But that is not going to happen. We would need tangible, meaningful productivity improvements for it to be possible. LLMs are sort of moderately useful, but companies adopting it see no real productivity increase (they see cost increase however).
In many ways it is a bullshit technology, being marketed way beyond its capabilities.
I don't think the 8-hour work day was accepted during the industrial revolution because job providers were convinced of productivity gains. It was won due to the leverage the labor unions and workers had over job providers.
The whole thing rests on whether workers can maintain their leverage in the AI era. That leverage is being eroded by the attempt by job providers to decouple humans from production and productivity (whether they will be succesful in this iteration is, as you say, debatable). The other axis is the attempt to erode democratic norms in order to prevent the majority from checking the power of a wealthy minority.
Once those two points of leverage are completely eroded, the only leverage left is mob power. And as history has shown on multiple occasions, excercising that leverage is ugly and messy.
Good point. Count me in.
> The whole thing rests on whether workers can maintain their leverage in the AI era. That leverage is being eroded by the attempt by job providers to decouple humans from production and productivity
Being brutally honest, I have zero concerns about AI displacing jobs en masse. I honestly believe it to be a deadend technology.
And I say this as someone that uses it daily.
I am a lot more concerned about the incoming economic downturn. Things have the potential get very ugly pretty soon. This will have a huge detrimental impact on labor rights for the foreseeable future. AI is just smoke and mirrors, something asshole CEOs can gesture at to spin layoffs as optimistic and sound like visionaries while mismanaging companies under their control.
sure, but don't expect companies to pay you the same as they did for the 5-day work-week.
And there lies the rub: yes, AI can increase productivity. But the gains of that productivity are being, and will be, captured entirely by shareholders, not employees.
There's room here for argument and further evidence gathering. I've read of cases where the leap in productivity fully justifies holding pay steady. But admittedly these leaps may or may not hold up over the longer run.
Many people might (say) trade a 10% pay cut for a 50% jump in weekly days off. Others can't afford to do that because, well, that's the state of things. But work is becoming a scarce resource, and if you look at the numbers, more is desired (especially. part-time) by pensioners, students, and others. If only health care was state-funded, so that it was not a deterrence to headcount.
Look at this: https://www.epi.org/productivity-pay-gap/
(there are plenty of other sources corroborating these findings)
Companies have been pocketing the gains in productivity for almost 50 years. How much more evidence do you need? You think they will stop now?