The architect of tomorrow’s AI must be as much a philosopher as a technician, capable of wrestling with questions that cannot be answered by code alone. What does it mean to create intelligence?
Hey Michael, this is a great essay, but perhaps too timid. I would suggest that without some appropriately “scaled” classical education, we leave any person adrift in a left-brain fog (in the true left-right sense, see McGilchrist), mistaking maps for the territory.
Much of today’s “wisdom,” as I’ve noted elsewhere, represents the brain (or person) as algorithm machine. This is bunk. Sure we “run” algorithms. Yet narrative, and interpretation of narrative is far more central to being human. Classical Education?
Try asking: Tell me about your family? No sane person will set into showing you family trees and process-flow diagrams for Christmas dinner. The sane person will tell you stories.
One doesn’t teach a child to play baseball by giving him a rulebook. Let him watch the game and try things out. It’s fascinating to read about how children actually learn to play games. Hint? It’s not the rules.
A seminal book to this point is Prof. Gerald Holton’s intimidating sounding, Thematic Origins of Scientific Thought. Holton was the first person given access to Albert Einstein’s personal writings and notes following his death. He was a prof of the history of science at Harvard. It is an elegantly written book, clear and enjoyable to wander through.
Based on his learnings from Einstein’s material and his studies of other situations, Holton posits two types of science, S1 and S2. Hoping that the good professor doesn’t read my pitiful simplification of his ideas, I’ll explain. S2 science is basically “how we explain things.” For example, I was taught that the Michaelson-Morley diffraction experiments in the late 1880s gave results incomprehensible from within the Newtonian framework. Einstein, to resolve the paradox, hit upon relativity a bit before 1910.
What Dr. Holton found, however, was that Einstein’s breakthrough ideas came not through some logical stepwise process, but sprung from a deep desire to find symmetry in universal laws. Only well after formulating his theories did it become clear that they provided an explanation for the M-M experiments.
Holton’s S1 science is “how things actually happen.” The S2 explanation I received in high school made perfect “sense” to students. First A, leading to B, leading to C. Ta da! But it was a mirage.
Much of today’s education teaches us to seek S2 explanations. Pharmaceutical companies spent uncounted billions on high-throughput screening of drug candidates aiming to increase their “innovation” by systematizing the search and assessment process. A far cry from Flemming’s discovery of penicillin. Over more than two decades HTPS has yielded unimpressive results. (But don’t ask the equipment vendors or consultants!!)
A classical education places the student in the messy world of S1. The real world! She learns to appreciate human genius not as a process, but as a gift and a mystery.
If I may, it's more than messy. Messy, perhaps, judged with respect to a flow chart or piece of software.
Yet it does have organizing principles, some of which you explicate in this and other CCR articles. It's just not the type of order understood in today's prevailing world view.
My challenge is that my own education is roughly what mathematicians call a "dense set." (Maybe in more ways than one.) My undergrad was straight-up engineering with only a passing nod to anything else. As a very smart high schooler, I managed to duck the classical elements the Xaverian Brothers sought to impart. After a few years in industry, it was off for my MBA.
Only post-40 have I tried to fill in the holes. So, like the dense set, I know a whole bunch, but don't cover all of anything. One particular shortcoming is not understanding many of the underlying frameworks of disciplines like philosophy. Too often I see ideas or phenomena that don't neatly fit into my mental models
Writing like yours helps me to consolidate and contextualize much of my own "messy attic."
The best argument is that Steve Jobs took one calligraphy class in his brief college time, and it inspired him to the Graphical User Interface. And all that implies. But I doubt that classical education in its principles or practice is anywhere to be found in Silicon Valley today. They are rather the tyrannical Oligarchs of AI currently. Quite the opposite of Apple’s 1984 commercial.
True, but as Mark Andreesen and others have pointed out, the techno-garcs of Silicon Valley often send their own children to low-tech or no-tech schools. See Alan Eagle.
Yes, I am aware of not allowing or severely limiting tech in their families. But I would be interested to know if anyone has surveyed SV right now to see what most of them are actually doing with their family education.
Love this! When I first got a job at a tech company after graduating from a liberal arts college, I wasn't sure how my education would play a role in tech. Now after working as a software testing manager for seven years, I don't know how people succeed without it! From knowledge of basic causality and logic to aesthetics and ethics, every level of the SDLC relies on well formed, critical thinking minds.
I used to be a computer scientist (a very creative field, as you mentioned) and a few months ago I was discussing AI with two physicists. I told them that we are in dire need of a “philosophy of AI.” One of them couldn’t get it at all, and the other agreed with me wholeheartedly. We are stumbling around in a minefield. We need classically-educated philosophers to work on this.
The article argues that a classical education, which includes ancient languages, philosophy, rhetoric, and mathematics, is an excellent preparation for careers in technology and computer science. Here are the main points:
Integrated Knowledge: Classical education emphasizes integrating different fields of knowledge, which mirrors how technological innovation often combines disparate ideas.
Creativity: Classical education fosters creativity, which is essential in technology. It helps students imagine possibilities and create new solutions.
Intellectual Rigor: Learning logic and philosophy teaches intellectual discipline, which is valuable in debugging code and designing algorithms.
Moral Philosophy: Classical education includes moral philosophy, encouraging students to consider ethical implications of technology.
Adaptability: The timeless principles of classical education prepare students to adapt to rapidly changing technology.
The article posits that the ideal architect of tomorrow’s AI needs to be both a philosopher and a technician, capable of answering complex human questions that cannot be solved by code alone.
Hey Michael, this is a great essay, but perhaps too timid. I would suggest that without some appropriately “scaled” classical education, we leave any person adrift in a left-brain fog (in the true left-right sense, see McGilchrist), mistaking maps for the territory.
Much of today’s “wisdom,” as I’ve noted elsewhere, represents the brain (or person) as algorithm machine. This is bunk. Sure we “run” algorithms. Yet narrative, and interpretation of narrative is far more central to being human. Classical Education?
Try asking: Tell me about your family? No sane person will set into showing you family trees and process-flow diagrams for Christmas dinner. The sane person will tell you stories.
One doesn’t teach a child to play baseball by giving him a rulebook. Let him watch the game and try things out. It’s fascinating to read about how children actually learn to play games. Hint? It’s not the rules.
A seminal book to this point is Prof. Gerald Holton’s intimidating sounding, Thematic Origins of Scientific Thought. Holton was the first person given access to Albert Einstein’s personal writings and notes following his death. He was a prof of the history of science at Harvard. It is an elegantly written book, clear and enjoyable to wander through.
Based on his learnings from Einstein’s material and his studies of other situations, Holton posits two types of science, S1 and S2. Hoping that the good professor doesn’t read my pitiful simplification of his ideas, I’ll explain. S2 science is basically “how we explain things.” For example, I was taught that the Michaelson-Morley diffraction experiments in the late 1880s gave results incomprehensible from within the Newtonian framework. Einstein, to resolve the paradox, hit upon relativity a bit before 1910.
What Dr. Holton found, however, was that Einstein’s breakthrough ideas came not through some logical stepwise process, but sprung from a deep desire to find symmetry in universal laws. Only well after formulating his theories did it become clear that they provided an explanation for the M-M experiments.
Holton’s S1 science is “how things actually happen.” The S2 explanation I received in high school made perfect “sense” to students. First A, leading to B, leading to C. Ta da! But it was a mirage.
Much of today’s education teaches us to seek S2 explanations. Pharmaceutical companies spent uncounted billions on high-throughput screening of drug candidates aiming to increase their “innovation” by systematizing the search and assessment process. A far cry from Flemming’s discovery of penicillin. Over more than two decades HTPS has yielded unimpressive results. (But don’t ask the equipment vendors or consultants!!)
A classical education places the student in the messy world of S1. The real world! She learns to appreciate human genius not as a process, but as a gift and a mystery.
Education is messy indeed. Great food for thought!
If I may, it's more than messy. Messy, perhaps, judged with respect to a flow chart or piece of software.
Yet it does have organizing principles, some of which you explicate in this and other CCR articles. It's just not the type of order understood in today's prevailing world view.
My challenge is that my own education is roughly what mathematicians call a "dense set." (Maybe in more ways than one.) My undergrad was straight-up engineering with only a passing nod to anything else. As a very smart high schooler, I managed to duck the classical elements the Xaverian Brothers sought to impart. After a few years in industry, it was off for my MBA.
Only post-40 have I tried to fill in the holes. So, like the dense set, I know a whole bunch, but don't cover all of anything. One particular shortcoming is not understanding many of the underlying frameworks of disciplines like philosophy. Too often I see ideas or phenomena that don't neatly fit into my mental models
Writing like yours helps me to consolidate and contextualize much of my own "messy attic."
The best argument is that Steve Jobs took one calligraphy class in his brief college time, and it inspired him to the Graphical User Interface. And all that implies. But I doubt that classical education in its principles or practice is anywhere to be found in Silicon Valley today. They are rather the tyrannical Oligarchs of AI currently. Quite the opposite of Apple’s 1984 commercial.
True, but as Mark Andreesen and others have pointed out, the techno-garcs of Silicon Valley often send their own children to low-tech or no-tech schools. See Alan Eagle.
Yes, I am aware of not allowing or severely limiting tech in their families. But I would be interested to know if anyone has surveyed SV right now to see what most of them are actually doing with their family education.
Michael, Thanks. This was a well written and well argued post. And just plain enjoyable to read. Blessings, Richard
Love this! When I first got a job at a tech company after graduating from a liberal arts college, I wasn't sure how my education would play a role in tech. Now after working as a software testing manager for seven years, I don't know how people succeed without it! From knowledge of basic causality and logic to aesthetics and ethics, every level of the SDLC relies on well formed, critical thinking minds.
I used to be a computer scientist (a very creative field, as you mentioned) and a few months ago I was discussing AI with two physicists. I told them that we are in dire need of a “philosophy of AI.” One of them couldn’t get it at all, and the other agreed with me wholeheartedly. We are stumbling around in a minefield. We need classically-educated philosophers to work on this.
"We are stumbling around in a minefield." Exactly!
Here's the TLDR summary from AI. :-)
The article argues that a classical education, which includes ancient languages, philosophy, rhetoric, and mathematics, is an excellent preparation for careers in technology and computer science. Here are the main points:
Integrated Knowledge: Classical education emphasizes integrating different fields of knowledge, which mirrors how technological innovation often combines disparate ideas.
Creativity: Classical education fosters creativity, which is essential in technology. It helps students imagine possibilities and create new solutions.
Intellectual Rigor: Learning logic and philosophy teaches intellectual discipline, which is valuable in debugging code and designing algorithms.
Moral Philosophy: Classical education includes moral philosophy, encouraging students to consider ethical implications of technology.
Adaptability: The timeless principles of classical education prepare students to adapt to rapidly changing technology.
The article posits that the ideal architect of tomorrow’s AI needs to be both a philosopher and a technician, capable of answering complex human questions that cannot be solved by code alone.
And this proves what?
It's a summary of the article that argues a classical educational is a great preparation for careers in tech.
Yes, but using AI to prove an argument for classical education is . . . like the Goliath telling David how to aim his slingshot?
Someone had to teach David how to aim. ;-)