I am really enjoying thinking with this! As an educator, “creative confidence” really resonates me as a top priority. Obviously we disagree about AI though so I’m wondering how you see it as promoting rather than discouraging connection. The third option you don’t mention for AI adoption is people no longer speaking much to each other or having original thoughts. I also think an analysis of how intensely a few powerful entities control generative AI is important for thinking through your questions here. I recently read Karen Hao’s Empire of AI, which is mostly a cautionary tale/takedown of OpenAI, but she also has a coda where she says AI doesn’t have to be like that and gives the example of a community-oriented group using AI to archive te reo Māori, which expanded my thinking about AI a lot—I feel like some of that group’s principles align with what you’re saying here.
As a poet (& someone who grew up in SG, but now lives elsewhere), I find these questions encouraging.
The spirit of education has always been, for me, about a posture that is open and curious about learning. I understand the need for measurable outcomes, but the spirit must always exceed the containers we make for it.
The pragmatic question I have in response to these helpful thoughts about education is this: What, then, will we pay people to do? Even surfers and artists have to make something that people are willing to pay for (for surfers, it's about selling entertainment; for artists, it's about selling the actual pieces or merch).
Everyone deserves a creative practice, but figuring out how to pay the bills remains a pressing problem for artists, surfers and, yes, poets.
I think Goodhart's law is well worth keeping in mind but I think we should pay a lot of attention on creating new kinds of metrics.
Goodhart's law has always been a problem: Even in knowledge based exams, one could memorise or just learn how to do specific problem types and leave a course not having developed a good understanding, or one could forgetting everything shortly after. Even in academic publishing metrics get gamed at the expense of research quality. What we should do in response is to just change our tests. Make them unpredictable and change every year, make them satisficing and not indefinitely maximisable, etc.
I think Singaporeans are world-class at picking an objective and completing it (top unis, top airport, cleanest city, students maximising grades, adults maximising money, etc.). This energy should be tapped on and redirected to the extent we can create metrics for aliveness or whatever other virtues we want to inculcate.
This is quite a difficult challenge but I think doing it this way will be more scientific and more practical. And tangentially - I think it's the heart of multiple societal issues right now - AI needs new evaluations for measuring both its safety and capability; people complain how we measure a society on its GDP but miss many other important things - we need new metrics.
And as we sit down to try to articulate what we really want to teach, we'll have a better idea of what precisely we want, and we'll get all the benefits of science too - we'll be able to measure how well our interventions are working and so on.
I wonder what you think! Would love to dive deeper and prototype or collect good thinking on what some of these metrics might look like.
I'm hoping to also compile a list of metrics that perhaps already exists in psych or social science literature that measures the 6 aspects outlined! I think they could be a good initial way to measure effectiveness of any interventions. Wdyt about collaborating on that?
I am really enjoying thinking with this! As an educator, “creative confidence” really resonates me as a top priority. Obviously we disagree about AI though so I’m wondering how you see it as promoting rather than discouraging connection. The third option you don’t mention for AI adoption is people no longer speaking much to each other or having original thoughts. I also think an analysis of how intensely a few powerful entities control generative AI is important for thinking through your questions here. I recently read Karen Hao’s Empire of AI, which is mostly a cautionary tale/takedown of OpenAI, but she also has a coda where she says AI doesn’t have to be like that and gives the example of a community-oriented group using AI to archive te reo Māori, which expanded my thinking about AI a lot—I feel like some of that group’s principles align with what you’re saying here.
As a poet (& someone who grew up in SG, but now lives elsewhere), I find these questions encouraging.
The spirit of education has always been, for me, about a posture that is open and curious about learning. I understand the need for measurable outcomes, but the spirit must always exceed the containers we make for it.
The pragmatic question I have in response to these helpful thoughts about education is this: What, then, will we pay people to do? Even surfers and artists have to make something that people are willing to pay for (for surfers, it's about selling entertainment; for artists, it's about selling the actual pieces or merch).
Everyone deserves a creative practice, but figuring out how to pay the bills remains a pressing problem for artists, surfers and, yes, poets.
I think Goodhart's law is well worth keeping in mind but I think we should pay a lot of attention on creating new kinds of metrics.
Goodhart's law has always been a problem: Even in knowledge based exams, one could memorise or just learn how to do specific problem types and leave a course not having developed a good understanding, or one could forgetting everything shortly after. Even in academic publishing metrics get gamed at the expense of research quality. What we should do in response is to just change our tests. Make them unpredictable and change every year, make them satisficing and not indefinitely maximisable, etc.
I think Singaporeans are world-class at picking an objective and completing it (top unis, top airport, cleanest city, students maximising grades, adults maximising money, etc.). This energy should be tapped on and redirected to the extent we can create metrics for aliveness or whatever other virtues we want to inculcate.
This is quite a difficult challenge but I think doing it this way will be more scientific and more practical. And tangentially - I think it's the heart of multiple societal issues right now - AI needs new evaluations for measuring both its safety and capability; people complain how we measure a society on its GDP but miss many other important things - we need new metrics.
And as we sit down to try to articulate what we really want to teach, we'll have a better idea of what precisely we want, and we'll get all the benefits of science too - we'll be able to measure how well our interventions are working and so on.
I wonder what you think! Would love to dive deeper and prototype or collect good thinking on what some of these metrics might look like.
I'm hoping to also compile a list of metrics that perhaps already exists in psych or social science literature that measures the 6 aspects outlined! I think they could be a good initial way to measure effectiveness of any interventions. Wdyt about collaborating on that?
Nice! Let's do it 👍