Slashdot Mirror
Armies of Helper Robots Keep Amazon's Warehouses Running Smoothly
jones_supa writes Amazon is continuing to maintain its vision of an automatic warehouse. Since acquiring robot-maker Kiva, a Massachusetts company, for $775 million in cash in 2012, the e-commerce retailer has been increasingly implementing automation at its gargantuan fulfillment centers. This holiday season, Amazon's little helper is an orange, 320-pound robot. The 15,000 robots are part of the company's high-tech effort to serve customers faster. By lifting shelves of Amazon products off the ground and speedily delivering them to employee stations, the robots dramatically reduce the manual labor to locate and carry items. The Kiva robots, which resemble overgrown Roombas, are capable of lifting as much as 750 pounds and glide across Amazon's warehouse floors by following rows of sensors. Because Kiva-equipped facilities eliminate the need for wide aisles for humans to walk down, eighth-generation centers can also hold 50% more inventory than older warehouses. As Amazon is doing well, the company says that increase of automation hasn't yet led to staff reduction.
A 3:35 video on youtube of their general operation, for those interested https://www.youtube.com/watch?...
Automation reduces the cost of labor to zero and the cost of product to the cost of raw materials.
Automation does not and never did reduce the cost of direct labor to zero, nor does it reduce the cost of any product to the cost of raw materials even in a state of the art factory. First off there is always a substantial amount of direct labor in any company, even a highly automated one and the amount is non-trivial. Second there is the important matter of overhead (Sales, Engineering, Marketing, R&D, capital equipment, utilities, rent, property, maintenance, management, etc) which you seem to be completely overlooking. Automation can minimize labor costs but it cannot eliminate them because it is not economical to automate all jobs even when it is technologically possible to do so.
Disclosure: I'm an industrial engineer and a certified cost accountant. I do this sort of stuff for a living.
The only thing stopping that is that it's still too expensive to build machines to do certain jobs. But that won't last forever.
That isn't likely to change in the lifetime of anyone reading this. You actually even explained why below. To eliminate all need for human direct labor you would have to invent a machine that is as flexible as a human and costs less per unit. In other words a human level AI on the cheap. That simply isn't likely to happen anytime soon. (and don't give me any BS about the so called singularity or other paranoid hypothetical dystopian futures) Any scenario where we get human level AI in a robot body for less money than a human would cost is simply science fiction for the foreseeable future.
Eventually, with the progress of technology, it will become very economical to replace workers with machines. Some jobs may require a robot that requires 20 years to pay itself off, that probably isn't worth it for a lot of businesses.
That is exactly why your hypothesis is wrong. The return on investment is simply too long for many projects to make certain types of automation economically feasible without the invention of robots with human level AI at absurdly cheap prices. To make automation practical you have to have large enough volume of product (measured in dollars) to achieve an economic return within the lifetime of the project. There are countless manufacturing projects where the lifetime and/or dollar value of the project is too short to justify complete automation and this will not change any time soon. The labor content of manufactured goods will minimize to a limit function but the need for human labor will not go to zero without invoking science fiction level advances in technology. If you need a model to look at check out agriculture. A lot of automation has gone into agribusiness but the equipment is VERY expensive (and not getting cheaper) and has not come even close to replacing the need for human labor and won't anytime soon either.
The notion that robots will replace all human workers in manufacturing is a paranoid delusion from people who aren't actually involved in manufacturing. I run a manufacturing company. I deal with this daily.
Plenty of companies (including manufacturers) have Amazon storefronts. Some of them use Amazon for fulfillment, some just use Amazon as a storefront. I don't see why MSI can't.
While Amazon's site for computer parts isn't nearly as good as NewEgg's (Amazon's spec search capability is pitiful), I've never had any difficulty telling who the seller for a particular product is. In your case, if it said "Sold By: MSI", you can be pretty sure that's who it was.
As far as not getting a shipping quote until checkout? That's pretty normal for lots of web stores. If you are going to charge for shipping at all, per-item shipping is certainly a choice, but plenty of web retailers do it differently. They can go by actual shipping cost, a rate based on total order size, etc. In Amazon's case, if the item is fulfilled by Amazon, you either go with the free shipping (or prime), or you pay according to their published shipping rate tables. If it's not fulfilled by Amazon, they just do whatever the retailer tells them to.
Personally, I find NewEgg's shipping to be the most confusing: depending on the individual item, shipping is either free (and slow), free (and less slow), per-item, or total-weight. And it's never clear which shipping rates are going to apply if your order contains items in multiple shipping categories.
Simple X-Y robots (that have been around for years) that pick regularly-shaped items off of shelves (usually decent-sized boxes) and drop them onto conveyors are pretty standard, and not that difficult. Picking up objects of an infinite variety of shapes and sizes, many of which are quite small, is something it's not possible for robots (at least not reasonably priced ones) to reliably do at this time.
This system (which brings the shelves to the workers, as workers are MUCH better at plucking small, irregularly-shaped items out of boxes) has fascinating challenges all of it's own, mainly related to traffic control, safety, and where to put the shelves after you are done. (A fixed location is very inefficient, but neither do you want to stick the shelf in the first available space.)
AMZN is not losing money because they are reinvesting back into their operations. That is not how accounting works.
I'm a certified accountant so I'm kind of giggling over you telling me "how accounting works".
That is not how accounting works. "Earnings" is a balance sheet operation
I presume you are talking about Retained Earnings which is on the balance sheet. Retained Earnings != Earnings. Earnings = Profit and that comes from the Income Statement, not the balance sheet.
"Investment" is a balance sheet operation
Investment is FAR more complicated than simply transferring items around a balance sheet and it touches the Balance Sheet, Income Statement and Statement of Cash Flows.
Think about this way – If I invest $100m of profits in US Treasury Notes, how does that affect my earnings?
It depends on why you are investing the profits into those Treasury notes and whether you are investing in them as a profit making venture or merely as a place to park cash intended for other uses. The accounting is substantially different depending on the purpose of the investment. Furthermore the effect on earnings can be substantial in future accounting periods which I should think would be obvious.
If I invest $100m in property, plant, and equipment, how does that affect my profits?
The effect on profit depends on the return on the investment though in the immediate period the effect is either neutral or negative most likely. You're making the mistake of only considering the current accounting period. If you buy a building and capitalize the expense it has some effect on the current accounting period but the real effect on profit is depends on what you can do with that asset. It may reduce future profits or enhance them. Without more information no one can say more.
What if I paid out a dividend?
Then you are returning earnings to the shareholders rather than reinvesting them in the company or other assets. Basically a dividend is an admission by the company that the expected return from available investment opportunities is low. The company is foregoing future opportunities so the long term effect on earnings to the company is either neutral or negative.
It does not – in all cases one's profit is 100m.
Not correct and even if it were you aren't considering the net present value of that $100m.
The issue is that AMZN is trading profit margin for market share. Expanding quickly today to reap the profits of tomorrow – in theory.
Which is another way of saying they are investing in the business. Amazon is introducing products (tablets, phones, etc), investing in infrastructure (warehouses, IT) and similar. They have a long term perspective and aren't worrying about quarterly results. Perhaps this will burn them in the end but my thesis that they are investing in the business remains intact.