Personally I am going to start a company where the "AI" is actually a bunch of people in a business complex somewhere writing, typing, or speed painting all the content in real time.
We will be the first AI company whose final product is based fully on the Wizard of Oz model.
Sorry to say, you'd be far from the first. This is literally what "IBM Watson" was. (The consulting service, not the computer that played Jeopardy.) A bunch of traditional consultants cosplaying as an AI.
“The factory of the future will have only two employees, a man and a dog. The man will be there to feed the dog. The dog will be there to keep the man from touching the equipment” – Warren G. Bennis.
There aren't a lot of tools available to government (or capital for that matter) which can sustainably juice employment. Only an organic increase in demand will do it without incurring inflation imo, and the only things on the horizon which will pull that off are AI (ironically) and fusion power.
UBI (mentioned in the article) may well work fine on a national-economic level, but it's existence threatens several business models which are predicated on the availability of masses of unemployed and unskilled people, which would be a self-reinforcing phenomenon.
Long story short is that AI is coming for capitalism. I believe any solution would have to entail the ideation of a new economic system, and what passes for the aristocracy these days is almost certainly not going to be on-board, inevitably leading to historically interesting times.
In Bullshit Jobs, Graeber wrote that the great job loss due to automation already happened; it was just swallowed by a ballooning middle management class. That ballooning happened for power reasons, higher up managers signaling their status by expanding their teams, while those teams don't have much to do.
Graeber used the terms of neo-feudalism for this supposed pattern: every manager has their own little fiefdom. If Graeber was right about the existence of this pattern, we might even see an expansion of those fiefdoms: even more bullshit jobs!
The term he uses is “box tickers” as a class of bullshit job of memory serves right (been a bit since I read it). But basically we created make-work for millions to keep people going into the office.
Why does anyone think the role of creating new jobs in the wake of losing jobs to due to AI should fall on the shoulders of the government, and why do people believe they could be successful at it?
> Governments will have to find a solution for knowledge sector workers whose jobs are automated away thanks to the advent of artificial intelligence, a leading expert in the field warned.
This is a horrible way to solve this possible future problem.
Automation always creates jobs that we cannot foresee. For example, the more jobs that are lost to AI the more openings there will be at the guillotine workshop.
The tags they put on this article are: Cryptocurrency Metaverse, and Cybersecurity. AI could definitely come up with a better categorization that that :p
Jokes aside, I think we are seeing this question come up every other day. Has anyone proposed a solution? Ideally, based on data?
An N-hour work week is like the front page of a newspaper: it has to be filled. If we say that a person should have to work so many hours to earn a living, that's the same as saying our society should have to work so many hours per person to function.
Unfortunately, that's not how the real world works. If farmers produce more food than people want to eat, the food will go uneaten and the farmers' time will have been wasted. If carpenters produce more housing than people want to live in, the housing will stand vacant and the carpenters' time will have been wasted.
Does this mean the farmers and carpenters get to work less? No, because if they only work 30 hours a week, they won't make enough money to function in the society that assumes people have to work 40 hours.
So instead some of the farmers and carpenters have to find other work that needs to be done. Which is great, but only so long as there really is other work that needs to be done. If there isn't, then the surplus farmers and carpenters have to find performative work--something to soak up 40 hours and appear valuable so they can get a paycheck.
Which is bad. Not just because it wastes the lives of the people who do that unneeded work, but because it even wastes the lives of the people who do the farming and carpenting. They have to work a full 40 too, even though everyone could be working fewer hours on jobs that matter.
I know, I know, "But if we adopt a 40-N hour work week, we'll lose lots of things that make our lives better. There will be fewer snacks to choose from at the grocery store; fewer video games, all of inferior quality; fewer computers and TVs and fewer everything that goes into making them."
And maybe so, although personally I don't think we'll lose anything we'll miss much. But we'll all have more of a life to improve for ourselves, instead of the way TV or Instagram dictates we should. My happiness will not suffer for lack of cheese puffs, and I'll actually enjoy learning how to cook a lot of things I used to buy because I will have the time and energy needed to enjoy it. So my washing machine won't have a phone app anymore. Big deal, I never used it anyway. And it least it's a lot more repairable now, since companies have been forced to figure out how to add competitive value to it without investing a lot more time in it.
I'm looking around at my world and damn near every part of it is designed to imprison me in some way. I got a bottle of shampoo saying it gave me 20% extra for free, because statistically speaking I'm more likely to think I'm getting more value than if it didn't, even though the price per ounce is the same no matter if it says it gave me 99% for free. Somebody was paid to come up with that trick. And I paid them to try that trick on me, even if I was just buying the lowest price per ounce.
Claims about the near term impact of AI are always vague. Exactly which jobs are going away, and how many of them are there? Is there empirical research supporting these conclusions?
From my point of view, the impact of AI in the past couple of decades has primarily been better translation, computer vision, speech recognition/synthesis, scientific computing, recommender systems, and the latest wave of generative AI. These are excellent tools, but they aren't capable by themselves of doing a job that a human does. Autonomous vehicles are an exception to this, and they do have the potential to disrupt the transportation industry. But it seems like there are still significant problems to be solved before we can deploy them on a large scale. We don't seem to be making much progress on creating truly generally intelligent agents, either. I think any claim that these technologies are going to have a large impact on the overall number of jobs in the next 5-10 years is highly speculative at this point.
The underlying issue is that the current hype and fears around AI are being used to push a specific political agenda. Even if I grant for the sake of argument that millions of jobs will be automated away in the next decade, that doesn't automatically imply that UBI is the solution. Why is UBI a better solution than expanding existing social programs? What impact will UBI have on inflation and other economic indicators? Will UBI have a positive or negative impact on societal problems like structural inequality?
Of course, it could turn out that the predictions about AI are true. And maybe UBI is in fact the best solution. All I want to see is a nuanced conversation on this issue, grounded in empirical data, that weighs the pros and cons of the alternatives. That's not what I'm seeing in the media narratives around AI right now.
> Is there empirical research supporting these conclusions?
No there's no empirical research because ChatGPT 3 is what - 3 months old ? Are you seriously saying this tool can't replace people in many many tasks?
>No there's no empirical research because ChatGPT 3 is what - 3 months old ?
Right, so these claims are based on speculation, not evidence.
>Are you seriously saying this tool can't replace people in many many tasks?
I can't think of a single profession where ChatGPT could be used without expert input/guidance. It's a useful tool, not an agent capable of doing work independently.
26 comments
[ 5.1 ms ] story [ 64.8 ms ] threadinspect element: ``` document.querySelector('.paywallActive').classList.remove('paywallActive') ```
``` document.querySelector('.paywallActive').classList.remove('paywallActive'); document.querySelector('.paywallFade').classList.remove('paywallFade'); document.querySelector('.paywall-selector').classList.remove('paywall-selector') ```
Optional: This also hides that subscription banner (or you can just scroll past it if it doesn't both you)
``` $(document).ready(function () { $(".pn-offer-grid__body").hide(); }); ```
We will be the first AI company whose final product is based fully on the Wizard of Oz model.
UBI (mentioned in the article) may well work fine on a national-economic level, but it's existence threatens several business models which are predicated on the availability of masses of unemployed and unskilled people, which would be a self-reinforcing phenomenon.
Long story short is that AI is coming for capitalism. I believe any solution would have to entail the ideation of a new economic system, and what passes for the aristocracy these days is almost certainly not going to be on-board, inevitably leading to historically interesting times.
Graeber used the terms of neo-feudalism for this supposed pattern: every manager has their own little fiefdom. If Graeber was right about the existence of this pattern, we might even see an expansion of those fiefdoms: even more bullshit jobs!
> Governments will have to find a solution for knowledge sector workers whose jobs are automated away thanks to the advent of artificial intelligence, a leading expert in the field warned.
This is a horrible way to solve this possible future problem.
Jokes aside, I think we are seeing this question come up every other day. Has anyone proposed a solution? Ideally, based on data?
Unfortunately, that's not how the real world works. If farmers produce more food than people want to eat, the food will go uneaten and the farmers' time will have been wasted. If carpenters produce more housing than people want to live in, the housing will stand vacant and the carpenters' time will have been wasted.
Does this mean the farmers and carpenters get to work less? No, because if they only work 30 hours a week, they won't make enough money to function in the society that assumes people have to work 40 hours.
So instead some of the farmers and carpenters have to find other work that needs to be done. Which is great, but only so long as there really is other work that needs to be done. If there isn't, then the surplus farmers and carpenters have to find performative work--something to soak up 40 hours and appear valuable so they can get a paycheck.
Which is bad. Not just because it wastes the lives of the people who do that unneeded work, but because it even wastes the lives of the people who do the farming and carpenting. They have to work a full 40 too, even though everyone could be working fewer hours on jobs that matter.
I know, I know, "But if we adopt a 40-N hour work week, we'll lose lots of things that make our lives better. There will be fewer snacks to choose from at the grocery store; fewer video games, all of inferior quality; fewer computers and TVs and fewer everything that goes into making them."
And maybe so, although personally I don't think we'll lose anything we'll miss much. But we'll all have more of a life to improve for ourselves, instead of the way TV or Instagram dictates we should. My happiness will not suffer for lack of cheese puffs, and I'll actually enjoy learning how to cook a lot of things I used to buy because I will have the time and energy needed to enjoy it. So my washing machine won't have a phone app anymore. Big deal, I never used it anyway. And it least it's a lot more repairable now, since companies have been forced to figure out how to add competitive value to it without investing a lot more time in it.
I'm looking around at my world and damn near every part of it is designed to imprison me in some way. I got a bottle of shampoo saying it gave me 20% extra for free, because statistically speaking I'm more likely to think I'm getting more value than if it didn't, even though the price per ounce is the same no matter if it says it gave me 99% for free. Somebody was paid to come up with that trick. And I paid them to try that trick on me, even if I was just buying the lowest price per ounce.
This kind of shit does not make me feel free.
(To be fair, let's not say "no clue". But overall their ability to help folks actually find jobs seems quite marginal).
From my point of view, the impact of AI in the past couple of decades has primarily been better translation, computer vision, speech recognition/synthesis, scientific computing, recommender systems, and the latest wave of generative AI. These are excellent tools, but they aren't capable by themselves of doing a job that a human does. Autonomous vehicles are an exception to this, and they do have the potential to disrupt the transportation industry. But it seems like there are still significant problems to be solved before we can deploy them on a large scale. We don't seem to be making much progress on creating truly generally intelligent agents, either. I think any claim that these technologies are going to have a large impact on the overall number of jobs in the next 5-10 years is highly speculative at this point.
The underlying issue is that the current hype and fears around AI are being used to push a specific political agenda. Even if I grant for the sake of argument that millions of jobs will be automated away in the next decade, that doesn't automatically imply that UBI is the solution. Why is UBI a better solution than expanding existing social programs? What impact will UBI have on inflation and other economic indicators? Will UBI have a positive or negative impact on societal problems like structural inequality?
Of course, it could turn out that the predictions about AI are true. And maybe UBI is in fact the best solution. All I want to see is a nuanced conversation on this issue, grounded in empirical data, that weighs the pros and cons of the alternatives. That's not what I'm seeing in the media narratives around AI right now.
No there's no empirical research because ChatGPT 3 is what - 3 months old ? Are you seriously saying this tool can't replace people in many many tasks?
Right, so these claims are based on speculation, not evidence.
>Are you seriously saying this tool can't replace people in many many tasks?
I can't think of a single profession where ChatGPT could be used without expert input/guidance. It's a useful tool, not an agent capable of doing work independently.