Rereading this, I'm noticing that your Mars example feels rather strange and doesn't quite seem to work.
The reason why worrying about environmental regulation might be bizarre is because there are other things that could cause you all to die in this scenario (such as running out of food, water or electricity) and the assumption is neglecting environmental regulation isn't going to kill you.
Notably, if failing in terms of environmental governance could actually kill you (because some industrialist was planning to pour toxic chemicals in your water supply), it'd quickly jump up your priority list.
Now, in the AI case, regulation hugely varies. Some is closer to "environmental regulation", whilst other regulations have been proposed precisely because it's supposed to keep you dying. Treating these distinct kinds of regulation uniformly is a notable weakness in the case you've presented.
Secondly, the case for "more important things to be done" that you've presented is quite sparse. What is the more important thing that you think needs to be done given that government regulation might be the difference between life and death? You gesture towards "constraining the state is a founding principle of America", but that's a single thing (as opposed to the whole host of more important things that was suggested) and the argument "it worked in the past, so it'll work for us now" is in tension with your assertion that we need to completely rethink our institutions for the AI age.
For what it's worth, I think you were onto something with this post. Even if I don't think it'd work (as per my other comment), it's both original and discussion-worthy. I might even agree with you if I thought timelines were longer.
In 1975, Lawrence Sanders wrote a near-future dystopian novel called The Tomorrow File. Its frontspiece was a quote from the second inaugural address of the fictional president Harold Morse in 1988 that your invocation of another Founding reminded me of:
"We can no longer afford an obsolete society of obsolete people."
I agree that agility is underrated compared to specific policy plans.
It's interesting and I'd love to see it applied in other contexts, but AI is moving too fast for this plan to be viable.
Once created, any private governance orgs would attempt to preserve their own existence and ward off any government regulation. This could be disastrous when things get real and when we need serious action (not lowest common denominator governance based on orgs either opting out or picking the most lax private governance org to safety-wash their actions).
Rereading this, I'm noticing that your Mars example feels rather strange and doesn't quite seem to work.
The reason why worrying about environmental regulation might be bizarre is because there are other things that could cause you all to die in this scenario (such as running out of food, water or electricity) and the assumption is neglecting environmental regulation isn't going to kill you.
Notably, if failing in terms of environmental governance could actually kill you (because some industrialist was planning to pour toxic chemicals in your water supply), it'd quickly jump up your priority list.
Now, in the AI case, regulation hugely varies. Some is closer to "environmental regulation", whilst other regulations have been proposed precisely because it's supposed to keep you dying. Treating these distinct kinds of regulation uniformly is a notable weakness in the case you've presented.
Secondly, the case for "more important things to be done" that you've presented is quite sparse. What is the more important thing that you think needs to be done given that government regulation might be the difference between life and death? You gesture towards "constraining the state is a founding principle of America", but that's a single thing (as opposed to the whole host of more important things that was suggested) and the argument "it worked in the past, so it'll work for us now" is in tension with your assertion that we need to completely rethink our institutions for the AI age.
For what it's worth, I think you were onto something with this post. Even if I don't think it'd work (as per my other comment), it's both original and discussion-worthy. I might even agree with you if I thought timelines were longer.
In 1975, Lawrence Sanders wrote a near-future dystopian novel called The Tomorrow File. Its frontspiece was a quote from the second inaugural address of the fictional president Harold Morse in 1988 that your invocation of another Founding reminded me of:
"We can no longer afford an obsolete society of obsolete people."
Be careful what you wish for, in other words.
I agree that agility is underrated compared to specific policy plans.
It's interesting and I'd love to see it applied in other contexts, but AI is moving too fast for this plan to be viable.
Once created, any private governance orgs would attempt to preserve their own existence and ward off any government regulation. This could be disastrous when things get real and when we need serious action (not lowest common denominator governance based on orgs either opting out or picking the most lax private governance org to safety-wash their actions).