Reality check from the WSJ article on Walmart distribution system automation. In the real world - not in the dreams of Silly-con-Valley, automation is hard, productivity gains are slow and it's been that way for a very long time. Slapping up huge numbers of datacenters to increase "compute" is not going to change this. That means that revenues and more importantly earnings to pay for all these datacenters won't come to fruition fast enough to keep this gigantic malinvestment-driven AI bubble from imploding. Four years into this craze and there are no signs of productivity improvements in the economy... and that should not be a surprise.