What's your current AI career strategy: Already changing careers, learning AI tools, ignoring predictions, planning for post-work society, or focusing on AI-resistant skills?
There's one thing that's often overlooked: humans always want things that are hard to get. When AI makes productivity abundant and all products and services become cheap, our needs and desires simply move on to other things that are hard to get. The attention of a stranger, a silent spot in the city, a roundtrip to the moon, two minutes with Oprah, watching every scifi movie ever created, a hamburger from moose meat, a hike passing through all European capital cities, a kiss from a porn star, a wedding at the bottom of the ocean, the honest empathy of a non-machone therapist. There's no end to the things that are hard to get, which means there's no end to human needs and desires. And where there's demand, there is supply. Where there's supply, there is labor.
Jurgen you speak truth here. Proverbs 27:20 comes to mind "Just as Death and Destruction are never satisfied, so human desire is never satisfied.” Great comment!
Yes. I’d add that AI challenges not just work, but also identity.
And too often, our response is reactive: save the economy, retrain the workers, don’t fall behind. But what if this moment isn’t just a threat … it’s a test of values? What do we protect, what do we surrender, and who gets to decide?
The real danger isn’t AI … it’s not a good vs evil story … but to your point … what we do with this powerful technology … I’ve been writing a lot in this too. Happy that this framing is coming across other writers
Great article. This quote really rang true: “Here's what puzzles me: Hinton spent 50 years building this technology, then acts surprised by its implications.”
While I appreciate Hinton sounding alarm bells, he did in fact create the monster.
He has forced big tech to finally admit there will be mass layoffs. The new solution is that the tools will make it easy for EVERYONE to become an entrepreneur because the tools will make it so easy. Currently entrepreneurs make up about 16% of the adult workforce.
I’m reaching with a potential solution to the impending jobs crisis that will be knocking at our door with the inevitable arrival of advanced AI and Robotics.
I believe if Australia was to adopt some of my recommendations. We could not only avoid the jobs catastrophe that no one yet has an answer for, but rather set Australia up with a policy that will encourage global adoption as we drive immigration to Australia as that is where the jobs will be.
I will highlight some opportunities definitely worth exploring.
AI don’t take Jobs they get allocated jobs. The Problem itself is quite simple to solve, but will the super wealthy let that solution play out at the loss of abundant wealth?
1.Put AI/Robots on the payroll! Or as least place a monetary value on their contribution. This contribution would scale depending on how many human jobs were taken.
2. AI/Robots would contribute to taxation. If an employer works robots past the standard daily hours/ overtime rates kick in and double over time after that. This stops abusing the protocol.
3. This principle equalises the equation somewhat especially in some industries where the AI/Robot advantage is minimal
4. The income stream to government is maintained.
5. Where a business replaces a worker with an AI/Robot that decision is scrutinised, recordered and the displaced worker receives part of that AI/Robots wage until re- employed.
6. Stress is removed from the welfare system, abundance arrives as discoveries will continue to be made and technologies improved.
7. More than I employee displaced! Apply a multiplier to the above.
8. Employers still win if the do their cost benefit analysis ie: where one AI or Robot displaces an employee the levee applied is one to one but if the robot can optimise the position leading to productivity gain this then benefits the employer the increased productivity flows to GDP and the benefits get realised across all systems.
This is a short / medium term solution to help the planet transverse this period of adjustment. Robots will eventually do everything but we buy some precious time to find a safer landing place.
Martin, you are so right and that punchline hits! 😂 But I’ll play peacemaker and say this: a good college education can make it easier to catch on to abstract stuff like AI and tech-so it has it’s sweet spot and advantages. That said, I also know some master tradespeople who could outthink and outwork a few 'tech geniuses' before their morning coffee. Nice comment!
I think I just got really lucky by growing up in Germany going to trade school and then university just to end up in America is a beach bum and then entrepreneur
Maybe AI will create a market for elite human knowledge. Would a consulting client pay the same fee for AI generated out as it would be a partner’s advice who has spent 20 years in an industry specific niche accumulating knowledge?
What do you think of Gary Marcus’s work and the recent Apple white paper critiquing LLMs and how poorly they perform outside the distribution of their input data? Interested readers can start with his “7 replies” post (https://open.substack.com/pub/garymarcus/p/seven-replies-to-the-viral-apple), but can also look at his summary post, the white paper itself, or the follow-up discussing a paper where LLM performance on hard coding tasks dropped to 0%(!) when controlling for data contamination.
A lot to address here. I think this article raises an important point.
I think many people are not ready to address these questions because there are a slew of assumptions most people make about the value and importance of work. You see this across the political spectrum.
In my opinion this story starts a long time before AI. The power of labor started to decline long before AI, in the early 1970s. Many people point to political trends to explain this (I.e. neoliberalism), but I think technological forces have been driving these trends since the beginning. Unionized labor was strong when companies needed many skilled workers all working together in the same factory which was located in the market where the final product was sold. During the Vietnam war new shipping methods were developed. Containerized shipping made outsourcing viable. It was then possible to use cheap overseas labor, which undermined Unions thereafter. Likewise most of the increased productivity per worker over the past 50 years has been driven by computers. Part of why labor hasn’t seen the benefits of this increase in productivity is that there is nothing special about what the workers are doing — the productivity is not coming from labor power as much as computer power, and thus there is no incentive to pay labor more (a.k.a. Anyone can answer emails).
The problem is that it is unhealthy for society for such an imbalance between labor and capital to persist. I’m a big fan of Graeber’s ‘Bullshit Jobs’ and through that book I’m aware of Keynes’ “Economic possibilities for our grandchildren”. The idea that automation would reduce our collective workload has been out there for years. However there is no incentive for capital to pay labor the same for fewer hours of work.
At my current job we are quite overstaffed and often have lots of free time. However facing economic uncertainty from the Trump administration my boss made sure to emphasize the importance of working 8 hours (even though there are days where we struggle to fill 4 hours). It’s literally absurd but on some level it makes sense from the employers perspective.
The cultural conversation still implicitly assumes lots of full time jobs where people are busy through the day, even though there is already less work today. Obviously AI will accelerate this trend. We probably should have been moving towards more time off, more benefits, etc for decades. I think in that alternate reality people would be less threatened by the emptiness of ‘joblessness’ because our culture would have had lots of time to adapt (more time for unpaid work, childcare, volunteering, hobbies, sports leagues, etc). Instead many cultures (especially the USA) are so work dominated that people can’t imagine an alternative.
Plumbers make their money fixing the pipes in white collar workers houses. We cannot build prosperity out of everyone crowding into a few high paid trade jobs. I’m pretty pessimistic — I think we either see a huge unmitigated spike in unemployment or a ‘bullshit jobs dystopia’ where most people are doing very menial office tasks.
In the long run I think we will be forced to confront this stuff but the basic fact is that employers have no incentive to redistribute income to people unless they provide value and the political discourse is a million miles away from addressing this. I think it’s likely we face a crisis of sorts before anything good comes of this.
Sharing my speculative scenarios on Substack. I've thought about how the use of LLMs for language translation might help knowledge worker weather this storm.
There's one thing that's often overlooked: humans always want things that are hard to get. When AI makes productivity abundant and all products and services become cheap, our needs and desires simply move on to other things that are hard to get. The attention of a stranger, a silent spot in the city, a roundtrip to the moon, two minutes with Oprah, watching every scifi movie ever created, a hamburger from moose meat, a hike passing through all European capital cities, a kiss from a porn star, a wedding at the bottom of the ocean, the honest empathy of a non-machone therapist. There's no end to the things that are hard to get, which means there's no end to human needs and desires. And where there's demand, there is supply. Where there's supply, there is labor.
Love that take
Jurgen you speak truth here. Proverbs 27:20 comes to mind "Just as Death and Destruction are never satisfied, so human desire is never satisfied.” Great comment!
Yes. I’d add that AI challenges not just work, but also identity.
And too often, our response is reactive: save the economy, retrain the workers, don’t fall behind. But what if this moment isn’t just a threat … it’s a test of values? What do we protect, what do we surrender, and who gets to decide?
The real danger isn’t AI … it’s not a good vs evil story … but to your point … what we do with this powerful technology … I’ve been writing a lot in this too. Happy that this framing is coming across other writers
Great article. This quote really rang true: “Here's what puzzles me: Hinton spent 50 years building this technology, then acts surprised by its implications.”
While I appreciate Hinton sounding alarm bells, he did in fact create the monster.
He has forced big tech to finally admit there will be mass layoffs. The new solution is that the tools will make it easy for EVERYONE to become an entrepreneur because the tools will make it so easy. Currently entrepreneurs make up about 16% of the adult workforce.
Right!? Big fat “duh” to Hinton 😉
AI’s Don’t Take Jobs the Rich Do
I’m reaching with a potential solution to the impending jobs crisis that will be knocking at our door with the inevitable arrival of advanced AI and Robotics.
I believe if Australia was to adopt some of my recommendations. We could not only avoid the jobs catastrophe that no one yet has an answer for, but rather set Australia up with a policy that will encourage global adoption as we drive immigration to Australia as that is where the jobs will be.
I will highlight some opportunities definitely worth exploring.
AI don’t take Jobs they get allocated jobs. The Problem itself is quite simple to solve, but will the super wealthy let that solution play out at the loss of abundant wealth?
1.Put AI/Robots on the payroll! Or as least place a monetary value on their contribution. This contribution would scale depending on how many human jobs were taken.
2. AI/Robots would contribute to taxation. If an employer works robots past the standard daily hours/ overtime rates kick in and double over time after that. This stops abusing the protocol.
3. This principle equalises the equation somewhat especially in some industries where the AI/Robot advantage is minimal
4. The income stream to government is maintained.
5. Where a business replaces a worker with an AI/Robot that decision is scrutinised, recordered and the displaced worker receives part of that AI/Robots wage until re- employed.
6. Stress is removed from the welfare system, abundance arrives as discoveries will continue to be made and technologies improved.
7. More than I employee displaced! Apply a multiplier to the above.
8. Employers still win if the do their cost benefit analysis ie: where one AI or Robot displaces an employee the levee applied is one to one but if the robot can optimise the position leading to productivity gain this then benefits the employer the increased productivity flows to GDP and the benefits get realised across all systems.
This is a short / medium term solution to help the planet transverse this period of adjustment. Robots will eventually do everything but we buy some precious time to find a safer landing place.
Kind regards,
Hamish Poppelwell
We spent decades telling kids to go to college so they wouldn’t end up plumbers. Joke’s on us :)
Martin, you are so right and that punchline hits! 😂 But I’ll play peacemaker and say this: a good college education can make it easier to catch on to abstract stuff like AI and tech-so it has it’s sweet spot and advantages. That said, I also know some master tradespeople who could outthink and outwork a few 'tech geniuses' before their morning coffee. Nice comment!
I think I just got really lucky by growing up in Germany going to trade school and then university just to end up in America is a beach bum and then entrepreneur
🤣👍
Maybe AI will create a market for elite human knowledge. Would a consulting client pay the same fee for AI generated out as it would be a partner’s advice who has spent 20 years in an industry specific niche accumulating knowledge?
What do you think of Gary Marcus’s work and the recent Apple white paper critiquing LLMs and how poorly they perform outside the distribution of their input data? Interested readers can start with his “7 replies” post (https://open.substack.com/pub/garymarcus/p/seven-replies-to-the-viral-apple), but can also look at his summary post, the white paper itself, or the follow-up discussing a paper where LLM performance on hard coding tasks dropped to 0%(!) when controlling for data contamination.
A lot to address here. I think this article raises an important point.
I think many people are not ready to address these questions because there are a slew of assumptions most people make about the value and importance of work. You see this across the political spectrum.
In my opinion this story starts a long time before AI. The power of labor started to decline long before AI, in the early 1970s. Many people point to political trends to explain this (I.e. neoliberalism), but I think technological forces have been driving these trends since the beginning. Unionized labor was strong when companies needed many skilled workers all working together in the same factory which was located in the market where the final product was sold. During the Vietnam war new shipping methods were developed. Containerized shipping made outsourcing viable. It was then possible to use cheap overseas labor, which undermined Unions thereafter. Likewise most of the increased productivity per worker over the past 50 years has been driven by computers. Part of why labor hasn’t seen the benefits of this increase in productivity is that there is nothing special about what the workers are doing — the productivity is not coming from labor power as much as computer power, and thus there is no incentive to pay labor more (a.k.a. Anyone can answer emails).
The problem is that it is unhealthy for society for such an imbalance between labor and capital to persist. I’m a big fan of Graeber’s ‘Bullshit Jobs’ and through that book I’m aware of Keynes’ “Economic possibilities for our grandchildren”. The idea that automation would reduce our collective workload has been out there for years. However there is no incentive for capital to pay labor the same for fewer hours of work.
At my current job we are quite overstaffed and often have lots of free time. However facing economic uncertainty from the Trump administration my boss made sure to emphasize the importance of working 8 hours (even though there are days where we struggle to fill 4 hours). It’s literally absurd but on some level it makes sense from the employers perspective.
The cultural conversation still implicitly assumes lots of full time jobs where people are busy through the day, even though there is already less work today. Obviously AI will accelerate this trend. We probably should have been moving towards more time off, more benefits, etc for decades. I think in that alternate reality people would be less threatened by the emptiness of ‘joblessness’ because our culture would have had lots of time to adapt (more time for unpaid work, childcare, volunteering, hobbies, sports leagues, etc). Instead many cultures (especially the USA) are so work dominated that people can’t imagine an alternative.
Plumbers make their money fixing the pipes in white collar workers houses. We cannot build prosperity out of everyone crowding into a few high paid trade jobs. I’m pretty pessimistic — I think we either see a huge unmitigated spike in unemployment or a ‘bullshit jobs dystopia’ where most people are doing very menial office tasks.
In the long run I think we will be forced to confront this stuff but the basic fact is that employers have no incentive to redistribute income to people unless they provide value and the political discourse is a million miles away from addressing this. I think it’s likely we face a crisis of sorts before anything good comes of this.
Hurray!
Sharing my speculative scenarios on Substack. I've thought about how the use of LLMs for language translation might help knowledge worker weather this storm.
https://open.substack.com/pub/daniellepierre/p/la-langue-de-la-prevoyance