At the same time, the plan is trying to prepare us for AGI and its associated risks as best its authors can without explaining that it is doing this.
I think actually doing this would include Liability.
I think overall this bill will be accelerating capabilities, give a brief percent of wording to safety esque things, which will make some intellectual safetyists act happy in public, for the hope that showing how happy they are with what they got will lead to more like it that’s closer to what they actually want.
Dean Ball seems to generally act against putting AI regulation in place for safety and seems to be good at making intellectual arguments to convince safetyists that actually this version of some safety regulation isn’t the right one.
I think actually doing this would include Liability.
I think overall this bill will be accelerating capabilities, give a brief percent of wording to safety esque things, which will make some intellectual safetyists act happy in public, for the hope that showing how happy they are with what they got will lead to more like it that’s closer to what they actually want.
Dean Ball seems to generally act against putting AI regulation in place for safety and seems to be good at making intellectual arguments to convince safetyists that actually this version of some safety regulation isn’t the right one.