Curious about the ‘delay the development’ via regulation bit.
What is your sense of what near-term passable regulations would be that are actually enforceable? It’s been difficult for large stakeholder groups facing threatening situations to even enforce established international treaties, such as the Geneva convention or the Berne three-step test.
Here are dimensions I’ve been thinking need to be constrained over time:
Input bandwidth to models (ie. available training and run-time data, including from sensors).
Multi-domain work by/through models (ie. preventing an automation race-to-the-bottom)
Output bandwidth (incl. by having premarket approval for allowable safety-tested uses as happens in other industries).
Compute bandwidth (through caps/embargos put on already resource-intensive supply chains).
(I’ll skip the ‘make humans smarter’ part, which I worry increases problems around techno-solutionist initiatives we’ve seen).
Curious about the ‘delay the development’ via regulation bit.
What is your sense of what near-term passable regulations would be that are actually enforceable? It’s been difficult for large stakeholder groups facing threatening situations to even enforce established international treaties, such as the Geneva convention or the Berne three-step test.
Here are dimensions I’ve been thinking need to be constrained over time:
Input bandwidth to models (ie. available training and run-time data, including from sensors).
Multi-domain work by/through models (ie. preventing an automation race-to-the-bottom)
Output bandwidth (incl. by having premarket approval for allowable safety-tested uses as happens in other industries).
Compute bandwidth (through caps/embargos put on already resource-intensive supply chains).
(I’ll skip the ‘make humans smarter’ part, which I worry increases problems around techno-solutionist initiatives we’ve seen).