This weekend I listened to your post with my 18-year-old son, and I wanted to share a challenge to your article. I feel your piece views the innovation of AI through a pre-AI lens. I agree—the line is extremely thin and blurry—but can we start shifting our perspective, focusing on how we should adapt to the future rather than lamenting how the future is disrupting the present?
I challenge anyone who believes we can opt out of the AI revolution. Why should we avoid it or renounce its benefits? Yes, AI is a major shock to our cognitive uniqueness and can be dangerous. Many disruptive innovations throughout history have threatened the status quo: the early bushman had better spatial cognitive skills than us, the industrial revolution eliminated jobs in manual textile spinning and weaving; automobiles replaced the need for urban stables and blacksmiths; computers and the internet phased out manual bookkeeping and typewriters. Now, AI is already replacing call center operators, paralegals, and basic software coding. It was never a zero-sum game. Yet we adapted—losing some things, gaining others—and called it progress.
The current skills gap lies in knowing how to use AI to enhance performance, rather than improving obsolete skills that are bound to be replaced.
A positive approach would focus on two main priorities:
1. Educational AI – teaching how to use AI as an enhancement tool, not as a substitute for our thinking.
It should become a core subject: learning how to drive AI instead of being driven by it. Most current articles depict a doom-and-gloom future, which only widens the fear gap and fuels a vicious cycle.
We should teach students how AI can take over tasks where critical reasoning isn't essential, freeing the brain to focus on higher-level thinking. For instance, in writing a lab report, don’t let AI generate the results outright (or generate an income map, as in Gary Marcus’s article). Instead, have students set up the principles and parameters for the experiment manually (starting with Boolean logic: AND, OR, IF), collect test data, and then use AI algorhytm to detect patterns—leaving interpretation and conclusions to human reasoning. In Business delegate all Standard Operation Processes (SOP) to AI interaction freeing up employees tasks from low impact ones. This way, queries are framed within a controlled environment based on word documents instructions translated into code by the AI and allowing it to thrive without overreaching. The common misuse of AI assumes it will replicate human logic or common sense in a vast sea of data—this is a probabilistic impossibility(several articles ref.1-4). It requires an understanding of how AI operates, the principles of algorithmic flowcharts, and the overarching logic of human reasoning.
Another clear example is the criticism of driver assistance systems or autopilot. When things go wrong, autopilot gets the blame. But we should instead examine our own role. Autopilot is just a tool—we are the ones using it. Imagine training aircraft pilots without ever teaching them how to operate the autopilot, leaving it up to them to figure out. Without proper education for operators, we will continue to see accidents increase rather than achieving the safety levels of the aviation industry. The choice between a vicious or virtuous cycle is ours.
2. Pragmatic AI – We should honestly assess which current human activities underutilize our cognitive potential. Using the brain for repetitive or supervisory tasks is a waste of mental power. We need to move past nostalgic ideals, like having a human sell movie tickets or scan barcodes all day at a supermarket.
We must be honest about the daily tasks that involve little to no critical thinking. I'm sure everyone has some of these and finds them frustrating. Imagine, if instantly all these demotivating tasks are handled by AI exactly as you’d like, freeing you to focus on what you love and what motivates you.
We must accept that some human activities can and should be done by AI-powered machines in exchange for more time spent on tasks that truly benefit from our cognitive abilities. If the human brain capacity is roughly equal across individuals, then using the same brain for delicate surgery and for clipping museum tickets is a monumental waste of potential.
The distinction between blue-collar and white-collar jobs is becoming less relevant. AI, when seen as an enhancement tool, cuts across all categories—every job has tasks that are too basic for our cognitive capacity. Our thinking should shift from "What job will AI replace?" to "Which simple or low-value tasks should AI be doing?"
The value added to each job could even rise, making work simpler and more efficient overall.(analysis on increased efficiency with AI ref.5-9)
However, your analysis provides extreme good insights regarding the short term impact on labor force and the future impact of the development of AI into AGI, hence we should be worried to understand better AI, probably the final question is, if short term losses are worth endure to understand how to cohabitate?
"If we want things to stay as they are, things will have to change." (Ref. Il Gattopardo)
Dear Paolo, thank you very much for this very high-value added comment. There is a lot here that I do agree with, and I need more time to craft a reaction that will do it justice, but in the meanwhile I wanted to thank you both for your attention and for this highly valuable contribution to the conversation -- thanks extended to your son as well, of course. More to follow -- Marco
p.s. I particularly appreciate the Gattopardo reference
Great post Marco. I think making manufacturing more attractive for young people is less hard than it seems.
The first obvious thing is that nowadays production hardly happens in the sooty factories described Dickens or in those hyper-dehumanizing conditions described by Charlie Chaplin in Modern Times. I recently collaborated with Autodesk and visited many of their fabrication labs where human creativity is unleashed as machines take away the negative mental space of repetitive work from them. It was exciting and immediately triggered into me an instinct to create and make something.
I think that as you reflect on what can be done - for example in education - to generate the drive to making things, one solution is to wash away from FabLabs, makerspaces, hackspaces, and repair cafes their not-so-subtle leftist-niche coating and preference for Etsy over large scale commercialization. There are unsuspected pleasures fostered by making things - even more when its done together. Co-creation is ideally super-powered by internet and it already works so well in the production and co-production of software through various platforms. We should create the GitHub of the making.
I strongly recommend to anybody this documentary on the making of a Steinway & Sons piano: https://www.youtube.com/watch?v=6rAhps4AkT8. If you are in a hurry please skip to 18:20 - the short interview to Eddie Salvadon, a plate fitter. Pause and reflect at his story of the concert he attended at Carnegie Hall some weeks before. Eddie should be invited everywhere to share to young people what making things really is.
Among the several damages done by us economists is the dehumanization of the concept of value added, reduced to a number, a percentage. Value added is transferring part of our soul into something we contribute making - a part of the soul which remains embedded in the object and travels with it across places and times; and that we always recognize as ours. This, Eddie Salvadon should share with the young generations.
Thanks Luca, great points. In fact, what you point out is that thanks also to digital technologies, manufacturing is beginning to undergo another transformation, one which has the potential to redefine economies of scale, bring more flexibility to its geographic distribution, and enhance the power of collaboration.
There is also, I think, a profound observation underneath all this. Our interaction with the physical world, including the act of making things, is an inextricable, fundamental part of our intelligence, and an important reason why Artificial General Intelligence remains elusive. We should cultivate it.
Another terrific article Marco! I really enjoyed the hallucinogenic post by Gary Marcus that you linked to! Hilarious.
But one thing I have definitely seen, at least in India, is the very rapid dependence that graduate students have developed to Chat GPT for submitting their assignments! It's crazy. Professors are tearing their hair out trying to check this fever but it's just catching on faster and faster. It's definitely going to have an impact on the future labour force in terms of actual usable skills. And in a developing India, less anv less people actually want to do anything that uses their hands! Everyone wants to be a Chat GPT assisted brainiac!😃
Thanks Bobby. Your observation is perhaps not surprising, given that India has already been trying to jump straight onto the software/service growth track. We keep boasting that what sets us aside from other species is our ability to learn and think, and now we are so eager to abandon it. If AI truly surpasses our cognitive abilities it will be less because AI has improved and more because human intelligence has deteriorated. And India definitely can't afford it, not yet at least...
Dear Marco, apologies for the very long comment,
This weekend I listened to your post with my 18-year-old son, and I wanted to share a challenge to your article. I feel your piece views the innovation of AI through a pre-AI lens. I agree—the line is extremely thin and blurry—but can we start shifting our perspective, focusing on how we should adapt to the future rather than lamenting how the future is disrupting the present?
I challenge anyone who believes we can opt out of the AI revolution. Why should we avoid it or renounce its benefits? Yes, AI is a major shock to our cognitive uniqueness and can be dangerous. Many disruptive innovations throughout history have threatened the status quo: the early bushman had better spatial cognitive skills than us, the industrial revolution eliminated jobs in manual textile spinning and weaving; automobiles replaced the need for urban stables and blacksmiths; computers and the internet phased out manual bookkeeping and typewriters. Now, AI is already replacing call center operators, paralegals, and basic software coding. It was never a zero-sum game. Yet we adapted—losing some things, gaining others—and called it progress.
The current skills gap lies in knowing how to use AI to enhance performance, rather than improving obsolete skills that are bound to be replaced.
A positive approach would focus on two main priorities:
1. Educational AI – teaching how to use AI as an enhancement tool, not as a substitute for our thinking.
It should become a core subject: learning how to drive AI instead of being driven by it. Most current articles depict a doom-and-gloom future, which only widens the fear gap and fuels a vicious cycle.
We should teach students how AI can take over tasks where critical reasoning isn't essential, freeing the brain to focus on higher-level thinking. For instance, in writing a lab report, don’t let AI generate the results outright (or generate an income map, as in Gary Marcus’s article). Instead, have students set up the principles and parameters for the experiment manually (starting with Boolean logic: AND, OR, IF), collect test data, and then use AI algorhytm to detect patterns—leaving interpretation and conclusions to human reasoning. In Business delegate all Standard Operation Processes (SOP) to AI interaction freeing up employees tasks from low impact ones. This way, queries are framed within a controlled environment based on word documents instructions translated into code by the AI and allowing it to thrive without overreaching. The common misuse of AI assumes it will replicate human logic or common sense in a vast sea of data—this is a probabilistic impossibility(several articles ref.1-4). It requires an understanding of how AI operates, the principles of algorithmic flowcharts, and the overarching logic of human reasoning.
Another clear example is the criticism of driver assistance systems or autopilot. When things go wrong, autopilot gets the blame. But we should instead examine our own role. Autopilot is just a tool—we are the ones using it. Imagine training aircraft pilots without ever teaching them how to operate the autopilot, leaving it up to them to figure out. Without proper education for operators, we will continue to see accidents increase rather than achieving the safety levels of the aviation industry. The choice between a vicious or virtuous cycle is ours.
2. Pragmatic AI – We should honestly assess which current human activities underutilize our cognitive potential. Using the brain for repetitive or supervisory tasks is a waste of mental power. We need to move past nostalgic ideals, like having a human sell movie tickets or scan barcodes all day at a supermarket.
We must be honest about the daily tasks that involve little to no critical thinking. I'm sure everyone has some of these and finds them frustrating. Imagine, if instantly all these demotivating tasks are handled by AI exactly as you’d like, freeing you to focus on what you love and what motivates you.
We must accept that some human activities can and should be done by AI-powered machines in exchange for more time spent on tasks that truly benefit from our cognitive abilities. If the human brain capacity is roughly equal across individuals, then using the same brain for delicate surgery and for clipping museum tickets is a monumental waste of potential.
The distinction between blue-collar and white-collar jobs is becoming less relevant. AI, when seen as an enhancement tool, cuts across all categories—every job has tasks that are too basic for our cognitive capacity. Our thinking should shift from "What job will AI replace?" to "Which simple or low-value tasks should AI be doing?"
The value added to each job could even rise, making work simpler and more efficient overall.(analysis on increased efficiency with AI ref.5-9)
However, your analysis provides extreme good insights regarding the short term impact on labor force and the future impact of the development of AI into AGI, hence we should be worried to understand better AI, probably the final question is, if short term losses are worth endure to understand how to cohabitate?
"If we want things to stay as they are, things will have to change." (Ref. Il Gattopardo)
Dear Paolo, thank you very much for this very high-value added comment. There is a lot here that I do agree with, and I need more time to craft a reaction that will do it justice, but in the meanwhile I wanted to thank you both for your attention and for this highly valuable contribution to the conversation -- thanks extended to your son as well, of course. More to follow -- Marco
p.s. I particularly appreciate the Gattopardo reference
Great post Marco. I think making manufacturing more attractive for young people is less hard than it seems.
The first obvious thing is that nowadays production hardly happens in the sooty factories described Dickens or in those hyper-dehumanizing conditions described by Charlie Chaplin in Modern Times. I recently collaborated with Autodesk and visited many of their fabrication labs where human creativity is unleashed as machines take away the negative mental space of repetitive work from them. It was exciting and immediately triggered into me an instinct to create and make something.
I think that as you reflect on what can be done - for example in education - to generate the drive to making things, one solution is to wash away from FabLabs, makerspaces, hackspaces, and repair cafes their not-so-subtle leftist-niche coating and preference for Etsy over large scale commercialization. There are unsuspected pleasures fostered by making things - even more when its done together. Co-creation is ideally super-powered by internet and it already works so well in the production and co-production of software through various platforms. We should create the GitHub of the making.
I strongly recommend to anybody this documentary on the making of a Steinway & Sons piano: https://www.youtube.com/watch?v=6rAhps4AkT8. If you are in a hurry please skip to 18:20 - the short interview to Eddie Salvadon, a plate fitter. Pause and reflect at his story of the concert he attended at Carnegie Hall some weeks before. Eddie should be invited everywhere to share to young people what making things really is.
Among the several damages done by us economists is the dehumanization of the concept of value added, reduced to a number, a percentage. Value added is transferring part of our soul into something we contribute making - a part of the soul which remains embedded in the object and travels with it across places and times; and that we always recognize as ours. This, Eddie Salvadon should share with the young generations.
Thanks Luca, great points. In fact, what you point out is that thanks also to digital technologies, manufacturing is beginning to undergo another transformation, one which has the potential to redefine economies of scale, bring more flexibility to its geographic distribution, and enhance the power of collaboration.
There is also, I think, a profound observation underneath all this. Our interaction with the physical world, including the act of making things, is an inextricable, fundamental part of our intelligence, and an important reason why Artificial General Intelligence remains elusive. We should cultivate it.
Another terrific article Marco! I really enjoyed the hallucinogenic post by Gary Marcus that you linked to! Hilarious.
But one thing I have definitely seen, at least in India, is the very rapid dependence that graduate students have developed to Chat GPT for submitting their assignments! It's crazy. Professors are tearing their hair out trying to check this fever but it's just catching on faster and faster. It's definitely going to have an impact on the future labour force in terms of actual usable skills. And in a developing India, less anv less people actually want to do anything that uses their hands! Everyone wants to be a Chat GPT assisted brainiac!😃
Thanks Bobby. Your observation is perhaps not surprising, given that India has already been trying to jump straight onto the software/service growth track. We keep boasting that what sets us aside from other species is our ability to learn and think, and now we are so eager to abandon it. If AI truly surpasses our cognitive abilities it will be less because AI has improved and more because human intelligence has deteriorated. And India definitely can't afford it, not yet at least...