I think that's roughly right. The big-picture trajectory of automation is also, from where I see things in AI, not that sensitive to minimum wage levels, at lea

I think that's roughly right. The big-picture trajectory of automation is also, from where I see things in AI, not that sensitive to minimum wage levels, at least if we're talking anything near the levels that have been proposed so far. California, for example, is proposing to raise the minimum wage from $10 to $15 by 2022. That annualizes out to 7% a year, or probably around 5% in real terms, assuming current ~2% inflation rates persist. That's just fiddling at the margins with the economics of automation, not a revolutionary change. Tech improvements are looking for orders of magnitude cost and productivity improvements, not +/- 5%. A wage floor going up 5% a year will, at most, move up projects which were already a sure thing, making them economically feasible a few years earlier than they would have otherwise been. But it's not going to make things feasible that are currently an order of magnitude on the wrong side of economic feasibility, or that are currently too unreliable to deploy at all. And those are the kinds of changes that major progress in automation is driven by.