The shelves are stocked the same way they are picked, by having the robot travel to a station where a human places the items on the shelves. One way to organize the warehouse would be to put the replenishment/put away stations near the incoming goods (e.g. receiving area) with the picking stations on the opposite side. not sure if this is how Amazon is configured or not.

URGENT: pls stop posting

Ok, I'll get on that right away.

Just one example of many of the need to properly model the dendritic computation is the hippocampal CA1 pyramidal neurons performing 2 different computations depending on the spatial distribution of the incoming signal (and level of synchrony). Either it performs input strength encoding or feature detection. You can't ignore this and end up computing the same thing, your end result will be different.

Well, if we consider the example of object recognition happening in dendrites of human visual system, that requires a neural network all by itself to emulate. It's absolutely doable, but there is a huge difference between integrate and fire and a non-linear NN style mapping of input to output. There are other examples of dendritic signal processing which all point to neural computation really being a level below the neuron, each neuron is a network in it's own right. So, instead of 302 neurons, to properly emulate, we may need 302 * 10,000 computing nodes ("neurons").

Don't get me wrong, I like it, my point was just a word of caution about what has been accomplished with this. The neuron is constantly getting more complex as they learn more and the older assumptions about being able to emulate at that level are pretty much gone by neuroscientists.

Researchers are finding that those dendritic computations are frequently very important. One example is that object recognition happens in the dendrites in human visual system. Another example is that backwards spiking in the dendrites is a key part of learning. Maybe you can emulate the dynamic learning algorithm but I seriously doubt they figured it out and are emulating it, which means the system is not dynamically adjusting the same way the worm does.

They did not emulate the functionality of a neuron. If you read up on the subject you will find that the neuron is a network all by itself with spikes moving forward and backward, local spikes on the dendrites, the dendritic tree performing multiple simultaneous linear and non-linear computations, etc. etc. etc. They used an extremely simplistic formula that completely skips over these computations that have been shown to be very important for the proper functioning of the neuron.

Kamikazes Per Hour?

They've installed the Kiva robots in a few DC's and they are continuing to expand their usage.

99.99% of all software patents are trivial for a skilled dev and have been coded by many parties in many forms in many environments over the last 60 years. There is very little software that is non-obvious to any skilled person in the field.

1 - Startup
2 - Cash in
3 - Sell out
4 - Bro down

Why not automatically scan the printed document and email it to the person?

Good point, I better print the email reminder and place that in the three ring binder that sits behind my desk.

Apparently the information at MS is not very fluid, so this is a serious attempt at making that information much more fluider

This one suspiciously coincides with the legalization of pot in WA state

I think this is really a statement about facilities, there is a lot of wasted vertical space so most likely they will be installing something like those Japanese "drawer" hotels (where the person lays in what looks like a human sized drawer). They should be able to put a screen and keyboard in there so people can work laying down. I'm guessing they can get 3x more people in each building with this system.