Discussion about this post

User's avatar
Nick Felker's avatar

How would regulations and even nationalization work in a year from now when open weight models can rival or surpass Fable?

Michael Alan Dover, PhD's avatar

Great discussion of social policy related to AI and how to link it to climate change and to broader progressive objectives. I've recommened your substack.

Zachary, you are correct about climate change policy running into a brick wall: the best work is informed by theory of human need (see https://www.iangough.com), but what single nation has resolutely addressed these issues? Yet anger is rising about "data centers" and tapping into it is key. The question is what is the policy goal?

I found this link of yours helpful: https://drive.google.com/file/d/1-buzxZ6uDZTQYqRxJ9HAttMpZfvuwvXH/view

However, beware of what I call programitization, when we propose throwing a particularistic program at a universally experienced danger.

Such a program would be proposed safety nets linked to AI-related job loss. The experience of the now deauthorized Trade Re-Adjustment Assistance Act of 1974, a very effect program, but since 2022 now essentially defunct, comes to mind.

How about bringing that back and expanding it to include AI-related job losses? But I worry: how do you prove the causes of job loss?

What about making TRAA-like programs universal, beefing up the whole unemployment compensation and job-retraining entitlements for all citizens and permanent residents? What about the proposal in Bill Quigley's Ending Poverty as We Know It to promote a constitutional amendment for the right to a job at a living wage? Shouldn't that be part of our "minimum program?"

Sen. Peter's has a bill to re-authorize TRAA: http://congress.gov/bill/119th-congress/senate-bill/1449/text. In the new Congress, and in campaigns this fall, the progressive pragmatist sweet spot should be to propose more sweeping version of such a bill.

However, I'm not sure about the question of AI-related job losses. I think we have to recognize that AI will likely substantially increase worker productivity, and could lead to the loss of what are known as "bullshit jobs" (see David Graber's work https://en.wikipedia.org/wiki/Bullshit_Jobs), and an increase in mechanization of manual tasks (take the advent now of robots working like door dash). Your warning against a progressive version of being a Luddite is a valuable point.

But the AI Denial approach is very worrisome! There are too many who want to shut down discussion of the very issue! Or accuse you of killing bunnies if you personally use it as what I call "Google on steroids." Having Norway buy one of the key firms is neat: but it is a competitive industry, could a government owned-entity compete?

1 more comment...

No posts

Ready for more?