Anything that improves the livelihood of the warehouse workers is a win in my book.
By the end of each shift I would be in a daze, wondering why the hell I thought this would be a good job, and by the end of my first month I started using heroin again (I had been clean for six months up until that point). Now I'm not saying I would have stayed clean if it wasn't for Amazon, but it definitely made relapse happen a lot sooner. And besides, the drug made me work like a machine. The 10 hour shifts that slogged by in sobriety began to blow by with blissful alacrity, and my numbers were excellent to boot. And eventually, well, I was fucking hooked on heroin again and I knew I had to stop. So I quit to focus on my recovery. It was the best decision I've ever made and I've been clean for over three years now.
That might be a stupid story, but the whole time I worked at Amazon, all I could think was "they are going to automate this job someday, and thank fucking god for that. What is taking them so long?" No human should be made to do such mindless work. It sucks. And for $13.75 an hour, it definitely isn't worth it.
This is how I felt as an undergrad working in a research lab. Endless, annoying pipetting. Labeling tubes. Bitchwork that the postdoc didn't want to do. I kept thinking... there's no way they can't come up with a robot to do this.
My guess is the technology to do it is there but the motivation isn't - most undergrads (me included) work in research labs for free with the hope of getting into med school.
Luckily, I'm in med school now so I don't have to worry about being taken advantage of in that way. But every time I think about how the system does this to so many premeds, it pisses me off.
You may enjoy https://en.wikipedia.org/wiki/The_House_of_God
And yeah residency is kinda of bullshit. Not looking forward to it :(
I don't think anyone goes into medical school these days without knowing about residency, though. At least, I would hope not.
At least I won't feel like a robot could be doing better than me, though? Although that might soon change with AI advancements, depending on what I specialize in.
Overall, if I could do it over again, I would have just gotten a CS degree. I'm afraid to read House of God because I feel jaded enough as it is.
Are you a physician btw?
While engineers (like myself) want to deal in absolutes, that is not the nature of disease and people. As such, AI may become a tool, but don't expect an AI in the near future to take over diagnostics. Additionally an AI isn't going to bring compassion to patients situation. To that point, learn the difference between empathy and compassion.
It's not that the motivation isn't there to get robots to do the bitchwork in science, it's that the money isn't there in an academic setting.
What I meant was - the motivation isn't there because you can just get people to do your bitchwork for way less than it would cost to invest in a robot.
PIs everywhere: “Why would I buy this robot when I could get an undergrad to do the work for free?”
Although if they hated taking the time to train undergrads maybe they’d give it a shot
Then again, they don’t train the undergrads, postdocs do
If censorship is the case, then may I make the point that instead of censoring people critical of y combination backed products/services, maybe you should listen to your potential client/customer base, as my opinion with this product is shared by others in academia. Your money would probably go a lot further.
There are better systems out there, but not every academic lab is willing to fork over the cash.
In the US, nobody is _made_ to do such work. However, until it is automated there will always be a price for it in the labor market.
> And for $13.75 an hour, it definitely isn't worth it.
Presumably the mindless nature of the work increases the wages. Do you recall the minimum wage in that region at the time?
I’m trying to arrive at is a logical sequence of steps like: the less desireable work is, the higher it’s price in the labor market. It’s more profitable to automate more expensive labor. Thus, by quitting you shortened the time-to-automate.
A billion of people around the world would be happy to work with that hourly rate.
http://pubdocs.worldbank.org/en/771271476908686029/Segal.pdf suggests that the number to be in the global 1% is around $50,000 PPP per person in a given household.
If that were the case, I would be happy to do anything, including relocating to the US and working as an Amazon's robot.
If we talk about working for that rate in their native country, I think it can be up to 5 billion. And a billion more will take the job without being too happy.
People's values can and will diverge on whether the bottom-line driven approach described above amounts to an ethically defensible system, but we shouldn't be under illusions as to how decisions like this are made.
Isn't the idea behind economy that if everybody tries to improve their bottom line then everybody wins (even if it was not intended)? And this story seems to support this idea.
This idea is given fairly often as a rationale that economies should be organised around capitalism, yes.
Whether the principle holds up in practice seems, to me, the kind of question that would be answered best by seeking a very wide range of data, rather than by attempting to draw conclusions from a handful of news articles.
Some areas of the economy that are often proposed as counterexamples to the above principle include healthcare and the environment, where critics claims that 'everybody trying to improve their bottom line' results in inefficiencies & unnecessary complexity, socialization of costs (eg pollution, morbidity & mortality), or in some cases direct harm.
Consider a hypothetical world where, every time a job was automated away, the person who lost their job got a pension equal to the difference in operating costs (they were paid $15/hr, the machines cost $6/hr, this person gets a $9/hr pension). In this world everyone looks forward to having their job automated and many people try to replace themselves or improve the efficiency of the machines that replaced them. And if this hypothetical world started doing this 100 years ago, everyone there is retired by now.
There are some flaws with that plan, but the gap between that world and ours shows that people are justified in being upset at the effects of automation.
I do understand the point about people being upset at not receiving a pension for the difference in their job being eliminated, but that's different from working toward less pay for not working.
Under the hypothetical, even if you want your current level of pay, you may want to automate yourself out, because that frees you up to spend your time on some other job, getting both paychecks.
(And if you really didn't want to have your job automated away... well it still works out better for you than our system of firing you and then nothing.)
If the best job you can get is a factory job, and new technology comes along that can eliminate your job, you are definitely not going to be so happy that your kind corporate overlords have freed you from the drudgery of work.
Another answer would be to have products packaged in standard, or at lease square, boxes that could be thrown into larger shipping boxes.
Or do away with the warehouse altogether and have products ship directly from manufacturers. I wonder if anyone has tried that?
1. Amazon Direct Fulfillment (dropship): Order is shipped directly from the manufacturer to the customer. Amazon pays for the shipping label and contractually "own" the item once it leaves the MFR doors. Amazon incentivizes Manufacturers by paying the freight and loosening shipment deadlines. 
2. Ships In Own Container (Amazon SIOC) certification puts the responsibility on the manufacturer to certify their products can be shipped without extra packaging. MFR pays for the testing/certification from a third-party. After July 3st 2019, Amazon will charge/fine manufacturers $2/unit received that are NOT SIOC certified and greater than a certain dimensional threshold. 
Its a tough economic balance for Amazon because they want to keep shipping stuff, but many MFRs will literally stop selling items to Amazon on July 31st because the costs of SIOC certification are greater than the estimated product profitability.
Source: I am a consultant for Manufacturers that sell to Amazon. 
Sorry, but how can this be a good idea? I'd hate if I ordered something expensive, for example a digital camera, and every porch pirate (or shady parcel delivery subcontractor) immediately sees "this thing is worth a couple grand", compared with the usual Amazon grey packaging.
More common these days is for manufacturers to produce items, and separately purchase very durable over-boxes that are sent to eCommerce customers like amazon/walmart/bestbuy/etc.
Shady subcontractors are Amazon's problem.
Blair, how do you Amazon's private label effort ? How will that end, would we buy mostly Amazon exclusive brands ?
Their retail team used to consist mostly of buyers. Most of those resources have shifted to private label.
Imagine a thousand people previously managing external brands now focused on finding/creating private label brands with the data of sales volume at their fingertips.
They have something like 150 private label brands right now.
1. Amazon changes the market(a huge challenge). it's now easier for manufacturers and other retailers to do e-commerce.
2. Amazon becomes the best in the world in packaging, and keeps the skill to itself. It's now harder to compete with Amazon.
Naturally Amazon prefers #2.
Of course, there's no reason #2 can be a check-and-balance against #3 becoming too expensive/independent.
Why would Amazon comoditize its greatest competitive advantage? (Yet, it seems that they are, so the question is not rhetorical.)
It's a lot easier for a manufacturer to send a boxcar of toilet paper to a warehouse (Amazon or otherwise) than to mail the case to you. Lacking Amazon or Safeway, they'd probably just recreate a similar system in the form of a co-op.
In any case, I don't doubt they've gamed and spreadsheeted the bejeepers out of dropshipping, returnable containers, standardized containers, third party (USPS) vs. self shipping vs. drones, etc.
I've been studying this lately, and it really seems to be more than that. Consumers love to compare goods based upon initial purchase price (or even loan payment price) with relatively little regard for the long-term cost of that good.
As evidence, take a look through AutoNation's annual report, in which they clearly make more gross profit on part sales & service, finance, and extended warranties than on actual car sales. It is to the point that it is questionable if the business would be viable only on the basis of new & used car sales alone.
A business that chooses to try to make profit without these addons will ultimately have to sell more expensive cars. Will consumers understand the difference?
Nothing wrong with that, but it could very well be the first clue that Amazon is behaving more like an incumbent than a disrupter as described in the Innovator's Dilemma.
If items are usually/never ordered by themselves, packing in a larger box might still always work out to be cheaper than paying for shipping for two boxes.
There's already enough of a problem with theft of generic 'Amazon'-branded boxes -- can you imagine if things were regularly shipped with $high_value-brand labeling on the outside (if not just glossy full-color retail box)?
Now they have included a free option to wrap inside another bigger Amazon box. I guess they got too many damaged items and complaints from customers.
This would also be very inefficient because you would have to store the postal-ready boxes at the fulfillment warehouses in an inefficient way instead of just putting them into bins. When shipping products to customers, also, it's often more efficient to pack multiple products into a single box to ensure both durability and cost efficiency.
Most of the goods available on Amazon also do not come direct from the factory. They come from various intermediaries. Not all products are shipped from the Amazon FBA warehouses, either. So, there are many reasons why it is impossible.
It is possible to ship to customers direct from the factory, as some other comments have mentioned. However, as you might imagine, this is impossible for many businesses. Processing customer orders is time and labor intensive particularly because it involves making many tiny shipments versus shipping out entire truckloads at a time only. Furthermore, contract manufacturing is very common. A given factory may be manufacturing products for many different brands at the same time and a certain brand can have its products manufactured in entirely different countries at the same time.
One problem is that showing what's inside helps thieves.
Erm... manufacturers have been shipping their own products for years... but they're not very good at it, compared to Amazon?
As in, they offer standard slow shipping (vs. same-day or next-day with Prime), shipping you have to pay for per item (vs. Prime), shipping that is often unreasonably expensive (vs. Prime) and have other issues resulting in a worse UX e.g. poorer website design, and/or signing up for individual accounts with every website (vs. Prime).
Even as big as they are, I don't think they have that much power. So maybe they could do that with some items, but they're still going to have to package some items.
In some cases, it might even be less efficient to do it this way. It means spending more up front on packaging, so the time value of money is a factor. Also, some products never sell, so why create packaging for them that will never be used? And an already-packaged product may take up more space on warehouse shelves. It might only be 10% more space or something, but 10% of everything adds up.
Also, sometimes people order more than one item. It is probably more economical to ship everything in one box. (Postage would be lower, and it might still cost less even if Amazon does the delivery themselves.)
They still put your FFP box into another bigger shipping box.
"Amazon.com works with manufacturers to box products in Certified Frustration-Free Packaging, which reduces the overall amount of packing materials used."
"It’s designed and tested to ship to customers in its own packaging without the need for additional Amazon packaging."
And something that Amazon does as well. If you're a third-party seller:
But this more-or-less hidden from the customer if Amazon is the seller.
Isn't that drop shipping?
Because Amazon saves significant money with fewer boxes.
Here are the savings
- Transportation Volume (no matter the vendor)
- Potential trips from warehouse to customer
- Package damage or loss
- Cost of packaging (they can charge their supply chain less OR make more margin if they can decrease this)
Even if Amazon uses their own delivery service.
Importantly, one of the major promotions they offer is delaying your 2 day guaranteed shipping until all items are ready to ship.
Amazon goal is same day delivery using low paid, Uber type contractors.
In my opinion, the first one to successfully do this on scale will be the next Amazon, just even bigger.
That being said, packaging for customer shipments is inherently different from B2B. The later usually uses some sort of bulk-level packaging for pallets and such. Cheaper for big volumes but unsuitable for B2C. Still, Amazon is pishing Ship-in-own-container (SIOC) wherever possible.
DISCLAIMER: I used to work there.
The seasonality of demand and inherent monotonous nature of the work itself makes it near-impossible to scale and retain the workforce required to fulfill orders.
You have to automate what can be automated, and continue to aggressively hire just to keep up with order growth.
It's actually pretty neat - creates boxes to size on demand.
Doesn't seem like this would be possible with humans at a store, and could really cut down on box inventory.
Amazon put it in one of their boxes, which did not have any labels about the package contents. Thieves are more likely to steal a box that says "expensive computer part" versus an unlabeled box.
Not to say that's always the case, but it depends on what the products box is like.
And maybe better protection for items not explicitly boxed to support mail-order/delivery? (Although as noted elsewhere in this discussion, maybe Amazon could use its power to require boxing better suited to their needs from the manufacturers...)
Amazon has about 100,000 robots now. Most are Kiva machines, moving storage racks to human pickers. Here's what that looks like. The pickers stay in one place, take things out of bins brought to them, and put them in outgoing bins. A laser pointer tells them what to pick. That's almost 10 years old, but Amazon has not yet been able to replace humans in that job.
It is pretty slick, and an obvious improvement to packing things in over-sized/generic boxes.
As is the case with your amazon example, putting the items in the storage containers and picking them for shipping is still handled manually. I'm really curious as to how this compares to the moving shelf idea efficiency-wise and how a possible "solution" (if it is a problem that needs solving...) for the picking could look like, e.g. standardized containers for small items that could be handled by a machine or something like that.
The Kiva system Amazon uses has a bigger size range. It's also more rugged. The only fixed infrastructure in the storage area is a flat cement floor with bar code tags for navigation. Shelving units and robot are easily replaced.
AutoStore has a human-carrying vehicle which can run on the AutoStore rails. It's 4x the size of the regular robots. So when something gets stuck, there's a way to unjam the grid. Still, if something falls out of a bin, it's a much bigger headache in AutoStore than it is in Amazon/Kiva.
You can think what you want about Amazon but I’m not complaining that I have 100’s of options for my phone case, all for 1/5 the price of the previous days buying them at Verizon or Apple.
Edit: would be awesome to hear about Amazon’s pricing practice from an employee
Do you have any source about this? If the costumer is already willing to pay a price there would be no point in lowering the price just because you automated the process. Or better said, theres no point in investing in machines if your profit margin is not going to get wider.
A profit-maximizing firm sets prices that maximize its expected profit. This seems tautological, but it's important and I think it's implications aren't appreciated.
If your prices are already beating the competition, and you already have a solid reputation for low prices, there is no incentive to continue lowering prices, even as you cut costs. This is why monopolies are considered "bad", because they eliminate incentives for firms to care at all about the welfare of consumers.
In undergrad economics classes, you might learn that prices are set equal to marginal cost (including opportunity cost). But that is only true under specific and relatively strong equilibrium conditions, which I don't believe apply to Amazon.
And once adopted by competitors, when everyone's margins have increased, these companies do compete on price.
This process doesn't happen immediately but it does happen.
Can a competitor actually catch up without this data?
If they could reach many more customers by lowering prices, even a monopoly could potentially find it profitable to continue lowering prices as they cut costs.
A monopoly could also continue to cut costs to head off the threat of antitrust litigation.
The point I was talking about has to do with the internal labor publicity at Amazon which rarely affects the consumer.
Yes there is. Lower prices means that consumers will buy more.
Unless you are going to argue that the elasticity of demand is perfectly inelastic. Such cases are rare.
I dont see why they would ever sacrifice this, when they can easily keep growing by taking this same mindset into new industries (groceries, cloud computing, pharmacies).
And no, I'm not implying that they are a monopoly.
I actually suspect that one big competitor to Amazon will not be another online store, but local stores providing free or cheap local delivery, and by extension services like Postmates.
And yes, most of our history is full of examples of newer technologies lowering cost, that's kind of the whole reason that or civilization is more advanced (materially) than it was in the past.
I think with Amazon building out their delivery and subscription options, they can hold back a lot of pressure to share the savings from automation with consumers.
Your margins is someone else opportunity. I believe Jeff said something similar along the line. By having a economy of scale in Warehouses, Automation, Transportation and consistently lower margin Amazon achieve a moat that can not be easily replicated even with billions of dollars.
But, on the other hand, if there are little to no competition, like a monopoly, then Amazon is obviously free to just keep their prices and increase their margins.
Better attribution of cost means more margin to play with and cost savings that can be pushed onto customers in a fairer sense - especially if your pricing is automated.
Interestingly Kiva may then selectively reduce the costs associated with bulky goods as they're probably better at moving big stuff and travel faster/shorter.
The mort apt comparison is the cost of the phone case when packed with human labor vs. packed by robots.
I also think that the full extent of this conversation, that consumer interest is currently in opposition with our responsibilities as a society, is rarely considered by the mainstream media as well.
It's not Amazon's role to provide jobs for Americans, it's their job to increase profits -- it's the government role it tax Amazon appropriately and use those funds to support Americans disrupted by the changing landscape. It's time for all of us to accept what the future brings and act appropriately.
Reuters only 12 days ago published, "Amazon dismisses idea automation will eliminate all its warehouse jobs soon." 
Automation will continue to increase, but we're still seeing numerous job openings in warehousing. 
For now, it looks like robots in warehouses are displacing work, not replacing jobs. We do believe that automation and robots will eventually lead to the creation of more stable jobs with less turnover.
I visited either YYZ4 or YYZ3 a few months ago and I was incredibly disappointed. For a technology company, their warehouses were anything but sophisticated.
Their storage robots were surprisingly lackluster, their outbound/cubiscan conveyor belt was incredibly slow and often stopped.
Of course, I am just an outsider looking in, and I have no visibility into the underlying software that optimizes these process, but when you look at companies with fulfillment tech like JD, you really wonder how Amazon's retail arm is able to survive at all.
If you're interested in helping make it better, Amazon's got an office full of developers here in Toronto working on these problems. It's a lot of fun, the pay is better than most in Toronto, and the 7 years I've been here have almost always involved me being mentored by really great folks. Feel free to reach out.
120 Bremner, between the ACC and Rogers Center.
This alias at company name dot com.
I had a coworker join in January. Very talented, great to work with. Wasn't really into what we did. So he found a team doing what he was interested in and transferred last week. Sorely missed, but better to have him happy in the company than unhappy until he leaves.
The culture itself is something special. The leadership principles mean something here. For certain types (of which I am one) it really works well. My only displeasure with the culture is the times the company makes choices that seem to go against the LPs.
It's great because Google et al tend to keep a lot to themselves, not offering hardware on the market (for example full version of the TPU).
I'm surprised that number is so low. There's only a couple dozen people out of 2000+ that actually put orders in boxes?
After 6 months or a year, the lessons learned about what worked and didn't will get rolled into the next round of automation, or an entirely different approach will be taken instead.
I used to work at a water filter factory, and this is exactly what happened there. One production line used the old, 7 person crew, and a new, more automated one was brought in for a different filter size that needed only 2 or 3 people. The ways that the two lines could cause batches to fail varied, and the company gave it some time to figure out which line ended up being more cost effective over X investment payoff period.
I didn't stick around long enough to see which one "won" out, but the automated process wasn't a clear winner by the time I had left.
> A key to its goal of a leaner workforce is attrition, one of the sources said. Rather than lay off workers, the person said, the world’s largest online retailer will one day refrain from refilling packing roles.
> The machines have the potential to automate far more than 24 jobs per facility, one of the sources said.
The answer is not actually obvious. There's a lot of research into efficiency, which is almost synonymous with job destruction.
Edit: Thanks for the answers below. I was thinking of academic areas of research. For example, Macroeconomics, Labor Relations, Governance, Management Theory etc.
People create new jobs about as fast as they destroy old ones, and that has little relation to the kind of jobs and technology involved. There are some second-order factors that create the bad kinds of unemployment, but the ones that currently are relevant have little relation to automation.
That's not saying that automation unemployed can not become relevant. It certainly can, and has done so a few times in history, with results varying from bad to disastrous. I'm just saying that job creation and destruction aren't different phenomena, and you delude yourself if you think somebody can study one without also studying the other.
But I do think things like UBI are still necessary, hand made items are always going to be in-demand, but unless you're selling only to the 1%, it's hard to charge enough.
And they are faster too
"They crank out 600 to 700 boxes per hour, or four to five times the rate of a human packer"
Automation is what we should be striving for, but only if workers and society see material benefits from it. ("Your labor potential can be reallocated to a position of higher need"/"You now have the opportunity to retrain" isn't a material benefit.)
How many "paid positions" does Microsoft Excel replace every day? A single for loop could replace a thousand jobs doing manual computation.
How many farm laborers are replaced by a single piece of modern equipment?
You can't "strive for" automation, while at the same time essentially locking in labor costs by taxing that automation as human labor.
I don't know, could you tell me? I don't think that we actually can point to people working in the position of calculator who are rendered unemployed by Excel version bumps. Arguing that we'd be putting Bob Cratchet and other clerks from the 1800's out of work rings of reductio ad absurdum.
MS Excel seems to lack the one-to-one correlation between a worker on a fulfillment line putting things in boxes and a robot on the same line doing the same work (and doing it more efficiently, since the 'bot custom-fits the packaging to the product).
The farm and mining automation example do ring truer, and in each case we've seen large increases in productivity (as well as consolidation) while seeing people in these communities falling out of work and not being able to get back into the labor market. To pick the only mining county I can think off off the top of my head, the effect of automation continues to be felt in local economies:
> You can't "strive for" automation, while at the same time essentially locking in labor costs by taxing that automation as human labor.
Office jobs are disappearing by the millions, just like physical ones. Lucikly, at least until now, new ones were created in their place.
Thought experiment: what if we (only) tax pollution? All taxes end up being paid by the consumer anyway. But this get the incentive straight to deliver products/services "cleanly".
Just a thought...
Of course, I understand how valuable short-term living arrangements are for a number of reasons; my issue is with long term landlords, who can evict a tenant who's been there for decades at any whim ... and they reap the profits, while the tenant is out on the streets with nothing to gain from their many years of payments.
I don't think taxation on automated labor really gets us anywhere. Wealth tax makes a lot more sense with fewer loopholes.
Mechanisation delivers benefits to society at large, although I agree that at an individual level redundancy is harmful. This harm is a humanitarian issue, which the whole of society is responsible for, not an ill which employers should be specifically penalised for.
This box robot saves material (little wastage of cardboard) and labor costs, and doesn't ever call in sick. Why wouldn't they want robots for everything?
Any machine eventually calls in sick when it undergoes maintenance and failure. Of course the failure modes are different than with human labor, but the point is machines also fail.
I think that robots are more fungible and predictable in failure than humans. You can plan around that, like plugging in spare hard drives and machines when they fail so throughput is not affected.
It always seemed to me that you could wipe out far more jobs than Amazon is likely to by simply removing the need to weigh items in grocery stores (like our local discount market). Generally, supersizing retail and the underlying supply chain has changed matters far more than laying off some warehouse employees.
I suppose that outsourcing has already 'automated' scads of US jobs in any case. The interesting automation will hit people who type for a living. Dunno how it will all it turn out, but I've got my comfy chair and a cocktail waiting.
That is a major feature. The inability of human packers to properly size boxes is merely wasteful for many products that come in their own packaging, but damaging and costly for Amazon's original product, books.
We've had so many book orders arrive damaged that we hardly ever shop at Amazon for books. They just toss them in a box that is too large, with some wadding material, and sent them off. Almost every time, the corners, covers and pages are damaged or torn. Returns more than once are frequent.
So, I'm very glad to see this on a micro level, and also on the macro-level of reducing waste, both in packaging material per box, and repeat shipments for shipping-damaged goods.
Billions of people are employed by outsourced manufacturing. When they are replaced, no one will be able to afford any kind of safety net for them. There is not enough money to pay for every one's stuff. China, India, Africa, and S. America have all grown on the backs of this outsourcing. When that goes away, we will have many billions of angry, hungry, restless young men.
That is how wars start and how civilizations fall. Previously, countries have typically dealt with an inability to feed their people by starting wars. Those that have not have typically collapsed or starved them (which, in turn, results in collapse). There will need to be some kind of solution. Even if they are all fed, people without purpose are always trouble. As the saying goes, idle hands do the Devil's work.
Here is the product link:
Here's the YouTube link.
Though, this is only the beginning. A few thousand jobs now, the majority of them later.
It's also totally possible that each warehouse will be made twice as efficient per employee... but also deliver twice the merchandise... maintaining the same total number of jobs.
Kind of the same way that automation of the past 100 years across the economy has still resulted in record-setting employment right now.
For example, if I were to be chronically unemployed and quit looking for jobs altogether, I would actually help the unemployment rate look better since I'm no longer part of that metric.
Part of the problem can be attributed to automation in jobs and job duties, thus requiring a smaller labor force. For those who were laid off by automation in manufacturing or other sectors, many would likely have a hard time finding a job in their established career. Expecting most of such workers to get a job an an on demand skill like maintaining machinery or software development is a herculean task. Most likely they are the ones who fall prey to the poor labor participation metric.
More importantly though, we're on the cusp of some pretty hairy water. Imagine we drop all of the truck drivers, taxi drivers, Uber and Lift drivers tomorrow. How many millions of people will that be - where can they go?
I'm not against automation in the slightest, but to think it won't have an impact on employment seems farfetched. Imagine a robot tomorrow, humanoid capable of basic tasks. How many millions of people would that displace almost instantly?
I don't think the past 100 years even remotely indicates how we'll handle this as a society. We aren't talking about taking away one job. We're talking about taking away a skill category. No more drivers. No more basic manual labor.
And that's not even talking about how growth would need to expand to handle the increase in workforce. Imagine our workforce doubles tomorrow - who is going to consume it all? Do we even want more consumption?
I don't think it's as cut and dry as looking at the street cleaners and how they handled it when horses stopped being used transportation. We're talking about fundamentally categories of work becoming obsolete, and having no jobs that some people qualify for.
With that said, I'm curious to live in a world where the population is forced to be educated. No low skilled jobs might have a benefit of (basically) no low educated people. However I am worried about how it will affect the current populace.
We did it with farmers. 80% of jobs in early 1800's, close to 40% in 1900, 20% by 1930, 5% by 1965, and 1.7% now. A whole "skill category" like driving doesn't even come close to farming.
Automation, and even displacing drivers, is going to continue happening gradually, not overnight. E.g. only the easiest driving jobs in the easiest areas are automated first, then gradually more difficult ones... There's no "overnight". It'll take many years, and it will be fine just like it has each decade over the past century.
That this isn't special, that we're not "on the cusp of some pretty hairy water", but rather that this is just par for the course.
I just don't see any statistics that suggest the scale or rate will be anything close to larger/faster than our economy can normally handle. If someone can show those comparative statistics exist, then I'll change my mind.
Ie, how many millions of food service workers, manual laborers, and drivers (for trucks, taxis, etc) could move upward in career depth? Where would we expect them to be absorbed?
I definitely think this is hairy waters, in the sense that I want to see studies done. Planning for how it can be handled. Etc. Calling it fine in my view would suggest we don't even need to study it - which seems unfair. This degree of automation, assuming we can get over some AI hurdles of course, seems unparalleled - at least enough to warrant real data.
They will be forced to acquire more education or skill oriented trades.
The concern is if humanity will be able to adapt to something it has never seen. Where creativity and thinking oriented jobs become more and more required.
If the company doesn’t create the job, the tax break is reduced by a certain percentage per job.
Usually, no, though the public justification for the tax breaks is usually in terms of expected jobs.
I never had any scuffs or damage on the hundreds books I've bought through Amazon though. How bad are the scuffs? My apartment is close to an Amazon outlet factory so it doesn't have to very far to arrive.
Seems like marketing copy regarding Amazon, possibly being behind in implementation of a common solution.
In this thread people are suggesting UBI, but I can't see that as anything more than an economic uncertainty (trying to predict how market forces will respond and adapt), which will be exploited by the oligarchy that is already exploiting the systems we have now (as you might expect anyone to). To think we could successfully create a good UBI system despite the fact that we can't get other very important pieces of governance right in many parts of the world seems ludicrous.
People are suggesting that new jobs will be created as jobs are automated but there's no proof of that except history and this time seems different -- humanity has never had the tools it has now, the step change in ability/utility over the last few decades is insane. In a world with high levels of automation, where does advancement come from? Up until now, it's mostly been differentiation through some form of work. Do creative endeavors become the new work/only way to differentiate if we spend most of our time on leisure? But then what if someone automates that away too?
It seems like once we evolve past work the economic system and the big & small players in it will freeze in place. If you had enough capital to be on the right side of the split then you get to stay on that side, but if you didn't you seem all but condemned. Dystopia seems inevitable. I'm not saying that I think work is good in and of itself but it's disconcerting to happens when that particular music stops -- it's run so much of the world up until now.
Another thought -- the good 'ol revolution/riot path is going to get closed down in a few decades I think. Weaponization of autonomous machinery will happen, and the efficiency of oppression will reach untold heights. Security will be a highly sought skill and those most able to pay for it (with resources/money/whatever) will be able to purchase it.
Sure hope I'm wrong about all/most of this.
If machines can do it, they should.
There are also some products that are marked as "fragile" and require a minimum amount of empty space in a box to prevent them from getting smashed. The exact definition of "fragile" and amount of space required is some kind of arcane black magic.
there are not enough low-skilled, difficult to automate jobs that can sustain the population.
Automation is a scary specter off in the distance but it's not as actively harmful to people's livelihoods.
> According to our estimates, one more robot per thousand workers reduces the employment to population ratio by about 0.18-0.34 percentage points and wages by 0.25-0.5 percent.
> No single factor such as tech-driven worker anxiety determines local political behavior. But there’s no mistaking that districts that voted Republican in the 2018 election are subject to higher levels of automation exposure.
There is far more to shipping costs than distance. A split shipment (two items in two boxes) will typically increase costs for an order by $3-6, whereas receive/stow/pick/pack labor costs are only on the order of $1/unit. And yet, nearly every order I place comes in multiple shipments.
Kiva has O(n^2) efficiency degradation with respect to the size of their fulfillment centers. They solve this problem by having smaller fulfillment centers. But that also means more fulfillment centers, which means more split shipments, and much higher shipping costs. They're literally spending dollars to save pennies.
Reverse logistics are incredibly expensive. I'm talking $10-20/unit in some cases. The Amazon marketplace, and their associated quality control problems with fake reviews and counterfeit products, have increased their product returns by several fold over the last decade.
15 years ago, Amazon's inventory turns were something like 17, now they're 6. That directly translates to increased holding costs...for every sqft of storage space they needed in 2005, they need three today to ship the same amount of product.
And fulfillment costs aren't the only place they're making dumb mistakes with priorities. They did their whole HQ2 search shenanigans, only to settle on the two cities that offered the highest tax subsidies. But if they chose to instead prioritize lower cost of living, they could have saved an order of magnitude more, considering the fact that they are planning on having several tens of thousands of high paid workers in their HQ2 cities. The COL increases alone from locating in expensive cities with obstructive housing policies will eat up far more than those tax subsidies are worth.
In the funniest of ironies, Amazon's growth-first mindset that defied Wall Street expectations has led to an invariably laser-focused expectation on revenue growth that they are now a slave to. The P/E premium that Amazon has relative to its industry competitors, carries an implied expectation that Amazon will one day be ~10x bigger than it currently is. And Amazon's equity-first compensation philosophy means that their compensation budget inherently relies on the stock price. Therefore, any decision that may be in the long term best interest of the company (such as dumping the brand perception nuclear bomb known as the Amazon Marketplace) is now no longer feasible: it will cut revenue, killing growth expectations, lowering stock prices, and forcing Amazon to pay more cash compensation, which will just cause prices to tank further.
If we were to judge how Amazon is doing based off of how well their PR releases jive with wall street ideas of what good business decisions look like, they are doing everything right. Unfortunately, wall street ideas are shit for actual profitability. And Amazon is suffering for it.
It was very impressive. Watching the yellow plastic tubs flying around on conveyor belts, being sorted into different tracks automatically. It reminded me of the scene in "Monsters, Inc." where are the bedroom doors are flying around on different tracks.
Work was divided into different stages:
1. Taking delivery from supply trucks, checking that contents were correct and in good shape. Contents placed into tubs that were conveyed to multiple different areas of the warehouse.
2. Storing items. Tubs of incoming items were delivered to a storage station. Robots brought storage pods (6-7 feet tall) with many cubbies up to the station. Different pods had different sized cubbies. A storage worker would scan each item from the tubs, and choose some cubby to put the item into. It looked like a camera watched which cubby they stored into to record it. cubbies could hold multiple different items. Apparently they are careful to never put two similar items in the same cubby, to make it easier for the next stage:
3. Picking items from storage. When an order needs to be fulfilled, the robots would bring a line of pods up to a person who would pick the needed item out of the correct cubby of each pod. All the items would go into a tub, to be conveyed to somewhere else for packing.
4. Packing the items for shipment. A person gets the items out of the tub, puts them in the correct sizes box, adds the air pockets for protection and seals up the box, which is then conveyed to another machine that weighs them and labels them for shipment.
5. Outgoing shipping distributed to different trucks based on geo location to ship to.
We walked around the warehouse with a few guides, and headphones to be able to listen to them. Our guide was excellent in being able to explain things and genuinely enthusiastic about the whole process. At the end they gave us branded water bottles too.
I asked our guide what he thought would happen as they began to automate more and more of the jobs. He said that they would shift people over to new roles and gave an example of a time in the past when they automated some job and instead split up roles to have a separate person do QA / checking of orders where before the same person had to fill an order and check it. He also spoke of another fulfillment center in Southern California where they automated a job, then later reversed course and went back to all manual because the people at that center decided that it was done better that way for them. I was surprised that decisions over automation were made locally at each center.
We hung around the building afterward, and bought lunch from a taco truck outside, and ate at some picnic tables where many employees also ate. I talked to some of them to try to get a less biased perspective from the employees. When I asked them how they liked working there, the most common answer I got was "It's a job." They said that expectations were high and they had to work hard. Some shifts were more desirable than others. One man said that it was tough, but he was very loyal because a year before Amazon had chosen to pay for his cancer treatment and he said they didn't have to. I left with a positive impression of the operation.
I'd recommend taking the tour if there's a fulfillment center nearby. You need to register in advance for a time, but it's all free.
I agree, low-skill jobs are shrinking, but we should not expect companies to solve that. We have a government and we vote. Ideally, we should all be asking the government to propose solutions to this problem, not expecting it to be inside the corporate charter.
Robots will make the same mistake, only faster. :/
Companies solve to a given tax burden and regulatory regime. Their solutions can be influenced by changing the rules or the rewards.
Personally, I find it likely that no amount of regulation will solve the problem of creating employment opportunity for low skill workers. I think that we either need to create those jobs through the government (e.g. WPA, CCC, etc), or raise taxes to improve the amount of dollars spent supporting underemployed people, or investing in the skills.
Stepping past the "zero-sum" claim, though, I agree that corporations should take responsibility for their incredibly powerful role in modern society, but I don't see a mechanism by which this could happen. The entire point of a regulated liberal democracy is to create a legislative landscape which modifies the incentive structures of businesses to "coerce" them into providing more social benefit than cost.
There is also nothing wrong with some level of redistributism, as long as it doesn't blindly ignore economic reality (a la communism). Many of the right-leaning nerds here on HN are in favor of universal basic income for that reason.
UBI's an option -- one that I'm in favor of -- but as long as it is an idea and not reality there are still millions of people getting shafted without a lot of recourse, short of restructuring one's entire life and starting over in a new career. An expensive and emotionally draining transition that weak-ass severance packages fail to adequately compensate for.
If corporate tax payers were actually paying taxes and not sitting on a dragon's hoard of cash I might be more in favor of increased automation, but the fact is reserves being as high as they are means that automation is not so necessary in many sectors short of unreasonable and unethical shareholder demands.
Balancing that return, is of course the relatively lackluster returns on cash. Companies sit on cash not because they enjoy counting it, but because they cannot find appropriately productive uses of it. They also hold it because it helps them cope with uncertainty (of economic reality, regulation, opportunity, etc). Taxing cash is fine, but companies will just change how they balance their uncertainty with other methods.
It sounds like what you are concerned with is that labor is not receiving an equal share of the output, compared with capital. this is likely because true labor productivity has not been great (https://www.bls.gov/opub/btn/volume-6/below-trend-the-us-pro...) and also because capitalists (including pension fund shareholders, often the workers themselves) are engaged in rent-seeking on behalf of their capital. I think in many cases this reflects a shift in the role of savings from the employee to the employer. In eras past, the employee received relatively more pay compared to shareholders, but was required to save for themselves. Corporates that offered pensions had to set aside money to managed those obligations directly. In modern times, corporations tend to create retirement benefits that come from an employee's salary and the company's money, but which have been underfunded due to the assumption of future returns. This "optimistic" assumption about rate of return on the 401k/retirement fund/pension funds of today drives a lot of capital into the role of rent seeking from corporations. Corporations respond by trying to meet their shareholder demands, and corporate officers tend to be the beneficiaries of generally capital friendly activity, since they themselves are shareholders.
There's a lot to unwind here, and a lot more that I haven't written, but I think it is best to not assume that there can be a single solution to a very complex problem which results from a system built of many interacting components.
I am not saying there is a single solution to the complex problem of low-level work disappearing. However, there are many countries where the business culture is more willing to hold on to labor (non-english-speaking western europe, the specific example I am thinking of is big-box retail in France and Germany, though I've also seen this with the lack of self-serve gas stations in parts of South America); these are still successful businesses doing well in their sectors and in many cases out-competing American entrants to their markets. (And yes the reasons for their success are often better cultural fit than foreign entrants versus comparing where the spend their money) Yet the American zeitgeist seems to be heading full-on into mass automation with its eyes wide shut, when there is already so much money floating around that it doesn't really seem necessary. What labor crises are there that necessitate the deployment of automation? (I don't think rising minimum wages nor increasing threat of unionization count as they do not raise costs above what can be absorbed by capital hoarding)
+ Lots of empathy
+ The willingness to learn and be wrong
+ A ton of training
Even then, human-based customer service is still subjected to the whims of:
+ Gut checks
+ Power trips
+ Misinformation/misinterpretation of policy (whether intentional or not)
+ Lots of emotion on both sides of the interaction
Not every single customer service interaction requires a human behind it. At some point, there are simpler and more efficient ways of solving a problem besides throwing another human's time at it. Robots can easily process replacements without requiring a human being hear about why it needs to happen. They can quickly identify when something is late and proactively issue a credit.
There isn't a lot of good reason to be optimistic about the future either; chatbots and fast food kiosks were a flop (and rightfully so) and most of these other styles of interaction feel forced and constricting.
Humans may have their flaws but you can't beat the fact that they're what other humans are comfortable around.
The problem we have is with the existing low-skill workforce: how can we re-train them in an effective manner.