Post #3833796

On one hand I agree that in a lot of ways our current tech surveillance regime looks like the kind of orwellian nightmare that people are raising red flags about. I’m admittedly less terrified by this than I am frustrated at how quickly and easily these systems get assembled when it comes to protecting shareholders from potentially unfavorable market conditions like “people using the software they purchased and installed on their computer in a way we didn’t expect”. Like, the US can’t get it’s shit together to do basically anything about domestic terrorism to the point where elementary schools are having domestic terror drills in order to feel like they’re doing literally anything to acknowledge the problem, but God forbid OpenAI has to acknowledge their product’s risk of aiding those terrorists when they’re trying to get their IPO off the ground. But also I think Scott is hilariously uninformed about how in-depth the kind of monitoring systems would need to be in order to function trans nationally and not be subject to individual jurisdiction (because otherwise the only one with authority to go after the secret illegal chinese AI bunker project would be China and that’s a non-starter. As I said in the OP I think this combined with the level of transparency that they’re discussing would basically amount to opening up all the books and internal communications directly to your competitors whether on a corporate or national level, which makes it a non-starter.