4 ms·
Reactionary politics aren't known for their well-thought-out and considered policy ideas. You're applying logic to what is largely an emotional topic for people
by ToucanLoucan 2y ago
Reactionary politics aren't known for their well-thought-out and considered policy ideas. You're applying logic to what is largely an emotional topic for people, no matter how many times they say shit like "facts don't care about your feelings," their policies are almost in whole, feelings, and the subject of immigration is a great example.
They're furious about illegal immigrants taking jobs from Americans, and they just flatly are not. The jobs don't pay enough for Americans to want them. This is demonstrated every time there's a crackdown on it and farmers in the south west have fields full of rotting crops that won't get harvested because the Americans who could do that job just... won't.
They're also furious about illegal immigrants getting access to government services while not paying taxes, which is, in the vast majority of cases, a double-falsehood. Not only do illegal immigrants regularly pay taxes (the IRS has specific sections of the forms specifically for non-citizens) but they are in fact completely unable to access virtually any government services, including those that one can make a solid case they deserve, like access to worker's rights protections/organizations, and the ability to go to the police when they're victims of crimes, which they are disproportionately to the rest of the population, again, because criminals know they can't go to the cops.
> And conversely, I wonder what will happen if and when we finally do get useful humanoid robots.
I mean, even the cheapest number Elon's thrown out for his AI-powered humanoid robot (which I can't stress enough, does not exist in any meaningful way relative to what he is selling it but nevertheless) puts it at costing years of what an immigrant does to perform the work, and that's not even going into the fact that having them out in fields vs. climate controlled warehouses is likely going to increase the amount of maintenance needed, if they could manage it at all without just falling directly over. And there are energy concerns too, you'd almost need a mobile charging station for agricultural applications of them.
To be honest if you were setting out to automate the picking of crops that thus far have not been mechanized, you'd probably be better off building a large machine that can move over several rows at a time and have a dozen arms underneath it that do the actual picking, versus a bunch of humanoid bots. The human form is exceptional don't get me wrong, but translating it to machinery has been one of the greatest fixations of our species since robotics of any sort came into being, and we still haven't really managed it. It must be pretty fucking hard to pull off.
- ben_w 2y ago> You're applying logic to what is largely an emotional topic for people, no matter how many times they say shit like "facts don't care about your feelings," their policies are almost in whole, feelings, and the subject of immigration is a great example. Kinda the opposite, I'm surprised that the facts (needing immigrants) are overriding the extremely apparent emotion. > I mean, even the cheapest number Elon's thrown out for his AI-powered humanoid robot (which I can't stress enough, does not exist in any meaningful way relative to what he is selling it but nevertheless) puts it at costing years of what an immigrant does to perform the work, and that's not even going into the fact that having them out in fields vs. climate controlled warehouses is likely going to increase the amount of maintenance needed, if they could manage it at all without just falling directly over. And there are energy concerns too, you'd almost need a mobile charging station for agricultural applications of them. Aye, but I don't actually mean Tesla's Optimus[0] — I was thinking more like Unitree's G1[1], Agility's Digit[2], or Figure's… whatever they'll be called when they stop using version numbers as product names[3]. The power limits are indeed a huge concern if this were to all happen in the next few years — the global electricity supply is only about 250 watts per capita, and even if we maintain the current aggressively rapid growth of renewables, I suspect that we'll be constrained by electrical capacity until 2032 at the earliest[4], even if absolutely every technical challenge in engineering is met. ("Without falling over" is probably much easier with a human remote controlling them). [0] Optimus would have been a more impressive demo if it had been sold as remote control rather than AI: the only thing Tesla actually demonstrated that would have been an interesting AI breakthrough would have been correctly parsing speech in an unconstrained environment, everything else was meh and that they even implied it was AI suggests that they're not even sufficiently aware of the state of the art to pull off a smoke-and-mirrors trick. [1] https://shop.unitree.com/en-de/products/unitree-g1 https://shop.unitree.com/en-de/products/unitree-g1 [2] https://www.agilityrobotics.com/get-started https://www.agilityrobotics.com/get-started [3] https://www.figure.ai/about-us https://www.figure.ai/about-us [4] 2032 is when PV and wind roughly equal current global electricity consumption