«because we need to carry the torch even if we offer almost no jobs that require philosophy degrees.»
Among the few "almost no jobs": upper-class people used to study and still study philosophy or history, classics, ... because their roles do require it: their role is to rule that is to lead and direct those working for them and insights as to knowledge, history, literature are very useful for that.
In another way *any* wide-ranging subject like philosophy can improve people's cognitive abilities and after all not many people who get degrees in vocationally oriented subjects like engineering or accounting or law or programming do work in those subjects for most of their careers, but the myth of "job relevant" subjects is hard to dispel.
Amen, both to the specific points of your essay and the general proposition that we all need to learn from, back up, and get support from grandparents and those who are older than us. (I'm 65 so I'm trying to do this for people who are younger...)
One of my most satisfying experiences both on a personal and professional level has been the Think and Thrive Lecture Series (formerly Senior College) in Alexandria, Minnesota:
100 or so people participate in these sessions, most of them twice per week each semester. I've been presenting at least one session, usually two, per year since 2009. The range of insight and experience the participants bring is beyond anything I give them. I've met doctors, lawyers, engineers, teachers, ministers, accountants, you name it, and learned from all of them.
I'm convinced that we need to replicate this kind of forum throughout the US so that we can bring this wisdom to the fore. I've suggested this type of program to deans, provosts, college presidents, alumni offices, and community ed people to connect town and gown, to bring together students, staff, faculty, retirees, and alumni, and to create conversations in all kinds of communities that need to be had. Unfortunately, I've gotten nowhere with the idea and stopped shopping it. Perhaps I should begin again...
Thank you and the Think and Thrive series looks fantastic. New kinds of communities are being formed, locally and in networks, as universities close. Keep shopping it!
I have proposed a tribe of LLM Boddhisatvas, taking random walks through agentic AI to try to infuse both meaning and ethics in their interactions with humans.
I wonder how much your vision of grandmothers matches with my vision.
Nonetheless, all I can do is what I can do. Now that the State of Utah has released me, explicitly favoring Tech Bros over my students, I can no longer transfer my grandfather-level knowledge to my students and junior colleagues. I will continue to do so, but as my own wandering academic Johnny Appleseed, planting trees of meaning that I will never sit under.
Thanks for this recap! This has been on my worry list :\. A classic tragedy of the commons problem, I seriously doubt any of the current commercial LLM platforms would consider degrading their systems for long-term benefit.
It's not too late, it's already happening organically. Yes scholars can more rapidly produce work with AI, and this is intended benefit, and yes layman can produce work that is convincing to another layman, and ideally challenges the work of the scholar, which is another benefit. However grandmother discernment is needed. This can be seen with plateau of AI generated content. Mass production of layman's work that nobody understands is a direct path to stagnation, and so it won't happen.
Great point on Proposition 15 — categorising the grandmother as the demographic analogue of an inflow independent of novice incentives is precisely the right framing.
My position, however, is that knowledge does not collapse but stagnates — it is already encoded in LLMs. This means the grandmother as demographic analogue of the knowledge inflow will delay stagnation rather than stabilise against it in the long run. Acemoglu's own premise : that novices have no learning incentive under high AI precision ensures the cohort who should become grandmothers at t+n accumulated no expertise at time t. The grandmother contribution is a depleting inflow, not a permanent floor. The stabilisation is temporal rather than structural.
Thank you and genuinely pleased the argument landed. The SSRN paper develops the cognitive architecture dimension more formally if that is useful. Would very much welcome a conversation at some point
«overlapping generations with post-reproductive specialists whose payoffs run through dependents»
The standard approach of academic Economics ignores entirely the fact that reproduction is important to "economic agents" and parents do not necessarily maximize their personal "utility" but also that of their children and grandchildren, sometimes at high cost to their own "utility". instead a noted this is the standard approach:
«The model treats all human learners as identical single-period decision-makers whose incentive to exert effort depends entirely on the private return from context-specific knowledge»
The reason why this approach is standard is because it is assumed that all inter-agent interactions happen exclusively through the markets and there are no economically relevant relationships among agents in the markets.
There are two related motivations for that approach:
* A technical one where any inter-agent and inter-generational relationships generates path-dependency in the model and therefore prevents the proof that unrestricted markets deliver the best income distribution, which is the purpose of academic Economics.
* A political one where that approach reflects the "liberal libertarian" ideology favored by the "sponsors" of much academic Economics where market ineractions should matter far more than institutions like marriage, parenthood, grantparenthood.
This post raises some fascinating points about the relationship between historical context and innovation in technology. It echoes themes from my piece on the idea that mathematical heroes might just be historical accidents, which you might find interesting here: https://theuncomfortableidea.substack.com/p/heroes-in-mathematics-might-be-nothing.
I read that the development of the AP Computer Science Principles course was born out of a desire for more gender parity in Computer Science AP courses. Giving wider access to the language of the principles allowed for more students who are fluent/interested in other disciplines to "speak" to programmers. It seems like we need that kind on interdisciplinarity for the grandmothers. Perhaps in philosophy as Michelle notes.
Someone needs to study things like philosophy because we need to carry the torch even if we offer almost no jobs that require philosophy degrees.
«Someone needs to study things like philosophy»
Traditionally that was upper-class people.
«because we need to carry the torch even if we offer almost no jobs that require philosophy degrees.»
Among the few "almost no jobs": upper-class people used to study and still study philosophy or history, classics, ... because their roles do require it: their role is to rule that is to lead and direct those working for them and insights as to knowledge, history, literature are very useful for that.
In another way *any* wide-ranging subject like philosophy can improve people's cognitive abilities and after all not many people who get degrees in vocationally oriented subjects like engineering or accounting or law or programming do work in those subjects for most of their careers, but the myth of "job relevant" subjects is hard to dispel.
hunched in corner, catatonically rocking back and forth...
You or your grandmother??
Grandma's fine.
Amen, both to the specific points of your essay and the general proposition that we all need to learn from, back up, and get support from grandparents and those who are older than us. (I'm 65 so I'm trying to do this for people who are younger...)
One of my most satisfying experiences both on a personal and professional level has been the Think and Thrive Lecture Series (formerly Senior College) in Alexandria, Minnesota:
https://www.alextech.edu/atcc-events/lectureseries
100 or so people participate in these sessions, most of them twice per week each semester. I've been presenting at least one session, usually two, per year since 2009. The range of insight and experience the participants bring is beyond anything I give them. I've met doctors, lawyers, engineers, teachers, ministers, accountants, you name it, and learned from all of them.
I'm convinced that we need to replicate this kind of forum throughout the US so that we can bring this wisdom to the fore. I've suggested this type of program to deans, provosts, college presidents, alumni offices, and community ed people to connect town and gown, to bring together students, staff, faculty, retirees, and alumni, and to create conversations in all kinds of communities that need to be had. Unfortunately, I've gotten nowhere with the idea and stopped shopping it. Perhaps I should begin again...
Thank you and the Think and Thrive series looks fantastic. New kinds of communities are being formed, locally and in networks, as universities close. Keep shopping it!
I have proposed a tribe of LLM Boddhisatvas, taking random walks through agentic AI to try to infuse both meaning and ethics in their interactions with humans.
I wonder how much your vision of grandmothers matches with my vision.
Nonetheless, all I can do is what I can do. Now that the State of Utah has released me, explicitly favoring Tech Bros over my students, I can no longer transfer my grandfather-level knowledge to my students and junior colleagues. I will continue to do so, but as my own wandering academic Johnny Appleseed, planting trees of meaning that I will never sit under.
Thanks for this recap! This has been on my worry list :\. A classic tragedy of the commons problem, I seriously doubt any of the current commercial LLM platforms would consider degrading their systems for long-term benefit.
It's not too late, it's already happening organically. Yes scholars can more rapidly produce work with AI, and this is intended benefit, and yes layman can produce work that is convincing to another layman, and ideally challenges the work of the scholar, which is another benefit. However grandmother discernment is needed. This can be seen with plateau of AI generated content. Mass production of layman's work that nobody understands is a direct path to stagnation, and so it won't happen.
Great point on Proposition 15 — categorising the grandmother as the demographic analogue of an inflow independent of novice incentives is precisely the right framing.
My position, however, is that knowledge does not collapse but stagnates — it is already encoded in LLMs. This means the grandmother as demographic analogue of the knowledge inflow will delay stagnation rather than stabilise against it in the long run. Acemoglu's own premise : that novices have no learning incentive under high AI precision ensures the cohort who should become grandmothers at t+n accumulated no expertise at time t. The grandmother contribution is a depleting inflow, not a permanent floor. The stabilisation is temporal rather than structural.
I have engaged related questions from a cognitive architecture angle here, if useful: https://structuralepistemology.substack.com/p/when-ai-makes-us-fluent-but-not-intelligent
This is very very good and I restacked it. I will look at your other work! Thank you. I’m happy to connect at some point.
Thank you and genuinely pleased the argument landed. The SSRN paper develops the cognitive architecture dimension more formally if that is useful. Would very much welcome a conversation at some point
«overlapping generations with post-reproductive specialists whose payoffs run through dependents»
The standard approach of academic Economics ignores entirely the fact that reproduction is important to "economic agents" and parents do not necessarily maximize their personal "utility" but also that of their children and grandchildren, sometimes at high cost to their own "utility". instead a noted this is the standard approach:
«The model treats all human learners as identical single-period decision-makers whose incentive to exert effort depends entirely on the private return from context-specific knowledge»
The reason why this approach is standard is because it is assumed that all inter-agent interactions happen exclusively through the markets and there are no economically relevant relationships among agents in the markets.
There are two related motivations for that approach:
* A technical one where any inter-agent and inter-generational relationships generates path-dependency in the model and therefore prevents the proof that unrestricted markets deliver the best income distribution, which is the purpose of academic Economics.
* A political one where that approach reflects the "liberal libertarian" ideology favored by the "sponsors" of much academic Economics where market ineractions should matter far more than institutions like marriage, parenthood, grantparenthood.
This post raises some fascinating points about the relationship between historical context and innovation in technology. It echoes themes from my piece on the idea that mathematical heroes might just be historical accidents, which you might find interesting here: https://theuncomfortableidea.substack.com/p/heroes-in-mathematics-might-be-nothing.
I read that the development of the AP Computer Science Principles course was born out of a desire for more gender parity in Computer Science AP courses. Giving wider access to the language of the principles allowed for more students who are fluent/interested in other disciplines to "speak" to programmers. It seems like we need that kind on interdisciplinarity for the grandmothers. Perhaps in philosophy as Michelle notes.