’t care much about animal rights (though not much worse in that regard than current meat production, IMO). As an aside: tasks is a better metric than jobs, as almost all jobs have many tasks some of which are easier to automate than others. [0] https://worldbuilding.stackexchange.com/questions/51746/when-will-uploaded-minds-be-a-reality/67531#67531 https://worldbuilding.stackexchange.com/questions/51746/when...