The advance team of innovators In every industry and every field need to clean the whiteboards and think through how to rebuild their worlds based on the new foundation of intelligent machines
I started reading by myself when I was 4. My brother never enjoyed reading the way I did. Public education was easy for me and a struggle for my brother. To this day my love of learning has centered around my ability to read. Just imagine a world where every child was tutored to read (and were taught to comprehend what they read) at a very early age. The human potential would be unprecedented.
Addressing just the application of AI to our educational system, it sounds great in theory, but it will cost money. Many families struggle to survive financially these days. How would this be implemented to all children equally? Government is cutting free lunches to hungry kids so I can’t imagine that they would be willing to make sure that even poorer families would get the same resources as those families with plenty of disposable income. I’m not arguing against it, just feel it is idealistic & not practical at this time. We need more compassion in our society.
Hi Jean. What if it costs next to nothing? What if all it requires is computer access (which almost every kid already has) and a handful of educators who link AI tutoring agents (which already exist) to each kids present level of knowledge in any - or every - subject area? If such is already or even mostly the case, significant compassion is a few clicks away.
I put the following 2 (messy) prompts into Claude to get a sliver of a sense of what’s going on currently. Incomplete though it is, try doing the queries yourself or make up better ones (low bar). (;-)
A) “I believe the Khan academy uses AI tutoring agents to teach kids? How successful are they in taking kids at lower end of bell curve of knowledge for their age, up to much higher end? How do they do it? Is it expensive or could any parent or teacher implement it at home or in a conventional classroom?”
B) “Are there other similar AI agents for K to 12 education? Or companies or start-ups developing such agents? And, has anyone demonstrated that the more successful, easier to use, and less expensive AI educational agents are based on active inference (think Karl Friston) models that don’t require loading a huge data base into the AI tutor?”
It costs next to nothing to give kids "free" lunch. The USA could have solved food insecurity in their own country at any point in the last 50 years and they chose not to. A hungry child can't learn effectively with or without a 1-1 tutor.
The USA as a society doesn't value equality and the tech-illiterate politicians don't have the skills to guarantee an equal education with AI tutors even if they had the will power and financial backing to do so.
Sure, go ahead and improve AI systems and create AI tutors but don't expect the end result will be anything resembling equality.
Equality: the state of being equal, especially in status, rights, and opportunities. For example, do all school aged children have the opportunity to access advanced AI models? Do all school aged children have the opportunity to take classes to learn how to use computers.
In my original comment I am stating the USA doesn't value equality so you will never see a day when all school aged children have access to computers and advanced AI tutors.
Hence “optimize the potential for each child” to quote myself. The point is that individualized AI tutors wouldn’t just be better, they would be cheaper than current education practices and thus better and more affordable for poorer neighborhoods.
By the way, I believe that national school spending is higher today for lower quintile households than it is for the median. So it really isn’t a given that we are currently short-changing poor students.
Hence “optimize the potential for each child” to quote myself. The point is that individualized AI tutors wouldn’t just be better, they would be cheaper than current education practices and thus better and more affordable for poorer neighborhoods.
By the way, I believe that national school spending is higher today for lower quintile households than it is for the median. So it really isn’t a given that we are currently short-changing poor students.
It seems to me that an AI tutor would be next to free. The cost in software is in the first unit, not the hundred-millionth. A better-designed, built up from scratch system with private tutors wouldn’t cost more than today’s schools, it would cost less, and be incomparably better.
Interesting as always. thanks for putting a stake in the ground. it gives us food for thought.
On the innovation curve diagram, the "tipping point" is also known as the "chasm" and while AI has seemingly (early majority is just beginning) made the jump across the chasm, most other industries have not begun.
i understand your advocating for every industry to hit restart. that said, industries like insurance, banking, and investing cannot do more than regulators will allow and regulators don't like "black box" technologies... of course, in addition to finance, the fed also regulates health, medicine, transportation, communications, agriculture, and energy... not much left. what the fed doesn't regulate, the states do regulate. regional and national companies have to deal with multiple states and that can be fun... 10 years after the rest of the world was using "square tape" (IBM 3480's), the State of Massachusetts still required regulatory data be sent to them on "round reels" (IBM 3420's), so my employer had about fifty 3480 tape drives, and one 3420 tape drive...
as far as rolling out AI goes, regulators aren't the only roadblocks. consumers set the pace. you can't go faster than the early majority, and that leaves out a lot of customers (that the regulators will protect - including the laggards).
one other thing. consolidation. as a former competitive intelligence analyst, i ask, who will win? every industry has consolidated in the past 200 years - with help from the fed which likes 3 to 7 competitors in each space. so, which providers will be standing in 5 or 10 years. i remember numerous start ups that were initially successful losing to today's juggernauts (My Space, Yahoo!, Blackberry, Netscape Navigator, Watson, Ask Jeeves, et al). In fact, there is a strong argument supporting that the first to market almost always loses... I ask this because most current companies will try to wait for the dust to settle before committing to a company that may or may not be around in 5 years...
i agree the purpose of education was to create manufacturing workers (write well enough to report accomplishments and problems, read enough to follow a procedure, enough math to count what you produced, etc.).
You lay out a lot of ideas here, as you often do. Keep them coming. My quick hit is that the economic implications of all this are just starting to dawn on those currently running and leading the economy. I will have more to say about all this in future essays. But the first step is getting people to reorient to how big a deal this is and how thinking fresh and starting from scratch is the best way to approach a seismic change of this sort.
Thank you Peter. In my own observations the commercialization and availability of AI represents a societal inflection point similar to the printing press, the telegraph, the telephone, the microscope, the automobile, and flight.
There must be a corresponding and immediate understanding of the impact of widespread AI use on human labor demand. Without a complete reimagining of economic systems, AI will tend to concentrate even more wealth -beyond the already extreme state which exists in our post-Reagan society today -and is antithetical to any meaningful democracy in the US and anywhere in the world. A dystopian authoritarian future controlled by Thiel, Musk, Zuckerberg, and a handful of others, already on record as to their thoughts about human empathy and the average person trying to live a decent life (as nothing more than a participant in a video game) is not a great progression for the future.
George, I hear you and am aligned with your thinking. A bit part of The Great Progression series and ultimate book will include much on that topic of the need for a fundamental rework of the economy as we make this transition. This was not the essay to do that, but I have one coming out in the next couple weeks that will.
In fact, just yesterday I ran a roundtable with about a dozen people focused on the bigger economic implications in a day long workshop on the arrival of AI agents. And tomorrow I am going to a gathering called "Reimagining the Economy in the Age of AI." Many people in the SF Bay Area are getting focused on that. It's encouraging, but we have a long way to go.
Thanks @Peter Leyden. We both know policy and legislation lag far behind technological evolution. Not that I would suggest putting the toothpaste back in the tube (and the promise of AI dates back to the beginnings of my early use of UNIX) so it’s not like AI is an overnight sensation. Even a few years of diminished human labor demand can shift concentrated wealth well beyond a tipping point for democracy (granted it may already be there). Glad economics has a seat at the strategy table and is in the forefront of your thinking.
funny, "there must be a corresponding and immediate understanding...". the items you list "the printing press, the telegraph, the telephone, the microscope, the automobile, and flight." did not have a corresponding or immediate understanding... the beloved federal government is rarely ahead of the public except in intelligence and defense.
the concentration of wealth that you speak of is a reflection of the value added by those providers receiving the wealth - surely your household doesn't use google, facebook, amazon, apple, or microsoft. people voluntarily transfer their money to the providers that offer them the best value. the ability to scale virtually instead of physically is what has led to the growing gap in wealth. surely your household still uses snail-mail, handheld calculators, goes to Walmart, and uses a phone with a dial...
better than the by-product of a post-FDR era which stood up barriers to innovation and competition, or post-LBJ era which expanded welfare enough to encourage people to live off of the state...
Earning wealth through risk, innovation, and effort is fine. Externalities, wealth transfers, using accumulated wealth to undermine democracy, justice, and equality is at a minimum immoral. The world is burning due to insatiable greed and willful ignorance.
i think the world is burning due to over-involvement of government in our lives and envy from the people that don't take risk, innovate, and/or put in the effort that want the government to make life "fair"...
Using your example of AI tutors, my thought is that it would be better for society if we instead apply AI to the industries for which people work only to pay the bills, freeing them up to tutor their own children and other children in their community.
Any task that benefits from human connection with each other or the natural world, involves expository writing, requires ethics and critical thinking, or involves poetry/art/music should be left to the humans. We like it. It's why we're here. AI can drive the ubers, guard the prisons, fulfill the endless online orders, and do other jobs that are dangerous/traumatic/soul-crushing.
My other general thought about starting fresh is that much time and energy (environmental and personal) could be saved by operating AI within a closed environment rather than running everything against the entire internet. An analogy to this is how we used to have to select a "library" or organized collection of digitized publications before executing a search for research papers back in the pre-google days. If you select every collection and then run the search, you get a bunch of useless results. This is pretty much why I add "-ai" to the end of every internet search at this point :)
But, I'm not an engineer or a computer scientist, so maybe I'm off base on that last point.
Education is a great place to start this innovation - literally changing the way children interact with technology and especially AI from the very first day of school. I've seen advertisements for this - it's already starting. Alpha School just opened a location near me and it's been filling my Facebook feed for weeks. Looks amazing - kids spend 2 hours a day with an AI tutor and the rest of the school day they practice life skills in a teamwork setting. Teachers become guides. Here's an example of this model: https://www.tiktok.com/@dylan.page/video/7501418546728078614
I started reading by myself when I was 4. My brother never enjoyed reading the way I did. Public education was easy for me and a struggle for my brother. To this day my love of learning has centered around my ability to read. Just imagine a world where every child was tutored to read (and were taught to comprehend what they read) at a very early age. The human potential would be unprecedented.
Addressing just the application of AI to our educational system, it sounds great in theory, but it will cost money. Many families struggle to survive financially these days. How would this be implemented to all children equally? Government is cutting free lunches to hungry kids so I can’t imagine that they would be willing to make sure that even poorer families would get the same resources as those families with plenty of disposable income. I’m not arguing against it, just feel it is idealistic & not practical at this time. We need more compassion in our society.
Hi Jean. What if it costs next to nothing? What if all it requires is computer access (which almost every kid already has) and a handful of educators who link AI tutoring agents (which already exist) to each kids present level of knowledge in any - or every - subject area? If such is already or even mostly the case, significant compassion is a few clicks away.
I put the following 2 (messy) prompts into Claude to get a sliver of a sense of what’s going on currently. Incomplete though it is, try doing the queries yourself or make up better ones (low bar). (;-)
A) “I believe the Khan academy uses AI tutoring agents to teach kids? How successful are they in taking kids at lower end of bell curve of knowledge for their age, up to much higher end? How do they do it? Is it expensive or could any parent or teacher implement it at home or in a conventional classroom?”
B) “Are there other similar AI agents for K to 12 education? Or companies or start-ups developing such agents? And, has anyone demonstrated that the more successful, easier to use, and less expensive AI educational agents are based on active inference (think Karl Friston) models that don’t require loading a huge data base into the AI tutor?”
Cheers!
Khan is excellent on most topics. Excellent.
It costs next to nothing to give kids "free" lunch. The USA could have solved food insecurity in their own country at any point in the last 50 years and they chose not to. A hungry child can't learn effectively with or without a 1-1 tutor.
The USA as a society doesn't value equality and the tech-illiterate politicians don't have the skills to guarantee an equal education with AI tutors even if they had the will power and financial backing to do so.
Sure, go ahead and improve AI systems and create AI tutors but don't expect the end result will be anything resembling equality.
Why would anyone expect equal educational results when children differ so dramatically in potential? Why would anyone want it, for that matter?
Perhaps a better goal than equality would be to simply optimize potential for each child based on their intelligence and proclivities?
I think you are confusing equality and equity.
Equality: the state of being equal, especially in status, rights, and opportunities. For example, do all school aged children have the opportunity to access advanced AI models? Do all school aged children have the opportunity to take classes to learn how to use computers.
In my original comment I am stating the USA doesn't value equality so you will never see a day when all school aged children have access to computers and advanced AI tutors.
Hence “optimize the potential for each child” to quote myself. The point is that individualized AI tutors wouldn’t just be better, they would be cheaper than current education practices and thus better and more affordable for poorer neighborhoods.
By the way, I believe that national school spending is higher today for lower quintile households than it is for the median. So it really isn’t a given that we are currently short-changing poor students.
Hence “optimize the potential for each child” to quote myself. The point is that individualized AI tutors wouldn’t just be better, they would be cheaper than current education practices and thus better and more affordable for poorer neighborhoods.
By the way, I believe that national school spending is higher today for lower quintile households than it is for the median. So it really isn’t a given that we are currently short-changing poor students.
It seems to me that an AI tutor would be next to free. The cost in software is in the first unit, not the hundred-millionth. A better-designed, built up from scratch system with private tutors wouldn’t cost more than today’s schools, it would cost less, and be incomparably better.
"we need more compassion". agree, but not provided by schools...
from a cost perspective:
- we no longer need large properties and buildings to construct, maintain, heat/cool, etc
- no need to bus rural students
- need fewer teachers
- need fewer administrators
other considerations:
- for any given topic, the content taught should be more consistent in quality
- less discrepancies caused by differences in teacher quality
- less dependence on parents to be extensions of teachers
- topics can be taught/learned at the pace of the students
- before progressing, competency in a topic can be enforced
- fewer school shootings
Interesting as always. thanks for putting a stake in the ground. it gives us food for thought.
On the innovation curve diagram, the "tipping point" is also known as the "chasm" and while AI has seemingly (early majority is just beginning) made the jump across the chasm, most other industries have not begun.
i understand your advocating for every industry to hit restart. that said, industries like insurance, banking, and investing cannot do more than regulators will allow and regulators don't like "black box" technologies... of course, in addition to finance, the fed also regulates health, medicine, transportation, communications, agriculture, and energy... not much left. what the fed doesn't regulate, the states do regulate. regional and national companies have to deal with multiple states and that can be fun... 10 years after the rest of the world was using "square tape" (IBM 3480's), the State of Massachusetts still required regulatory data be sent to them on "round reels" (IBM 3420's), so my employer had about fifty 3480 tape drives, and one 3420 tape drive...
as far as rolling out AI goes, regulators aren't the only roadblocks. consumers set the pace. you can't go faster than the early majority, and that leaves out a lot of customers (that the regulators will protect - including the laggards).
one other thing. consolidation. as a former competitive intelligence analyst, i ask, who will win? every industry has consolidated in the past 200 years - with help from the fed which likes 3 to 7 competitors in each space. so, which providers will be standing in 5 or 10 years. i remember numerous start ups that were initially successful losing to today's juggernauts (My Space, Yahoo!, Blackberry, Netscape Navigator, Watson, Ask Jeeves, et al). In fact, there is a strong argument supporting that the first to market almost always loses... I ask this because most current companies will try to wait for the dust to settle before committing to a company that may or may not be around in 5 years...
i agree the purpose of education was to create manufacturing workers (write well enough to report accomplishments and problems, read enough to follow a procedure, enough math to count what you produced, etc.).
what should the purpose of education be now?
You lay out a lot of ideas here, as you often do. Keep them coming. My quick hit is that the economic implications of all this are just starting to dawn on those currently running and leading the economy. I will have more to say about all this in future essays. But the first step is getting people to reorient to how big a deal this is and how thinking fresh and starting from scratch is the best way to approach a seismic change of this sort.
Thank you Peter. In my own observations the commercialization and availability of AI represents a societal inflection point similar to the printing press, the telegraph, the telephone, the microscope, the automobile, and flight.
There must be a corresponding and immediate understanding of the impact of widespread AI use on human labor demand. Without a complete reimagining of economic systems, AI will tend to concentrate even more wealth -beyond the already extreme state which exists in our post-Reagan society today -and is antithetical to any meaningful democracy in the US and anywhere in the world. A dystopian authoritarian future controlled by Thiel, Musk, Zuckerberg, and a handful of others, already on record as to their thoughts about human empathy and the average person trying to live a decent life (as nothing more than a participant in a video game) is not a great progression for the future.
George, I hear you and am aligned with your thinking. A bit part of The Great Progression series and ultimate book will include much on that topic of the need for a fundamental rework of the economy as we make this transition. This was not the essay to do that, but I have one coming out in the next couple weeks that will.
In fact, just yesterday I ran a roundtable with about a dozen people focused on the bigger economic implications in a day long workshop on the arrival of AI agents. And tomorrow I am going to a gathering called "Reimagining the Economy in the Age of AI." Many people in the SF Bay Area are getting focused on that. It's encouraging, but we have a long way to go.
Thanks @Peter Leyden. We both know policy and legislation lag far behind technological evolution. Not that I would suggest putting the toothpaste back in the tube (and the promise of AI dates back to the beginnings of my early use of UNIX) so it’s not like AI is an overnight sensation. Even a few years of diminished human labor demand can shift concentrated wealth well beyond a tipping point for democracy (granted it may already be there). Glad economics has a seat at the strategy table and is in the forefront of your thinking.
funny, "there must be a corresponding and immediate understanding...". the items you list "the printing press, the telegraph, the telephone, the microscope, the automobile, and flight." did not have a corresponding or immediate understanding... the beloved federal government is rarely ahead of the public except in intelligence and defense.
the concentration of wealth that you speak of is a reflection of the value added by those providers receiving the wealth - surely your household doesn't use google, facebook, amazon, apple, or microsoft. people voluntarily transfer their money to the providers that offer them the best value. the ability to scale virtually instead of physically is what has led to the growing gap in wealth. surely your household still uses snail-mail, handheld calculators, goes to Walmart, and uses a phone with a dial...
better than the by-product of a post-FDR era which stood up barriers to innovation and competition, or post-LBJ era which expanded welfare enough to encourage people to live off of the state...
Earning wealth through risk, innovation, and effort is fine. Externalities, wealth transfers, using accumulated wealth to undermine democracy, justice, and equality is at a minimum immoral. The world is burning due to insatiable greed and willful ignorance.
i think the world is burning due to over-involvement of government in our lives and envy from the people that don't take risk, innovate, and/or put in the effort that want the government to make life "fair"...
Using your example of AI tutors, my thought is that it would be better for society if we instead apply AI to the industries for which people work only to pay the bills, freeing them up to tutor their own children and other children in their community.
Any task that benefits from human connection with each other or the natural world, involves expository writing, requires ethics and critical thinking, or involves poetry/art/music should be left to the humans. We like it. It's why we're here. AI can drive the ubers, guard the prisons, fulfill the endless online orders, and do other jobs that are dangerous/traumatic/soul-crushing.
My other general thought about starting fresh is that much time and energy (environmental and personal) could be saved by operating AI within a closed environment rather than running everything against the entire internet. An analogy to this is how we used to have to select a "library" or organized collection of digitized publications before executing a search for research papers back in the pre-google days. If you select every collection and then run the search, you get a bunch of useless results. This is pretty much why I add "-ai" to the end of every internet search at this point :)
But, I'm not an engineer or a computer scientist, so maybe I'm off base on that last point.
My thoughts on this article: https://stevenscesa.substack.com/p/the-real-barriers-to-an-ai-future
Let the debate escalate . . .
Education is a great place to start this innovation - literally changing the way children interact with technology and especially AI from the very first day of school. I've seen advertisements for this - it's already starting. Alpha School just opened a location near me and it's been filling my Facebook feed for weeks. Looks amazing - kids spend 2 hours a day with an AI tutor and the rest of the school day they practice life skills in a teamwork setting. Teachers become guides. Here's an example of this model: https://www.tiktok.com/@dylan.page/video/7501418546728078614