12 ms·

Caveat: I don't agree with the author, I'm just trying to understand both sides here. But, It seems to be a just an Economists (albeit a lefty one) looking for

by pj_mukh 11y ago

Caveat: I don't agree with the author, I'm just trying to understand both sides here. But, It seems to be a just an Economists (albeit a lefty one) looking for solutions in previous data. 1) That is the only conclusion that can be gleaned from the past century of data. Projections as to the effect of an "exponential explosion in AI/ML" have no data to support it. 2) I don't think he advocates for either. Though as a lefty he definitely doesn't advocate for busting of social programs (esp Healthcare) 3) Very True. 4) Labour doesn't necessarily ask for increased welfare, they ask for perfect employment. Which the author implies is a larger force for productivity boosts than something like UBI. It also gives the masses some control (therefore keeping the pitchforks away). Historically, weakening labour has led to slumps in productivity for a myriad of reasons (that he could've done a better job at laying out). And we have no evidence that a UBI would increase productivity, but we do know it will destroy labour unions. Without productivity increases (towards zero marginal cost), UBI may not be able to deliver on its lofty promisies. On the flip side: Increasing power to labour would mean another concentration of power that can (and has) be corrupted easily.

kapnobatairza 11y ago
1) I think that the difference comes from how you look at the data. In the past, jobs that were displaced were replaced by other forms of unskilled or minimally skilled labour. The total pool of unskilled/minimally skilled jobs never shrunk. It is still too soon to tell what effect intelligent automation will have on that total pool. It could very well be that for every job automation technology displaces, it creates ten more. But if those jobs are highly skilled jobs that the existing labour force can't transition into? Then we have a problem. 4) I think that perfect employment is a good goal, but how we get there is something else entirely. We can create more jobs, we can "re-educate" workers, we can decrease hours to create more shifts, but all of this still leaves a lot of people in the cracks. I think that one of the main benefits of a UBI is that it guarantees that nobody falls through the cracks. I don't see how UBIs would entirely destroy labour unions either, since labour unions are often about more than setting minimum hours and negotiating standard compensations. Unions provide a lot more utility than that so I don't see workers abandoning their unions just because they get a check in the mail from the government every month. If that was the case, the union probably had very little utility to begin with IMO. I don't know if a UBI is the right solution for solving poverty or worker inequality. But I don't think the author makes a convincing argument for an alternative, and I don't think doing nothing is a solution either.
aninhumer 11y ago
>1) That is the only conclusion that can be gleaned from the past century of data. Projections as to the effect of an "exponential explosion in AI/ML" have no data to support it. I think a more useful way to think about the problem of automation isn't that it replaces jobs directly, it's that it devalues human capital. New jobs will involve more skill, which means each person will need more education and training. It might get to the point where training everyone so they can work for a living is less efficient than just taxing and paying for some people's sustenance directly.
- eru 11y ago
  Wasn't deskilling the thing people complained about with new technologies?

by pj_mukh 11y ago

kapnobatairza 11y ago

1) I think that the difference comes from how you look at the data. In the past, jobs that were displaced were replaced by other forms of unskilled or minimally skilled labour. The total pool of unskilled/minimally skilled jobs never shrunk. It is still too soon to tell what effect intelligent automation will have on that total pool. It could very well be that for every job automation technology displaces, it creates ten more. But if those jobs are highly skilled jobs that the existing labour force can't transition into? Then we have a problem. 4) I think that perfect employment is a good goal, but how we get there is something else entirely. We can create more jobs, we can "re-educate" workers, we can decrease hours to create more shifts, but all of this still leaves a lot of people in the cracks. I think that one of the main benefits of a UBI is that it guarantees that nobody falls through the cracks. I don't see how UBIs would entirely destroy labour unions either, since labour unions are often about more than setting minimum hours and negotiating standard compensations. Unions provide a lot more utility than that so I don't see workers abandoning their unions just because they get a check in the mail from the government every month. If that was the case, the union probably had very little utility to begin with IMO. I don't know if a UBI is the right solution for solving poverty or worker inequality. But I don't think the author makes a convincing argument for an alternative, and I don't think doing nothing is a solution either.

aninhumer 11y ago

>1) That is the only conclusion that can be gleaned from the past century of data. Projections as to the effect of an "exponential explosion in AI/ML" have no data to support it. I think a more useful way to think about the problem of automation isn't that it replaces jobs directly, it's that it devalues human capital. New jobs will involve more skill, which means each person will need more education and training. It might get to the point where training everyone so they can work for a living is less efficient than just taxing and paying for some people's sustenance directly.

eru 11y ago
Wasn't deskilling the thing people complained about with new technologies?