I'll admit you changed my opinion of AI's future impact on employment, but I don't think anything will change my attitude on AI in general. As the old saying goes - garbage in, garbage out. Knowing the effect censorious, prejudicial algorithms and their creators have on what you get when using it, makes me suspicious of anything it spits out.
I am not convinced that non-AI assisted research is any less garbage in, garbage out. AI just gets there faster. What one does with that extra information and the extra time available to confirm it is as varied as human ability, personality and critical thinking will allow.
This does not mean every worker is safe or every transition is painless. The displaced clerk, teller, farmhand, or analyst deserves seriousness, not slogans. But fear is not strategy. Freezing technology to protect yesterday’s job is how nations become poor, dependent, and beaten by rivals that choose growth. America’s answer should be training, ownership, entrepreneurship, energy abundance, and freedom to build. AI will punish pass-through workers and reward operators, principals, builders, and people who own judgment. Ricardo’s ghost still haunts the debate because panic sounds rigorous. History says otherwise. The machine comes. The economy grows.
I know and understand this whole argument and grant its acuity. But I’m still hanging on what seems different this time. Automation is well on the way to taking out truck drivers and warehouse workers in the same way it has done with manufacturing line jobs. Those are workers who will not be moving laterally into another unskilled blue collar vocational, like farm workers moving into Henry Ford’s factories. We already know the auto workers and steel mill workers of the 90s mostly died of despair and addiction—they did not find new and higher employment. Wiping out such a vast layer of manual labor seems different this time in the scale of human wreckage it will leave behind without remedy. Granted there will be new and unforeseen job categories called forth by the wealth accreting upward in the hands of those capable of manipulating AI systems of various sorts. But the working class is paying the full cost of this for the upcoming generation, and at a scale greater than in the other transitions
I get what you’re saying Mr Anderson and agree there is a lot of that in Americans. One if the things that define us generally as a nation. But I’m concerned for the half of workers—by definition—that are below average. Often they are in jobs that don’t require initiative, creativity, or self reliance because they don’t have any of those qualities. It has always seemed to me an aspect of free market capitalism that takes the harshness off a bit— that there can be a place for the lesser abled but decent people who need the dignity of a job. AI is a technocratic and soulless advance. Very little of it appeals to me
You discredit the resilience and intellectual capacity of workers. We are living through a time when Trump and his administration are reducing regulations in an effort to increase opportunity. We are only thwarted by our own resistance to change.
I've been suffering in transition, if we can call it that, for about 60 years, as technology constantly changed what I could offer the market at an acceptable price. I used to type term papers for a living on electric typewriters (having learned on manual ones). I'm pretty good at glazing windows. No one needs those skills any more.
I'm now retired from the practice of large-scale corporate law, but for the first time, in my old age, I can offer ordinary consumer law help to my neighbors at a reasonable price to them and an attractive wage for me, because AI does the grunt work very, very fast, even in areas that are fairly unfamiliar to me. Much earlier in my career, I had to do legal research by scanning not-very-helpful paper digests and out-of-date treatises. Then came computers and LEXIS or WestLaw research, an improvement, but still extremely slow and limited. For all the wrinkles and quibbles we can find in AI legal research, it is astounding faster and less likely to miss something crucial because of the limitations of old-fashioned word searches and digest keys. Just the savings in expensive time and effort in reworking correspondence, memoranda, and pleadings every time the document needs to reflect a change in factual assumptions or the thrust of a legal argument is more than I could possibly have imagined a year or two ago.
Unstated conclusion: the people I serve now at a price that suits both them and me simply couldn't have afforded my services before, and for the most part would have made do without the legal help altogether. Now they can afford to get help that wouldn't have existed in the economy at all for all practical purposes.
The one problem not addressed is that AI can also stand for Any Indian. Use AI as a reason (excuse) to fire workers but then hire cheaper labor from India and other places overseas. You can still make the circulation argument but the circulation is not in America. Is it a question of AI or H1B and other classes of visas?
Ask the draft horses my grandfather grew up with about mechanization. They weren't replaced, but the tractors are. We're still using the 1933 Cat D2, everything else is modern. Put me in the air-conditioned cab of the Case 4WD that does most of the work, don't make me use horses again.
Am I worried about AI? Yes, but for security reasons, not economic reasons. I was in school when Linux was new and promised open-source robustness. Many code-writers and eyeballs would catch all mistakes. Mythos found a raft of exploits in the code, code that underlies most of the Internet. All those code-writers and eyeballs weren't enough to make the code safe.
When Mythos and the next generation get turned toward the genetic code that underlies life, look out. Will we follow their advice to clean up the errors they find? Or will they just go ahead and do it?
The next evolution of the job market is the least of my concerns.
My own view is that the situation with AI is different. It is not so much a tool or a machine as an extension of the human mind , with all the potential good and bad that entails. How will we regulate and manage a tool that is orders of magnitude (it soon will be) more intelligent than the smartest human being? How will we prevent our adversaries from using AI against us, rather than with our own AI. Human beings could become witnesses rather than protagonists to their own very existence. I’m not saying it will happen, I don’t know, but the threat is real.
No two ways about it: this was a fun and illuminating essay.
I'll admit you changed my opinion of AI's future impact on employment, but I don't think anything will change my attitude on AI in general. As the old saying goes - garbage in, garbage out. Knowing the effect censorious, prejudicial algorithms and their creators have on what you get when using it, makes me suspicious of anything it spits out.
I am not convinced that non-AI assisted research is any less garbage in, garbage out. AI just gets there faster. What one does with that extra information and the extra time available to confirm it is as varied as human ability, personality and critical thinking will allow.
The counterfactual/premise assumption is 100% subjective. This is why economics, like psychology, is considered a "soft science."
This does not mean every worker is safe or every transition is painless. The displaced clerk, teller, farmhand, or analyst deserves seriousness, not slogans. But fear is not strategy. Freezing technology to protect yesterday’s job is how nations become poor, dependent, and beaten by rivals that choose growth. America’s answer should be training, ownership, entrepreneurship, energy abundance, and freedom to build. AI will punish pass-through workers and reward operators, principals, builders, and people who own judgment. Ricardo’s ghost still haunts the debate because panic sounds rigorous. History says otherwise. The machine comes. The economy grows.
I know and understand this whole argument and grant its acuity. But I’m still hanging on what seems different this time. Automation is well on the way to taking out truck drivers and warehouse workers in the same way it has done with manufacturing line jobs. Those are workers who will not be moving laterally into another unskilled blue collar vocational, like farm workers moving into Henry Ford’s factories. We already know the auto workers and steel mill workers of the 90s mostly died of despair and addiction—they did not find new and higher employment. Wiping out such a vast layer of manual labor seems different this time in the scale of human wreckage it will leave behind without remedy. Granted there will be new and unforeseen job categories called forth by the wealth accreting upward in the hands of those capable of manipulating AI systems of various sorts. But the working class is paying the full cost of this for the upcoming generation, and at a scale greater than in the other transitions
I get what you’re saying Mr Anderson and agree there is a lot of that in Americans. One if the things that define us generally as a nation. But I’m concerned for the half of workers—by definition—that are below average. Often they are in jobs that don’t require initiative, creativity, or self reliance because they don’t have any of those qualities. It has always seemed to me an aspect of free market capitalism that takes the harshness off a bit— that there can be a place for the lesser abled but decent people who need the dignity of a job. AI is a technocratic and soulless advance. Very little of it appeals to me
You discredit the resilience and intellectual capacity of workers. We are living through a time when Trump and his administration are reducing regulations in an effort to increase opportunity. We are only thwarted by our own resistance to change.
I've been suffering in transition, if we can call it that, for about 60 years, as technology constantly changed what I could offer the market at an acceptable price. I used to type term papers for a living on electric typewriters (having learned on manual ones). I'm pretty good at glazing windows. No one needs those skills any more.
I'm now retired from the practice of large-scale corporate law, but for the first time, in my old age, I can offer ordinary consumer law help to my neighbors at a reasonable price to them and an attractive wage for me, because AI does the grunt work very, very fast, even in areas that are fairly unfamiliar to me. Much earlier in my career, I had to do legal research by scanning not-very-helpful paper digests and out-of-date treatises. Then came computers and LEXIS or WestLaw research, an improvement, but still extremely slow and limited. For all the wrinkles and quibbles we can find in AI legal research, it is astounding faster and less likely to miss something crucial because of the limitations of old-fashioned word searches and digest keys. Just the savings in expensive time and effort in reworking correspondence, memoranda, and pleadings every time the document needs to reflect a change in factual assumptions or the thrust of a legal argument is more than I could possibly have imagined a year or two ago.
Unstated conclusion: the people I serve now at a price that suits both them and me simply couldn't have afforded my services before, and for the most part would have made do without the legal help altogether. Now they can afford to get help that wouldn't have existed in the economy at all for all practical purposes.
The one problem not addressed is that AI can also stand for Any Indian. Use AI as a reason (excuse) to fire workers but then hire cheaper labor from India and other places overseas. You can still make the circulation argument but the circulation is not in America. Is it a question of AI or H1B and other classes of visas?
Ask the draft horses my grandfather grew up with about mechanization. They weren't replaced, but the tractors are. We're still using the 1933 Cat D2, everything else is modern. Put me in the air-conditioned cab of the Case 4WD that does most of the work, don't make me use horses again.
Am I worried about AI? Yes, but for security reasons, not economic reasons. I was in school when Linux was new and promised open-source robustness. Many code-writers and eyeballs would catch all mistakes. Mythos found a raft of exploits in the code, code that underlies most of the Internet. All those code-writers and eyeballs weren't enough to make the code safe.
When Mythos and the next generation get turned toward the genetic code that underlies life, look out. Will we follow their advice to clean up the errors they find? Or will they just go ahead and do it?
The next evolution of the job market is the least of my concerns.
My own view is that the situation with AI is different. It is not so much a tool or a machine as an extension of the human mind , with all the potential good and bad that entails. How will we regulate and manage a tool that is orders of magnitude (it soon will be) more intelligent than the smartest human being? How will we prevent our adversaries from using AI against us, rather than with our own AI. Human beings could become witnesses rather than protagonists to their own very existence. I’m not saying it will happen, I don’t know, but the threat is real.