Fair question! I buy the McClements and Wiseman argument that growth post-AGI could be transformative, but not explosive (e.g. 3-9%), and that robotics is a significant hurdle to achieving explosive growth. Without that, we should expect to have much more money, but not enough tax revenue for governments to suddenly supplement or replace the wages of white collar workers. Once you "solve robotics", tax revenues might catch up, but it's not obvious that governments would spend to 100% replace former workers' wages (or equivalent purchasing power) if those people don't offer return on their investment.
Internet companies like Google, Meta, etc. without engineers will have higher profit margins than Saudi Aramco. If the next president after Trump will be a democrat, he will probably put unprecedently high taxes on them. And if the situation is dire enough even a Republican may do this.
Overall, the fate of white collar strata outside the US seems worse because many of them are dominated by US internet companies and those high taxes from Google, Meta, etc. won't be replenished back into those countries' budgets (e.g.: UK).
Many people will lose their jobs and will not be able to pay for their disproportionately large apartments with a pool and air conditioning, possibly becoming homeless. If companies around the world start laying off people, this will significantly reduce consumption, and the question is where the economic growth will come from if consumption is decreasing. Reducing consumption will lead to the government collecting less taxes. So it is not clear where the government will get money for UBI and shitty jobs. Most people will become homeless so I would not be sure that construction workers, plumbers, electricians, and other blue-collar jobs will be safe. If people do not have money, they will not be able to go to the doctor and will die from a simple disease, and also will not buy a car, computers, smartphones, etc. One thing I can say for sure is that I do not see political and economic means to solve this crisis.
You are going to kick yourself for wasting all your time writing this stuff. AI can’t even interpret an excel sheet without ignoring and missing gigantic important inputs. It’s insanely dumb and inefficient. This whole article reads like science fiction
I agree with Luke when he says maybe. But insanely dumb or science fiction? Hardly. Even if none of this happens tomorrow or 10 years from now, it's *exactly* what AI frontier companies aspire to and intend. How, when or if they succeed is in question BUT determined minds, under the reduced AI regulation of now, can result in momentum. Not to mention the AI you describe, which can't even interpret an excel sheet without ignoring and missing gigantic important inputs;" this is a mere bump in the road at this time. As AI/AGI continues to advance and develop, such inconveniences will be a speedy little blip in the past.
Whether or not I fully accept the scenario in its extreme forms (I'm undecided, though it seems plausible), I think it's relevant to note that people often underestimate the possibility that demand for labour could rise significantly even as we are outcompeted.
A typical hopeful response to your post might point to the new opportunities AI could create. I certainly think there will be some of those, along with productivity boosts for some workers along the way. AI could very well expand the total amount of work that can be done and is worth doing. I can't imagine a time when there is no work (last I checked, entropy ensures that), but I can imagine a time when we are no longer the most desirable workers in various domains.
Well done Luke...This is great ammunition as we conduct conversations, internally and externally, about the Future of Work, Worker and the Workplace. Many implications to consider here.
> we simply do not have enough money for UBI
I’m confused by this part - in a world where white collar productivity doubles (the agent runs at half the cost), why wouldn’t we have enough money?
Fair question! I buy the McClements and Wiseman argument that growth post-AGI could be transformative, but not explosive (e.g. 3-9%), and that robotics is a significant hurdle to achieving explosive growth. Without that, we should expect to have much more money, but not enough tax revenue for governments to suddenly supplement or replace the wages of white collar workers. Once you "solve robotics", tax revenues might catch up, but it's not obvious that governments would spend to 100% replace former workers' wages (or equivalent purchasing power) if those people don't offer return on their investment.
https://inferencemagazine.substack.com/i/155018281/conclusion
Internet companies like Google, Meta, etc. without engineers will have higher profit margins than Saudi Aramco. If the next president after Trump will be a democrat, he will probably put unprecedently high taxes on them. And if the situation is dire enough even a Republican may do this.
Overall, the fate of white collar strata outside the US seems worse because many of them are dominated by US internet companies and those high taxes from Google, Meta, etc. won't be replenished back into those countries' budgets (e.g.: UK).
Keyword here being we. *We* don't have that money, the owners of the AI agents which automated that white collar productivity do.
Many people will lose their jobs and will not be able to pay for their disproportionately large apartments with a pool and air conditioning, possibly becoming homeless. If companies around the world start laying off people, this will significantly reduce consumption, and the question is where the economic growth will come from if consumption is decreasing. Reducing consumption will lead to the government collecting less taxes. So it is not clear where the government will get money for UBI and shitty jobs. Most people will become homeless so I would not be sure that construction workers, plumbers, electricians, and other blue-collar jobs will be safe. If people do not have money, they will not be able to go to the doctor and will die from a simple disease, and also will not buy a car, computers, smartphones, etc. One thing I can say for sure is that I do not see political and economic means to solve this crisis.
Great post. I would really like to work on trying to figure out this problem that government is not nearly close to prepared to tackle.
This far more realistic scenario keeps me up at night. The consequences go beyond me and my partner losing my job in the next decade.
You are going to kick yourself for wasting all your time writing this stuff. AI can’t even interpret an excel sheet without ignoring and missing gigantic important inputs. It’s insanely dumb and inefficient. This whole article reads like science fiction
Maybe! But I’d rather be wrong in public than be right in a Google Doc.
I agree with Luke when he says maybe. But insanely dumb or science fiction? Hardly. Even if none of this happens tomorrow or 10 years from now, it's *exactly* what AI frontier companies aspire to and intend. How, when or if they succeed is in question BUT determined minds, under the reduced AI regulation of now, can result in momentum. Not to mention the AI you describe, which can't even interpret an excel sheet without ignoring and missing gigantic important inputs;" this is a mere bump in the road at this time. As AI/AGI continues to advance and develop, such inconveniences will be a speedy little blip in the past.
Bleak but necessary outlook. Enjoyed this! (to an extent...)
Whether or not I fully accept the scenario in its extreme forms (I'm undecided, though it seems plausible), I think it's relevant to note that people often underestimate the possibility that demand for labour could rise significantly even as we are outcompeted.
A typical hopeful response to your post might point to the new opportunities AI could create. I certainly think there will be some of those, along with productivity boosts for some workers along the way. AI could very well expand the total amount of work that can be done and is worth doing. I can't imagine a time when there is no work (last I checked, entropy ensures that), but I can imagine a time when we are no longer the most desirable workers in various domains.
Thanks for sharing it :)
Well done Luke...This is great ammunition as we conduct conversations, internally and externally, about the Future of Work, Worker and the Workplace. Many implications to consider here.