Malcom Gladwell, Tyler Cowan and Arnold Kling go into a bar ...
I am all for the promotion of excellence and exceptionalism so I raise a toast to each of these individuals. I'm skeptical, however, any of them has figured out the steps to being the elite in one's field. But each of them provides insights on how an individual can become highly proficient and and enjoy a prosperous life.
All of what they say can be summarized in one sentence:
Have Discipline and Passion, Strive for Excellence and Surround Yourself with Successful People.
Is AI a necessary element? Arnold and Tyler and others say YES. I'm not certain about that. I know a man who is extremely prosperous in life, who built a multi-million dollar business. I have NEVER known this man to own a personal computer or a tablet. He always just owned a phone and a truck. He was successful because he used that phone and vehicle to contact other successful people so that he could get work and grow his business.
On average it does pay off to become proficient in technology, and today AI is one of those technologies. But proficient at what? How hard is it to become proficient at interacting with AI? At what point is "learning AI" like learning BASIC programming? Yea, you do it at some level of education but then you move on to more developing deeper skills and knowledge.
By the way, Arnold, AI is a model. How much should a person base their learning on a model? How much more valuable is it to learn from the real thing?
Yesterday's post pointed out that a certain amount of rote learning is useful for higher level learning. I strongly agree and I worry that will get dropped in the AI learning push.
Average has always been over. The Alpha school approach to learning will not be a panacea for average folks, but it will give wings to students who have the intention to learn. Academia as it is now is a dead man walking.
Does Alpha school teach shop class? I see a future where the pinnacle of society - the people most prized and desired - are plumbers and HVAC technicians.
Pinnacle, not so much. Endless good paying work? Oh yeah. We got that now. I see it even getting better. A lot better. "Better" presupposes high skill and business acumen. (I got 50 years in the trades. It's never been better.)
I'd say yes, in terms of cultivating in students the meta-skill of knowing how to do general task research, how to most efficiency assess a situation, decide if it's worth handling on their own or hiring a professional, and then pointing them in the direction of how to learn what to do, the steps and tools, enough to get started, etc. Kids learning how to really deploy powerful AI tools well will be able to "Be the AI's hands" when asking one to "teach me about my system and the problem then show me how to perform this repair."
I say this as someone who has gotten good at doing plumbing and (some) HVAC repairs on my own with the help of "references" which have gotten better and better, first books; then internet search, forums\,and videos; and now "AI". At this point I usually only run into trouble when (1) I would need expensive, specialized tools that are not worth buying because I would only use them very rarely, (2) Do not get above the 99% threshold of feeling confident I'm taking all important safety precautions with deadly things like electricity or large pieces of glass, (3) Need the kind of "finesse" of motor-skills and proprioception that are hard to acquire without a lot of training, practice, familiarization, and regularly-refreshed direct experiences, (4) The thing I'm working on is non-standard in some way that makes it likely I'll mess things up by following the usual procedures, or (5) the thing was designed intentionally to be hard for amateurs to repair on their own (my impression is that this issue is getting worse over time).
Sometimes it's worth it to have pros do a job just to have a team of some strong guys be able to lug bulky, heavy objects in and out of awkward places. That may be less of an issue soon when lots of capable "Handlers" - humanoid or insectoid robots with fine manipulation ability - are available for rent.
Besides tool cost, the other two things I factor in is labor warranty and regulatory requirements on DIY. I.e. the best AI tutorial in the world isn't going to buy me a car lift, count as a bonded electrician for a new panel (which my local building code requires), nor do I want to risk "learning" on $20k in one time use materials, i.e. the plumber warranties his work at usually a fixed firm price. Modern DIY culture via YouTube has been a net positive for people and the AI will only make it better but I do worry sometimes people go to far in on "I can do anything, I don't need to hire an actual expert" and AI will compound that problem; and yes yes I know, not a modern phenomena.
Centaur chess was a transient phase. Humans are now strictly dominated by AI. You’re right that once AI is ubiquitous, everything you’re describing collapses to a constraint-satisfaction problem—and that’s exactly why the Alpha/Cowen framing fails at the limit. When the solver is cheap and universal, skill, “AI fluency,” and even centaur-style collaboration stop being differentiators. The only thing that matters is who gets to set the constraints and who controls the actuators that make solutions real: compute, capital, platforms, enforcement rails, and signing authority. In that world, “objective selection” is not a higher cognitive layer; it is just another optimization step whose outcome is determined by power, ownership, and access control. Ambition and conscientiousness don’t produce winners; they produce high-output operators inside someone else’s constraint system.
Pushed all the way, this isn’t an education story at all but a political-economic one. As AI collapses performance differentials, advantage concentrates in whoever controls scarce physical resources and the institutional mechanisms that bind plans to action. Whether objectives are chosen by humans or by AI is secondary; what matters is who owns the keys and enforces the constraints. Alpha isn’t training students to win in an AI economy—it’s training them to function efficiently within it. In the limit, sovereignty flows to constraint-setters, not solvers, and pedagogy becomes a sorting mechanism for operators unless it is paired with control over the underlying rails.
“As AI collapses performance differentials, advantage concentrates in whoever controls scarce physical resources and the institutional mechanisms that bind plans to action.”
I suspect your text was written (at least in part) by an LLM.
In any case, I believe that your quoted text is incorrect.
Or more precisely, not more true going forward than in the past.
Of course there is benefit to starting with more resources.
But higher level human attributes (intelligence, work ethic, creativity, willingness to differ from the crowd) will likely have *more* opportunity going forward, not less.
That upside of “Average is over” will not be removed by AI, it will be enhanced.
At least I agree with AK’s take on this. Of course, neither you nor I nor the AI you collaborated with on the above can know for sure; this is merely opinion.
My thesis is that first they will, then they won't. So I suppose the best play is to leverage the brief centaur period to acquire physical assets until the music stops.
I do agree that you are correct about chess, of course.
And your implicit corollary that for certain, even many, tasks this will be true going forward I agree with.
Where we disagree is about human productivity. As machines get better at certain tasks, humans can move along - and net become more productive - to other tasks that machines cannot do better, at least better alone.
Using your music analogy, it’s not that the music stops, but it’s true that it will be ever changing.
Completely consistent with the technological changes of the last few hundred years.
If AI (beynd merely better LLMs) is different, then the chances of it replacing more “middle of the market” jobs more quickly is far higher than previous technologies. But that is very different from replacing those “Average is Over” at the top.
Might that happen in the much further future? Anything of course is possible. But IMO the singularity in the shorter term is imo quite unlikely, and we won’t get there simply from extending LLM technology.
Sure, and as humans move to other tasks, those tasks too will be automated until it's no longer profitable to do so. So you're right back to my central thesis, unless you can point to some hard line that makes valuable tasks impossible to automate in principle. All of this may take a long time, of course, and there's plenty of play in the joints.
Where we disagree is imo there will be ”always” be tasks for humans as well, and they will be economically more productive than before - at least for those “above average”.
I feel like this is in combination with your post yesterday. Technology seems so obviously negative for children I can't fathom someone having a different view. But it might be useful for highly gifted high school students and above.
Is AI going to make my state government honest? Is it going to eliminate the grift and misplaced priorities? Is AI going to pave roads, build housing, fix broken schools, etc?
How is AI going to fix the monopolistic health industry?
The problems we have are a consequence of laws and policy. AI will not change policy!
So how is AI going to be revolutionary? What in society is actually going to change if the powers that be are opposed to change?
But would the Beatles have been better or worse if they had had access to AI?
I dug Fernand Braudel’s civilization & capitalism trilogy out of a shipping box about a year ago to look something up and its been sitting stacked on a corner of the desk ever since as a frequently used reference work. I referred to it again today thinking about this substack post today because it called to mind Braudel’s reference in the second volume to Hubert Bourgin’s 4 category classification of industrial activities in the 15th to 18th centuries. Your LLM is better than mine if it can pull up any information on this, so I’ll have to use Braudel’s summary, which identifies the following classifications:
Category A – tiny family workshops, each with a master-tradesman, 2-3 journeyman, and 1 or 2 apprentices, with groups of workshops usually existing in clusters
Category B – workshops which were scattered, but connected to each other, what Bourgin describes as “dispersed factories” with work directed by a coordinator/merchant entrepreneur who advanced the raw material and saw that it was moved from one stage of processing to another culminating in a finished product which the entrepreneur then marketed
Category C – concentrated industry bringing together, usually under one roof with labor divided among the workforce and direct supervision which created the potential for increased productivity andtsy improved product quality
Category D – factories, which add water/steam/electric to create mechanical energy and machine tools to category C.
So how would AI impact each of these categories?
In category A, AI appears to be being used in the manner Dr. Kling suggests is necessary to avoid being left behind. I wish I new more about the typical Etsy enterprise, but the LLM was at least able to provide the following:
“While there is no single "typical" Etsy enterprise due to the platform’s diverse seller base, research highlights key patterns among successful Etsy sellers. The majority of Etsy sellers are women—81% according to Etsy’s 2020 seller census—and many operate small, niche-focused shops. The most popular product categories include Homeware & Home Furnishings, Jewelry & Personal Accessories, and Craft Supplies.
Top-performing shops often leverage data-driven research tools to identify winning products and optimize listings. Tools like InsightFactory, Alura’s Shop Analyzer, eRank, and EverBee provide real-time data on trending searches, sales volume, competition levels, and keyword performance. These tools help sellers pinpoint low-competition, high-demand niches—especially valuable for print-on-demand (POD) and dropshipping models.
A notable trend is the rise of sellers using AI-powered analytics to forecast demand, analyze reviews, and automate marketing. For example, Sale Samurai offers predictive market data, while Putler and Etshop provide comprehensive shop performance tracking and ad-ready product sourcing.
In terms of revenue, some sellers achieve $10k+ per month, often by focusing on scalable digital products or high-margin physical goods. A documented case saw a seller generate over 1,300 sales in under two months using trend data from InsightFactory.““
So will AI make hyper-niche enterprises more common and economically feasible? Maybe? Might AI enable the US to enjoy a renaissance of specialized shops offering bespoke products at competitive prices? The hostile legal environment would have to be overcome, but, one imagines it is not outside the realm of the possible.
The same might be said of category B as well. And one wonders with the pattern we see of start-ups being bought out, if this is not already a common pattern. The question is whether AI will be able to replace the merchant entrepreneur in identifying sources of inputs to finish and market. AI will be useful, but might the degree of specialization required to produce highly marketable finished goods be enough to maintain a role in the world for merchant entrepreneurs?
In categories C and D, robots and automation it seems will continue to reduce the demand for employees involved in mass production with AI likely to supplant the overhead management and bureaucracy functions. One question might be whether mass-produced products will continue to be able to compete successfully against an AI enabled future of individualized and bespoke products? The legal environment and government regulations offer C and D a variety of advantages but perhaps they will also create the conditions favorable to a thriving informal economy hidden away from the surveillance state.
Manual work will stay underpaid that's for sure. Creative professions and coding/IT are at real risk. While AI is not so advanced (yet) it is formidable just now. Have testing several models for personal coding projects and within 2 months I already have 20+ apps that solve real problems and make life easier. The project and apps are here: https://nonconfirmed.com/app/
Yes it seems the distinction between people who find greater agency by using AI effectively, and those who increasingly find themselves in reactive modes will grow.
"Everyone else will be left behind to some extent."
I think the only reason this is true are the hedge words that are included, "to some extent."
For sure the skill sets that "win" will change. But the optimal skill sets have been changing for hundreds of years, arguably at an increasing pace of change. Who has those new optimal will always be changing.
Regardless of AI opportunities, there will still be some need for workers doing lots of jobs that may never require much AI expertise, even if they use AI. These might include low skill mundane jobs including cleaning, lawn care, care for elderly and disabled, etc., people who interact with customers and clients, managers, executives, builders, repairers, and surely plenty more.
Can someone recommend a book, a course, an app or a blog to teach myself how to use AI in my life? I use AI sometimes to make decisions. I can't say it's revolutionary in my life.
Youths who have outstanding "internal pressure" tend to find their ways to success (even despite great flaws in highly selective universities). Doubtless, AI technology shock will enable many 'driven' youths to achieve more, faster — citius, altius, fortius.
Might some of these talented, motivated youths figure out ways to motivate the rest, or to increase the productivity of the rest? Is there no prospect that AI entrepreneurship will produce a rising tide that lifts most boats?
The productivity gains that AI will deliver will indeed lift all boats in terms of economic wealth. And so this very much includes the median standard of living.
But that is different than the “inequality” of productivity of those with top-end skills versus those in the middle.
The next gen Elon Musks of the world will very likely cause the per person productivity of his/her employees to be a lot higher, yes.
Malcom Gladwell, Tyler Cowan and Arnold Kling go into a bar ...
I am all for the promotion of excellence and exceptionalism so I raise a toast to each of these individuals. I'm skeptical, however, any of them has figured out the steps to being the elite in one's field. But each of them provides insights on how an individual can become highly proficient and and enjoy a prosperous life.
All of what they say can be summarized in one sentence:
Have Discipline and Passion, Strive for Excellence and Surround Yourself with Successful People.
Is AI a necessary element? Arnold and Tyler and others say YES. I'm not certain about that. I know a man who is extremely prosperous in life, who built a multi-million dollar business. I have NEVER known this man to own a personal computer or a tablet. He always just owned a phone and a truck. He was successful because he used that phone and vehicle to contact other successful people so that he could get work and grow his business.
On average it does pay off to become proficient in technology, and today AI is one of those technologies. But proficient at what? How hard is it to become proficient at interacting with AI? At what point is "learning AI" like learning BASIC programming? Yea, you do it at some level of education but then you move on to more developing deeper skills and knowledge.
By the way, Arnold, AI is a model. How much should a person base their learning on a model? How much more valuable is it to learn from the real thing?
Agreed there will always be plenty of alternatives to reach success and even also to wealth.
Yesterday's post pointed out that a certain amount of rote learning is useful for higher level learning. I strongly agree and I worry that will get dropped in the AI learning push.
Average has always been over. The Alpha school approach to learning will not be a panacea for average folks, but it will give wings to students who have the intention to learn. Academia as it is now is a dead man walking.
This is basically my opinion as well. Some people have always excelled at using new tools and many/most don't.
Does Alpha school teach shop class? I see a future where the pinnacle of society - the people most prized and desired - are plumbers and HVAC technicians.
Pinnacle, not so much. Endless good paying work? Oh yeah. We got that now. I see it even getting better. A lot better. "Better" presupposes high skill and business acumen. (I got 50 years in the trades. It's never been better.)
Probably not the pinnacle but it seems they almost certainly won't be left behind.
I'd say yes, in terms of cultivating in students the meta-skill of knowing how to do general task research, how to most efficiency assess a situation, decide if it's worth handling on their own or hiring a professional, and then pointing them in the direction of how to learn what to do, the steps and tools, enough to get started, etc. Kids learning how to really deploy powerful AI tools well will be able to "Be the AI's hands" when asking one to "teach me about my system and the problem then show me how to perform this repair."
I say this as someone who has gotten good at doing plumbing and (some) HVAC repairs on my own with the help of "references" which have gotten better and better, first books; then internet search, forums\,and videos; and now "AI". At this point I usually only run into trouble when (1) I would need expensive, specialized tools that are not worth buying because I would only use them very rarely, (2) Do not get above the 99% threshold of feeling confident I'm taking all important safety precautions with deadly things like electricity or large pieces of glass, (3) Need the kind of "finesse" of motor-skills and proprioception that are hard to acquire without a lot of training, practice, familiarization, and regularly-refreshed direct experiences, (4) The thing I'm working on is non-standard in some way that makes it likely I'll mess things up by following the usual procedures, or (5) the thing was designed intentionally to be hard for amateurs to repair on their own (my impression is that this issue is getting worse over time).
Sometimes it's worth it to have pros do a job just to have a team of some strong guys be able to lug bulky, heavy objects in and out of awkward places. That may be less of an issue soon when lots of capable "Handlers" - humanoid or insectoid robots with fine manipulation ability - are available for rent.
Besides tool cost, the other two things I factor in is labor warranty and regulatory requirements on DIY. I.e. the best AI tutorial in the world isn't going to buy me a car lift, count as a bonded electrician for a new panel (which my local building code requires), nor do I want to risk "learning" on $20k in one time use materials, i.e. the plumber warranties his work at usually a fixed firm price. Modern DIY culture via YouTube has been a net positive for people and the AI will only make it better but I do worry sometimes people go to far in on "I can do anything, I don't need to hire an actual expert" and AI will compound that problem; and yes yes I know, not a modern phenomena.
"That either made you very good if you had what it took to be good, or you gave up"
Plenty of people do activities without expertise, never getting better, but enjoy it immensely.
"They were determined to make the most of their abilities."
This is closer to truth. It still takes great talent to reach their level as well as a bit of determined practice.
Centaur chess was a transient phase. Humans are now strictly dominated by AI. You’re right that once AI is ubiquitous, everything you’re describing collapses to a constraint-satisfaction problem—and that’s exactly why the Alpha/Cowen framing fails at the limit. When the solver is cheap and universal, skill, “AI fluency,” and even centaur-style collaboration stop being differentiators. The only thing that matters is who gets to set the constraints and who controls the actuators that make solutions real: compute, capital, platforms, enforcement rails, and signing authority. In that world, “objective selection” is not a higher cognitive layer; it is just another optimization step whose outcome is determined by power, ownership, and access control. Ambition and conscientiousness don’t produce winners; they produce high-output operators inside someone else’s constraint system.
Pushed all the way, this isn’t an education story at all but a political-economic one. As AI collapses performance differentials, advantage concentrates in whoever controls scarce physical resources and the institutional mechanisms that bind plans to action. Whether objectives are chosen by humans or by AI is secondary; what matters is who owns the keys and enforces the constraints. Alpha isn’t training students to win in an AI economy—it’s training them to function efficiently within it. In the limit, sovereignty flows to constraint-setters, not solvers, and pedagogy becomes a sorting mechanism for operators unless it is paired with control over the underlying rails.
“As AI collapses performance differentials, advantage concentrates in whoever controls scarce physical resources and the institutional mechanisms that bind plans to action.”
I suspect your text was written (at least in part) by an LLM.
In any case, I believe that your quoted text is incorrect.
Or more precisely, not more true going forward than in the past.
Of course there is benefit to starting with more resources.
But higher level human attributes (intelligence, work ethic, creativity, willingness to differ from the crowd) will likely have *more* opportunity going forward, not less.
That upside of “Average is over” will not be removed by AI, it will be enhanced.
At least I agree with AK’s take on this. Of course, neither you nor I nor the AI you collaborated with on the above can know for sure; this is merely opinion.
My thesis is that first they will, then they won't. So I suppose the best play is to leverage the brief centaur period to acquire physical assets until the music stops.
I do agree that you are correct about chess, of course.
And your implicit corollary that for certain, even many, tasks this will be true going forward I agree with.
Where we disagree is about human productivity. As machines get better at certain tasks, humans can move along - and net become more productive - to other tasks that machines cannot do better, at least better alone.
Using your music analogy, it’s not that the music stops, but it’s true that it will be ever changing.
Completely consistent with the technological changes of the last few hundred years.
If AI (beynd merely better LLMs) is different, then the chances of it replacing more “middle of the market” jobs more quickly is far higher than previous technologies. But that is very different from replacing those “Average is Over” at the top.
Might that happen in the much further future? Anything of course is possible. But IMO the singularity in the shorter term is imo quite unlikely, and we won’t get there simply from extending LLM technology.
Sure, and as humans move to other tasks, those tasks too will be automated until it's no longer profitable to do so. So you're right back to my central thesis, unless you can point to some hard line that makes valuable tasks impossible to automate in principle. All of this may take a long time, of course, and there's plenty of play in the joints.
Where we disagree is imo there will be ”always” be tasks for humans as well, and they will be economically more productive than before - at least for those “above average”.
Sure. That can be true even as in absolute terms the return from human labor tends to zero.
I feel like this is in combination with your post yesterday. Technology seems so obviously negative for children I can't fathom someone having a different view. But it might be useful for highly gifted high school students and above.
Is AI going to make my state government honest? Is it going to eliminate the grift and misplaced priorities? Is AI going to pave roads, build housing, fix broken schools, etc?
How is AI going to fix the monopolistic health industry?
The problems we have are a consequence of laws and policy. AI will not change policy!
So how is AI going to be revolutionary? What in society is actually going to change if the powers that be are opposed to change?
But would the Beatles have been better or worse if they had had access to AI?
I dug Fernand Braudel’s civilization & capitalism trilogy out of a shipping box about a year ago to look something up and its been sitting stacked on a corner of the desk ever since as a frequently used reference work. I referred to it again today thinking about this substack post today because it called to mind Braudel’s reference in the second volume to Hubert Bourgin’s 4 category classification of industrial activities in the 15th to 18th centuries. Your LLM is better than mine if it can pull up any information on this, so I’ll have to use Braudel’s summary, which identifies the following classifications:
Category A – tiny family workshops, each with a master-tradesman, 2-3 journeyman, and 1 or 2 apprentices, with groups of workshops usually existing in clusters
Category B – workshops which were scattered, but connected to each other, what Bourgin describes as “dispersed factories” with work directed by a coordinator/merchant entrepreneur who advanced the raw material and saw that it was moved from one stage of processing to another culminating in a finished product which the entrepreneur then marketed
Category C – concentrated industry bringing together, usually under one roof with labor divided among the workforce and direct supervision which created the potential for increased productivity andtsy improved product quality
Category D – factories, which add water/steam/electric to create mechanical energy and machine tools to category C.
So how would AI impact each of these categories?
In category A, AI appears to be being used in the manner Dr. Kling suggests is necessary to avoid being left behind. I wish I new more about the typical Etsy enterprise, but the LLM was at least able to provide the following:
“While there is no single "typical" Etsy enterprise due to the platform’s diverse seller base, research highlights key patterns among successful Etsy sellers. The majority of Etsy sellers are women—81% according to Etsy’s 2020 seller census—and many operate small, niche-focused shops. The most popular product categories include Homeware & Home Furnishings, Jewelry & Personal Accessories, and Craft Supplies.
Top-performing shops often leverage data-driven research tools to identify winning products and optimize listings. Tools like InsightFactory, Alura’s Shop Analyzer, eRank, and EverBee provide real-time data on trending searches, sales volume, competition levels, and keyword performance. These tools help sellers pinpoint low-competition, high-demand niches—especially valuable for print-on-demand (POD) and dropshipping models.
A notable trend is the rise of sellers using AI-powered analytics to forecast demand, analyze reviews, and automate marketing. For example, Sale Samurai offers predictive market data, while Putler and Etshop provide comprehensive shop performance tracking and ad-ready product sourcing.
In terms of revenue, some sellers achieve $10k+ per month, often by focusing on scalable digital products or high-margin physical goods. A documented case saw a seller generate over 1,300 sales in under two months using trend data from InsightFactory.““
So will AI make hyper-niche enterprises more common and economically feasible? Maybe? Might AI enable the US to enjoy a renaissance of specialized shops offering bespoke products at competitive prices? The hostile legal environment would have to be overcome, but, one imagines it is not outside the realm of the possible.
The same might be said of category B as well. And one wonders with the pattern we see of start-ups being bought out, if this is not already a common pattern. The question is whether AI will be able to replace the merchant entrepreneur in identifying sources of inputs to finish and market. AI will be useful, but might the degree of specialization required to produce highly marketable finished goods be enough to maintain a role in the world for merchant entrepreneurs?
In categories C and D, robots and automation it seems will continue to reduce the demand for employees involved in mass production with AI likely to supplant the overhead management and bureaucracy functions. One question might be whether mass-produced products will continue to be able to compete successfully against an AI enabled future of individualized and bespoke products? The legal environment and government regulations offer C and D a variety of advantages but perhaps they will also create the conditions favorable to a thriving informal economy hidden away from the surveillance state.
Or perhaps, that is all just wishful thinking.
Manual work will stay underpaid that's for sure. Creative professions and coding/IT are at real risk. While AI is not so advanced (yet) it is formidable just now. Have testing several models for personal coding projects and within 2 months I already have 20+ apps that solve real problems and make life easier. The project and apps are here: https://nonconfirmed.com/app/
Yes it seems the distinction between people who find greater agency by using AI effectively, and those who increasingly find themselves in reactive modes will grow.
"Everyone else will be left behind to some extent."
I think the only reason this is true are the hedge words that are included, "to some extent."
For sure the skill sets that "win" will change. But the optimal skill sets have been changing for hundreds of years, arguably at an increasing pace of change. Who has those new optimal will always be changing.
Regardless of AI opportunities, there will still be some need for workers doing lots of jobs that may never require much AI expertise, even if they use AI. These might include low skill mundane jobs including cleaning, lawn care, care for elderly and disabled, etc., people who interact with customers and clients, managers, executives, builders, repairers, and surely plenty more.
Can someone recommend a book, a course, an app or a blog to teach myself how to use AI in my life? I use AI sometimes to make decisions. I can't say it's revolutionary in my life.
Try asking an AI. If nothing else, the process of learning how to refine your questions to get good answers will be a good self-teaching exercise.
ETA: This is not a smartass answer. It's the real deal. You don't learn arithmetic tables by reading blogs or textbooks, but by doing arithmetic.
Youths who have outstanding "internal pressure" tend to find their ways to success (even despite great flaws in highly selective universities). Doubtless, AI technology shock will enable many 'driven' youths to achieve more, faster — citius, altius, fortius.
Might some of these talented, motivated youths figure out ways to motivate the rest, or to increase the productivity of the rest? Is there no prospect that AI entrepreneurship will produce a rising tide that lifts most boats?
Imo these are entirely different points.
The productivity gains that AI will deliver will indeed lift all boats in terms of economic wealth. And so this very much includes the median standard of living.
But that is different than the “inequality” of productivity of those with top-end skills versus those in the middle.
The next gen Elon Musks of the world will very likely cause the per person productivity of his/her employees to be a lot higher, yes.
Most boats are not the same; most boats are average boats. Most boats are not ocean-going boats.