4 ms·

So far we've automated physical tasks. It was possible to go into intellectual jobs as an alternative avenue, since automating that wasn't possible. However now

So far we've automated physical tasks. It was possible to go into intellectual jobs as an alternative avenue, since automating that wasn't possible. However now we're automating mental work. Humans can currently go into jobs requiring higher education or creativity but 1) this won't work for everybody and 2) this is not a new avenue outside of the reach of automatization. The machines have a good chance of catching, especially since a lot of those humans moving into more specialized jobs will be tasked with making better machines, which in turn endangers more jobs.

decayy 10y ago
We seem to be ignoring the current theoretical limitations of A.I. I'm not an C.S or A.I expert. But what I can remember from university is that we need a pretty big breakthrough in computer science to get to that stage. I'm not saying it won't happen. But who knows.
- majewsky 10y ago
  Even if we cannot automate everything, we are going to be in big trouble (on a social level) if the work left to humans only occupies anything less than about 50% of adult people. The relevant thing will be the ratio of decay rate of jobs that are automated away vs. creation rate of new jobs, but that's hard to estimate because technological advances are hard to estimate.
  - decayy 10y ago
    It's true. The past is not a perfect indicator of future. This time it could be different. The transition wont be smooth and people in large number will actively try and stop technological progress.
- eva1984 10y ago
  It is already happening. The whole deep learning thing is no joke. They cannot replace human, but can achieve close to or better than human level of performance in certain specific tasks. Before that, close to human performance is never achieved for certain tasks, and largely considered impossible in near future just 5 years ago, like image recognition/playing go/speech recognition. It is already eliminating jobs, gradually, not overnight but steadily, and this time it is from higher above, say, Wall Street traders.It is very easy to see there is a trend out there, that data + smart algorithms is going to take over a lot decision-making processes, that currently belongs to analyst. The theoretical limitation of AI is bullshit. We don't get know what intelligence is, how do you think we quantify and draw a line around it?

decayy 10y ago

We seem to be ignoring the current theoretical limitations of A.I. I'm not an C.S or A.I expert. But what I can remember from university is that we need a pretty big breakthrough in computer science to get to that stage. I'm not saying it won't happen. But who knows.

majewsky 10y ago
Even if we cannot automate everything, we are going to be in big trouble (on a social level) if the work left to humans only occupies anything less than about 50% of adult people. The relevant thing will be the ratio of decay rate of jobs that are automated away vs. creation rate of new jobs, but that's hard to estimate because technological advances are hard to estimate.
- decayy 10y ago
  It's true. The past is not a perfect indicator of future. This time it could be different. The transition wont be smooth and people in large number will actively try and stop technological progress.
eva1984 10y ago
It is already happening. The whole deep learning thing is no joke. They cannot replace human, but can achieve close to or better than human level of performance in certain specific tasks. Before that, close to human performance is never achieved for certain tasks, and largely considered impossible in near future just 5 years ago, like image recognition/playing go/speech recognition. It is already eliminating jobs, gradually, not overnight but steadily, and this time it is from higher above, say, Wall Street traders.It is very easy to see there is a trend out there, that data + smart algorithms is going to take over a lot decision-making processes, that currently belongs to analyst. The theoretical limitation of AI is bullshit. We don't get know what intelligence is, how do you think we quantify and draw a line around it?