Hacker News new | past | comments | ask | show | jobs | submit login
Amazon rolls out machines that pack orders and replace jobs (reuters.com)
450 points by Vaslo 42 days ago | hide | past | web | favorite | 485 comments



Anyone parading around like this is bad news has never worked in an assembly line or warehouse job. This might reduce the workforce overall long-term, but honestly this is hard back-breaking long-workday labor.

Anything that improves the livelihood of the warehouse workers is a win in my book.


I worked for Amazon in what was then a brand new warehouse for a few months in 2015. I worked 10 hour shifts, from 20:00 to 07:00 (the extra hour was for lunch and breaks). My job was picking orders. It seemed easy at first (and it was, but mentally, not so much). All I had to do was stand in one place while robots brought me pod after endless pod of products. There was nothing to it, and that was the problem. The pods were divided into indexed bins containing their random assortment of products and the computer told me to get product X from bin Y. Again and again and again. I was literally just a robot arm with a human brain attached to it, and for what? To torment it [me, the head]? I was in hell.

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.


> I was literally just a robot arm with a human brain attached to it, and for what?

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.


Since I can't tell if you are being sarcastic or not, are you aware of what residency is?

https://en.wikipedia.org/wiki/Medical_resident_work_hours

You may enjoy https://en.wikipedia.org/wiki/The_House_of_God


That's why I was careful in my wording. I said I wouldn't be taken advantage of "in that way"

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?


Spouse is a mid-level provider, so I pick up some of it via proxy.

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.

https://zdoggmd.com/ama-005/


>Luckily, I'm in med school now so I don't have to worry about being taken advantage of in that way

Lol.

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.


I think we mean the same thing but we're expressing it differently!

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.


There is a pipetting robot now: https://opentrons.com/


This is awesome! But...

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


I was wondering why my previous comment was dead as quickly as I had submitted it. I did a quick Google search and discovered OpenTrons is y combinator backed.

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.


Eh, I get mixed reviews from people who own this, which is why I opted not to get one for my lab. Some labs never use it to its fullest capabilities, either and use it for show. A student is cheaper, and arguably more reliable when fully trained.

There are better systems out there, but not every academic lab is willing to fork over the cash.


Looks like the Theranos Edison. Except it works apparently.


> No human should be made to do such mindless work.

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.


I'm glad you made that decision for yourself and I think such perspectives are missing from the discourse about working-class folks dealing with addiction.


Your story seems like it's right out of the game "What Remains of Edith Finch".


I did this exact work as well in 2014. Mind numbing task. Seeing the robots and the overall warehouse system work was fascinating. But fuck that place, worse job I've ever had.


Having dealt with addiction, I can confirm that particular jobs are unusually likely to incline me to go back to those addictions, or be tempted to. The closest thing to a pattern I've found is that these jobs are either boring-but-require-attention, or they're complex and emotionally draining. The former often involved Amazon warehouse type jobs, and the latter were often dysfunctional IT departments.


Too bad this can't be outsourced.


Automation and mechanization have always been about replacing dehumanizing, backbreaking, mindless toil. Maybe not intentionally, but that was the result.


"And for $13.75 an hour, it definitely isn't worth it."

A billion of people around the world would be happy to work with that hourly rate.


Despite the fact that I might not be a rich man in this country, I am a very rich man in this world. I feel quite lucky and grateful to have been born in the USA. But to be fair, that was not a very good wage for where I live.


"accept your scraps, peasant, there's folks with no plausible way of ever enjoying them who would"


Not at the US cost of living. Nor would those billions of people be paid a that rate if their cost of living is lower than the US.


I think it's quite likely, but unknowable, that folks would accept the income:cost of living ratio of the US if it came with the associated sociological benefits. I think you're underestimating how disproportionately impoverished many people around the world are.


Well people here keep ranting about 1 percenters without ever realizing that >$100K earners are among one percenters in the world.


Actually, the threshold to be in top 1% income bracket worldwide is $32,400

https://www.investopedia.com/articles/personal-finance/05061...


This number is both old and in nominal terms. A person making 50k USD in China is likely richer than a person in the US making 100k USD in the US.

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.


Except that the Chinese person is a prisoner in their own country, and the American could move to China to have a higher quality of life outside of the polluted cities.


You're right that a few people do, although I think most people realise how deeply fortunate we all are to have won to at least some extent a genetic and [entirely separate] geographic lottery.


All in countries with a lower cost of living.


I upvoted you because at first I decided you were a king writing this and telling us to feel more lucky and I felt angry, but then I pretended you were the poorest laborer writing this and telling us to feel more lucky, and I agreed.


That is the first time someone has thought about me as a king. Even metaphorically and for a brief moment :)


Sure, if they could stay where they are. $13.75 buys a lot more in Bangalore than it does in the US.


Would you?


No, but I am not suffering from hunger, lack of clean drinking water, lack of basic medical help or violence.

If that were the case, I would be happy to do anything, including relocating to the US and working as an Amazon's robot.


To add further there will be billion who would do as a daily rate.


I was thinking about people moving to the US and working there at an Amazon warehouse.

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.


I worked two assembly line jobs. One feeding a robot with pre cut steel sections. The other wiring trucks. Both were non stop but just for 8 hour shifts. Both took physical energy and flexibility but were not back breaking (no more than the Amazon jobs anyway). I enjoyed them both and felt physically better at the end of the day than I do when I come home from a days coding. Everyone who worked with me needed the job and most were proud to have it. It was part of their identity.


An industrial job often pays a lot better than amazon ever would


If we are being honest, this isn't being installed to "improve the livelihood of warehouse workers", it is being installed in anticipation of improving the bottom line for Amazon's shareholders. There may be benefits or detriments to the workforce along the way; considerations like the "wellbeing of the workforce" are effectively given a weight of zero in these decisions.

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.


Of course, nobody tries to only "improve the livelihood of warehouse workers".

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.


> if everybody tries to improve their bottom line then everybody wins

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.


Yes let’s make everything manual in an entire factory


Arguing that we should be honest with ourselves about what precisely motivates corporate decisions to automate in the society we live in is not remotely the same thing as arguing against automation itself; neither in this specific case nor in the more general case (of automation across wider society).


Automation itself is a good thing. The way our economy is structured to respond to automation is predictably bad for the workers though.

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.


Not sure I follow this logic. Are you saying everyone tries to have their job eliminated so they can get paid less money to not work? Some people would like that, but a majority couldn't justify the reduction in their standard of living. And there are quite a few people who find work meaningful and wouldn't want to be paid to not work.

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.


I threw that part in there because I was expecting someone to confuse themselves into thinking there's no motivation for progress under such a system.

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.)


I mean that's kind of a narrow view isn't it? Sure these are not fun jobs to have. I worked ten hour shifts at an injection molding company when I was in my 20's. Ten hours of open the door, remove the part, close the door, put the part in the bin, repeat. I would welcome a world in which no human ever had to be part of a production line again. But you know what also sucks? Not being able to pay your bills, feed your kids, obtain medical care, feel even a shred of dignity. So until we have some idea of how to structurally handle this transition, and until we've developed a mindset where the people who run corporations take such things into account, such unbridled enthusiasm for the elimination of "bad" jobs rings hollow.


On the contrary, this feels like a response of someone who maybe worked part time or had a summer job in a warehouse or factory, but knew they'd have opportunity for better options later on.

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.


I'm surprised that Amazon still uses boxes. They have the market power to insist manufacturers comply with amazon-specific packaging requirements. Why not just insist that all products be packaged in postal-compatible packaging? Then the amazon warehouse robot need only slap a label on the product and throw it in the shipping bin.

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?


You mentioned two processes that have been in place with Amazon Vendors/Manufacturers for many years.

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. [1]

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. [2]

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. [3]

1: https://www.andersonassociates.net/guides/amazon/what-is-dro...

2: https://www.andersonassociates.net/2018/frustration-free/

3: https://www.andersonassociates.net


> Ships In Own Container (Amazon SIOC) certification puts the responsibility on the manufacturer to certify their products can be shipped without extra packaging.

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.


Its usually a brown cardboard overbox with a barcode that has extensive droptesting.

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.


I often see an indicator on Amazon that the product ships in packaging that reveals what it is, with an option to select Amazon packaging instead.


If you're worried about porch pirates - use a locker instead. They're already stealing some of your packages anyway.

Shady subcontractors are Amazon's problem.


Its already the reality for most TVs and furniture (at least from Target), child car seats/strollers are the other big one that's always ship-alone


I've usually seen an option to select the kind of packaging at time order is placed.


This box of rubber gloves is SIOC and hopefully isn't too attractive to porch pirates: https://www.amazon.com/gp/product/B01JGLWP3S


A side question:

Blair, how do you Amazon's private label effort ? How will that end, would we buy mostly Amazon exclusive brands ?


Yes its going to continue in a huge way.

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.


So there are 2 options:

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.


what about #3 - amazon tells sellers to ship to buyers, does no packaging, still makes money

Of course, there's no reason #2 can be a check-and-balance against #3 becoming too expensive/independent.


That's #1. It depends on the sellers accepting the extra tasks, and doing them competently while harming Amazon's brand if they don't. It also prepares those same services to sell though other intermediaries that can not compete today with Amazon's logistics.

Why would Amazon comoditize its greatest competitive advantage? (Yet, it seems that they are, so the question is not rhetorical.)


Smart dudes have wrestled over those problems for years. I think the strongest case for dropshipping would be new cars. The dealerships have presented a problem for people for years, but they are politically well connected.

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.


> Smart dudes have wrestled over those problems for years. I think the strongest case for dropshipping would be new cars. The dealerships have presented a problem for people for years, but they are politically well connected.

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?


This generalizes: most individuals can't or won't think past the first order effects of any decision. Marketers know this and prey on it accordingly.


Interesting, and could be 'fixed' by the trend of leasing cars, where you pay a fixed fee for maintenance and use of cars.


My impression is that Amazon is kind of biased in that regard. There are valid reasons to keep the warehouse logistics in-house. And they did invest a lot of energy into that. It just bugs me that Dropshopping was never really pushed while the logistical powers to be ever so slightly pushed against it.

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.


They have to strike a balance to optimize between shipping costs, time spent packing, packing material costs, claims due to damage during transit and insurance/theft costs.

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)?


True. In fact, they tend to ship big items in their original retail packaging. E.g. vacuum cleaners. Not good. In my case I got 2 damaged items out of 3 shipments with said boxing.

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 be impractical. For one thing, scuff-resistant packaging material is also necessarily less attractive than glossy and colorful packaging that holds high detail photographs & illustrations better.

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.


Or yet another complication: reconfiguring to manufacture a new product usually takes a lot of work. A product will be only produced for a week or two, but sold over a period of months or years. The factory isn't in the storage and shipping business, so why not send it to somebody who is?


I think you are diminishing the importance of brands and brand loyalty for companies. A lot of effort is spent to have boxes designed for retail; size, shape, color, etc. matter a lot for this use case. Resistance to weather and being thrown around are not the main consideration of this use case. Standard size boxes that are shippable might be good for Amazon and the consumer but companies always want the brand front and center.


Surely a box could be both shippable and branded?

One problem is that showing what's inside helps thieves.


... and costs privacy.


The vast majority (95%, including lots of oddball pcbs, fasteners, etc) of my recent Amazon deliveries have come from an Amazon van. I'm almost surprised they don't offer a reusable bin or something of that nature. It feels wasteful to throw away so many boxes.


When I was in Seattle for an internship about a decade ago, I would get books and other sortable items thrown in with my Amazon Fresh deliveries— they would show up in the same rugged plastic tote as the pantry items, and the tote would be picked up again the next day. It was a great system.


Especially with the recent news that a lot of what we thought was recyclable really isn't. I presume that clean brown corrugated is still OK, but I don't know that for sure.


oh it is "recyclable", just (for several materials) takes more energy to recycle it than toss it in a landfill. You get to feel good that cardboard gets a "new life", instead of a new one (which extracts more carbon from the atmosphere) being created, we can burn more CO2 to have a second truck travel down your street to pick up renewable cardboard and haul it around town till it eventually gets cleaned and processed (at the cost of even more energy).


I mean, they probably don't have the market power to do that. Online retail is still only ~10% of total retail (in the US). Amazon might be dominant within online retail, but they still aren't quite enough I feel.


> Or do away with the warehouse altogether and have products ship directly from manufacturers. I wonder if anyone has tried that?

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).


All products? That would be 12 million different products according to this site: https://www.retailtouchpoints.com/resources/type/infographic...

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.)



No, that's not what Frustration Free Packaging is. "FFP" is just easy to open cardboard product packaging, basically no blister packs.

They still put your FFP box into another bigger shipping box.


This depends on the product. Significant amounts of FFP are meant to reduce waste and not require overboxing.


from the description:

"Amazon.com works with manufacturers to box products in Certified Frustration-Free Packaging, which reduces the overall amount of packing materials used."

Also:

"It’s designed and tested to ship to customers in its own packaging without the need for additional Amazon packaging."


I'm not an expert, but regarding your last question, I believe it's an industry-standard practice:

https://en.m.wikipedia.org/wiki/Drop_shipping

And something that Amazon does as well. If you're a third-party seller:

https://sellercentral.amazon.com/gp/help/external/201808410

But this more-or-less hidden from the customer if Amazon is the seller.


The article wouldn't apply to drop-shipping situation. Drop-shipping means the upstream provider delivers directly to the downstream consumer, but that means it never goes through Amazon's warehouse and is thus not packaged or re-packaged by Amazon.


Right. But the question the person above asked was "or do away with the warehouse altogether and have products ship directly from manufacturers."

Isn't that drop shipping?


> Why not just insist that all products be packaged in postal-compatible packaging?

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.


I think doing away with the warehouses isn’t compatible with conquering the last of retail; same day delivery. Even next day is cost prohibitive for drop shipping as it requires UPS or FedEx.

Amazon goal is same day delivery using low paid, Uber type contractors.


I'm surprised they don't have reusable crates that they can pick up when they ship the next shipment to you - at least for subscribe & save orders. Like the milkman used to pick up the bottles.


Dropshopping might just turn out to he the biggest missed opportunity for the fulfillment arm of Amazon. Just imagine a company that successfully bridges the gap between any supplier who wants to connect and the consumer using the internet. Imagine Amazon bought or developed Shopify themselves and integrated the POS tech for any shop with their fulfillment and retail tech. Damn, that would be a killer.

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.


As someone in consumer-goods warehousing and distribution, automation is the only way forward.

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.


Video of the CarltonWrap mentioned in the article: https://www.youtube.com/watch?v=U1yw05mWnXw

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.


If it results in the elimination of oversized boxes it should save on cardboard, also reducing shipping volume. This would be financially and environmentally be a good thing. Maybe they could also include in the box sizing algorithm the optimization of shapes for efficient packing.


IIRC, some boxes are intentionally oversized, in order to completely fill the truck, so that things don't move around. That's not going to change because of this new machine.


Within reason, shipping weight is a bigger deal than volume.


Right but from a recycle / waste perspective, its a great thing. If manufacturers and distributors use less materials in their product packaging, that's less materials that end up in landfills. Not everyone recycles (even if the material, like cardboard, is recyclable), so the best way for Amazon to help fix waste is by reducing the raw output of cardboard.


Absurd how most of the items on the belt are already in a cardboard box. Seems so wasteful.


Part of it could be privacy too. I ordered a computer monitor that was in its own cardboard box. However, that box had clearly printed "Gaming Monitor" on the box. This item was sitting on my porch for a few hours before I got home.

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.


Better to support returns?

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 already (for some items, in the UK at least) makes you explicitly choose to have them re-box it in their own packaging, otherwise they just ship it in whatever box it's in. I imagine they'll do the same here, just not feed these items into the box machine in the first place.


This is a minor improvement in packaging. It's not automated picking, which will be a big deal if Amazon cracks that. They've been trying and failing for years now. Amazon gave up on their robotic picking challenge after 2017. Here's the winner of the last challenge.[1] Way too slow and not reliable enough.

Amazon has about 100,000 robots now. Most are Kiva machines, moving storage racks to human pickers. Here's what that looks like.[2] 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.

[1] https://www.youtube.com/watch?v=AljePt7Mh6U

[2] https://www.youtube.com/watch?v=3UxZDJ1HiPE


Here's a vid of the actual packing machine noted in the article: https://youtu.be/0eqtqKDbV1Q

It is pretty slick, and an obvious improvement to packing things in over-sized/generic boxes.


Yo that's crazy! custom sized packaging regardless of what's being packed? impressive


Here's how one of the biggest online electronics retailers in Switzerland handles storage / logistics for their smaller items [1].

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.

[1] https://www.youtube.com/watch?v=Scql9_7RjfU


That's AutoStore. It has a limited size range. It's most useful for electronic parts and medical supplies, which tend to fit in AutoStore bins.

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.


Articles about Amazon are quick to mention the workplace conditions or robots automating something. Rarely do they mention that machines equal lower prices for the consumers 9 times out of 10.

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


>machines bring lower prices

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.


It's worth expanding on this. Because I think it gets lost in the rhetoric, whether deliberately or by accident.

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.


The reason that this tends to lower prices for consumers is that these innovations rarely stay confined to the companies that invent them. If this lowers costs for amazon to fulfill orders, then these techniques tend to get adopted by their competitors.

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.


Might not be true as more problems are solved through machine learning. Amazon has or could generate the largest and most diverse dataset to train the object recognition problems needed to automate a warehouse. They also have the end to end pipeline to evaluate accuracy (damaged or incorrect orders, manual intervention needed in the warehouse).

Can a competitor actually catch up without this data?


Sure. The usual strategy would be a joint venture between the second, third, etc. biggest competitors to share the development (and relevant data) for such a technology.


Who are Amazon's most likely competitors?


Walmart, alibaba, jet.com, Target, and ebay all spring to mind


I believe jet.com is owned by Walmart.


wal-mart or aliexpress


>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.

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.


I agree with the general idea. In your experience, when has Amazon acted negatively towards the welfare of the consumer?

The point I was talking about has to do with the internal labor publicity at Amazon which rarely affects the consumer.


> there is no incentive to continue lowering prices,

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.


Thats the nice thing about Amazon, because of the "flywheel" (customers will always want lower prices and wider range) they've built a reputation of operating like a utility, and customers love them for it.

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).


Right, but amazon isn't known for low prices nor is it a monopoly. This will allow them to lower prices and become known for the lowest prices while still preserving their margin (if that's the market they want to target).


They aren't known for low prices? They are known for having the same prices as everyone else, with free or cheaper shipping, meaning that effectively the price is lower.

And no, I'm not implying that they are a monopoly.


They have that reputation, but be watchful. Lots of things are more expensive at Amazon, now. What they're selling me is convenience more than anything else. I can often find lower prices elsewhere but it's more work.


I'd argue that convenience is part of the price.

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.


Sure there is. If they don't do it, someone else will be able to outcompete on price, eventually at least.

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.


There's also the "number of purchases * purchase price" equation which means you can make more money by reducing price if it means (enough) more people by the product, or buy more of them.


Yeah, one of Bezos’ oldest mantras is “Your margin is my opportunity”.


Amazon has been quite the monopolistic company lately. Just because a company started winning by outdoing the competition doesn't mean that will continue to be their only way of succeeding. Even the most principled companies will adopt a few anticompetitive practices here and there when they get big.

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.


Bezos also said his top 3 priorities are lower prices, faster delivery, and wider selection.


Right now they have no need to lower prices. They can hold their lower production costs as a threat to any competitor who might be tempted to enter a price war, without ever needing to actually lower their prices.


>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.

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.


Lowering costs gives a business the possibility of cutting prices, for example to try to increase market share. Whether they take it or not depends on the situation. But Amazon often does.


If you have number of competitors, then the ones that automate things can lower their prices due to decreased costs - they will gain a competitive edge, when all else is equal - as things tend to be in this industry (price and shipping are really the only variables, for the consumer)

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.


Customers are always willing to pay the lowest possible price if they are not paying that already.


I don’t think you need an academic paper for everything


When you collect as much data as Amazon, you can start attributing incremental worker time spent picking/carrying down to the SKU level. This means that cost of worker time is less for things like books (a picker can carry many and move quickly) while a 60" TV might require two pickers moving slowly.

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.


Lol, I think Americans figured out that whole story already. We've already seen decades of American jobs being slashed to lower costs, with something tacked on about reducing the cost to the consumer.


Another advantage is there that the frequent claim that amazon abuses its employees will finally be put to rest as there will be mostly machines doing the work. No workplace abuse is a big win.


It seems less relevant to compare the cost of a case at Verizon/Apple than purchasing through Amazon. Why? It's a completely different experience -- you get to see the selection in person, ask questions, try the case on your phone to confirm it actually fits, you know there is near-0 odds of it being a counterfeit or unscrupulous seller (unlike Amazon).

The mort apt comparison is the cost of the phone case when packed with human labor vs. packed by robots.


I think you are correct in that this is often not brought into such conversations.

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.


The problem isn't lower prices, it is lack of money by consumers who lose their jobs. The displacement of retail jobs is massive and really hurting people.


The industrial revolution helped spur the Progressive Era and the modern welfare state, the current technological evolution will likely bring about the next wave of Progressive actions. The US is more conservative than Europe, but sooner rather than later we are going to have to reckon with the fact that employment based health care doesn't cut it when the blue-collar jobs of yesteryear are eliminated by automation.

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.


Contrarian (and biased) point of view based on our current success serving Fortune 500 customers - automation in warehouses is not a near-term solution, and acquiring and retaining talent is still a huge pain.

Reuters only 12 days ago published, "Amazon dismisses idea automation will eliminate all its warehouse jobs soon." [1]

Automation will continue to increase, but we're still seeing numerous job openings in warehousing. [2]

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.

[1] https://www.reuters.com/article/us-amazon-com-warehouse/amaz...

[2] https://www.bls.gov/news.release/jolts.nr0.htm


About time!

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.


It was YYZ4. YYZ3 doesn't do tours.

If you're interested in helping make it better, Amazon's got an office full of developers here in Toronto[0] 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[1].

[0]120 Bremner, between the ACC and Rogers Center.

[1]This alias at company name dot com.


Off topic, but you're literally the first Amazon developer I've ever seen say anything positive about working there. Do you feel that the work culture is unusually good for your office compared to Amazon as a whole?


The hiring bar at Amazon is high enough that anyone unhappy would have little trouble getting a job elsewhere. Plus, there's an internal policy that any dev in good standing is welcome to join any team that wants them. So most managers treat their devs very well.

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.


Any article that you suggest for knowing more about JD's fulfillment tech?


The technology seems similar: https://www.youtube.com/watch?v=HYjc9h8oSsY (This is actually Alibaba but JD should be the similar)


It's awesome that the boxing machine is not patented by Amazon but is owned by an independent (? for now?) company. It means that others can access this tech and that Amazon won't close it before competitors. If it works, I believe Amazon won't wait for long with the takeover though.

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 would like to remind you such machines are not built because it requires some genius touch in design, they are not built because there is little demand for them. But yeah, tool being not locked under some patent by some monopoly or the other is a good thing.


Do you have any pointers for articles stating that this is the case? I'd love to read more about such solutions - I thought that machines stealing our real jobs is still a matter of tech not being good enough.


"That would amount to more than 1,300 cuts across 55 U.S. fulfillment centers for standard-sized inventory." ... "removing at least 24 roles at each one, these people said. These facilities typically employ more than 2,000 people."

I'm surprised that number is so low. There's only a couple dozen people out of 2000+ that actually put orders in boxes?


The machines aren't going to operate perfectly at first. This is likely a trial run on very specific types of goods (dimensions, weights, etc).

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.


It sounds like a gradual transition to me, not replacing all packers overnight.

> 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.

and

> The machines have the potential to automate far more than 24 jobs per facility, one of the sources said.


'standard-sized inventory' is probably doing a lot of heavy lifting. If this the same machine several of the videos there's no bubble wrapping by the machine and no option to use dunnage to put several wildly different sized objects in one box, or deal with like a lamp or a toilet plunger. There's still gonna be manual packing it's just now there will by SIOC, automatic-boxed and manual-boxed categories when presorting for shipping instead of just SIOC and boxed.


I guess 24 roles could also mean 24 roles like lead packer, junior packer etc ...


It's no stretch to assume by "role" they might mean a job title, and that each role could possibly be being filled by 3 people (3 * 8 hour shifts for a 24 hours continuous operation), so now you're looking at 72 jobs. I've no idea if Amazon does run 24 hour operations though.


Looks like they mean 24 people at each center, as 24 * 55 = 1320.


I know almost nothing about how these places work, but I would've thought there would be way more people putting stuff in boxes to ship.


Is there a field of study associated with job creation?

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.


Automation means some people are richer (mostly the machine sellers and end-consumers), with more to trade for the products of some other kind of job.

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.


I think hand-made "artisan" products is going to continue to be a large part of the future of "Jobs". Automation can mass produce high quality items that aren't unqiue, and 3D printers can make unique products that aren't high quality... but humans right now are the best at making high quality unique items.

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.


Certain segments of the agriculture industry seem promising, since smaller scale operations could benefit from efficiencies in logistics and distribution that were traditionally drivers of consolidation.


Entrepreneurship.


Looks like it's not just Amazon "JD.com Inc and Shutterfly Inc have used the machines as well, the companies said, as has Walmart Inc. Walmart started 3.5 years ago and has since installed the machines in several U.S. locations"

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"


I was debugging firmware for a Dutch supermarket chain as a summer job while in high school in the 80s and at that time they were ‘very close’ to replace all packing jobs with robots. Did Amazon pull it off finally?


I'm all for this as soon as we change the laws so companies pay payroll taxes on automation commensurate to the number of paid positions taken by the equipment.

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.)


This would kill the entire economy almost immediately.

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.


> How many "paid positions" does Microsoft Excel replace every day?

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:

https://www.bestplaces.net/economy/county/kentucky/harlan

https://www.kentucky.com/news/local/coal/article44121909.htm...

> You can't "strive for" automation, while at the same time essentially locking in labor costs by taxing that automation as human labor.

Why not?


Reductio ad absurdum is not a fallacy, it's a perfectly valid logical argument (that applies in this case).


Email, MS Office and cell phones have put almost all secretaries out of their jobs. Automated billing system (in a telco) has put my aunt out of her job.

Office jobs are disappearing by the millions, just like physical ones. Lucikly, at least until now, new ones were created in their place.


Tax on labour (or housing) is quite weird if you look at it with fresh eyes. Don't we want to support people working and living in a house? Then why tax those? Is it a relic from the past?

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...


"Taxing housing" isn't weird at all. We don't tax housing. If you rent, you pay rent to your landlord and that's it. You don't also pay taxes on that rent to the government. I'm guessing what you actually mean is that we tax property. This seems inherently fair to me. Why should you get to squat on a piece of land for free in perpetuity, especially when for most land, the value of the land is inherently improved by government-provided services?


Semi-related ... I feel like we should tax rent-seeking way more than we do. Selling a piece of property/house is one thing, but to extend your metaphor, why should one get to seek rent on a piece of property in perpetuity (if the tenant stays long term)?

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.


Most paces have some tax for living there, usually related to the worth of the house. Weather that's collected at the owner or at the rent-payer, that's not the point as "all taxes are transferred to the consumer" anyway.


That seems tough to determine. Does an autonomous bulldozer count as one employee because you would have had one person driving it? Or does it count as five employees because you run it 24/7? Or does it count as five hundred employees because that’s how many people you’d need if they were using shovels?


Similarly, this also leaves open a big loophole - if you implement a tax on jobs eliminated, the incentive shifts to starting companies that that don't have jobs in the first place. Then you quickly arrive at the sub-contracted functions shell game where a company pays for a service rather than for specific workers to avoid the automation tax, and the fulfillment companies shuffle around as new automation tech allows them to eliminate more jobs.

I don't think taxation on automated labor really gets us anywhere. Wealth tax makes a lot more sense with fewer loopholes.


Shovels? How many people would it take to do the job without shovels is the question you should be asking yourself.


I find this notion ridiculous. There are many industries like farming and mining which are now highly mechanised and employ a tiny fraction of what they did. As a result, food and minerals are cheap, fewer workers have to put their lives at risk doing dangerous jobs, and the labour force for other endeavours is increased.

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.


Considering how badly Amazon treats their employees, I'm surprised they haven't allocated vast amount of resources already to automating everything in their warehouses. The number of times they've appeared in the news in the past regarding their treatment of employees, you'd think they despise humans and want to replace everyone with robots and that they would have done it already.


Well, if they could they would. They're keeping the workers they absolutely need and automating everything else possible. The endgame is a dark warehouse (no lights, the robots just sort of know what to do) but the workers still need some money so they keep working even though their numbers keep dropping.

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?


> This box robot (…) doesn't ever call in sick.

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.


>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.


I should go back and read some older books on the wondrous future of automation. The idea always was, of course, that machines would do the boring part of a job. What people actually bring to the picnic is a subset of a job that is hard to do with a computer and/or machine manipulation, so you end up with deskilling of supermarket cashiers or bank teller gigs.

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.


"...builds boxes to size..."

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.


So the 'pick' part of 'pick and pack' still needs humans. This machine glues a cardboard box together removing the need for the guy with the tape. The article stresses that this is about efficiency rather than reducing head-count. This is a good thing as that means more optimally packed items being ferried around the country. However, for other online sellers that do not have this semi-automated packing this is not so good news, they don't have the scale Amazon does to have the robot support engineer on site. They have to pack the old ways, with lots of labour.


I think it’s safe to assume that this machine will be getting improvements. The “pick” part will surely also get automated at some point. It’s quite obvious to anybody who is willing to think about it for a while, to deduct that these kind of jobs will eventually all get automated. What the various governments should be doing now, is to start preparing the society to a new reality in which no-skill and low-skill jobs are gone. This will happen at some point. What do we do about it? The only thing that is happening so far are various timid Universal Income pilots. But those often hit substantial political pressure and are being labeled as socialist or even communist, and therefore unacceptable. So how do we prepare?


Picking things is surprisingly hard. It will by automated at some point, but it may be one of the last jobs to disappear from warehouses.


The biggest issue is not America, it is the third world.

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.


Fantastic, less human labor in one sector means new opportunities in another.


This is a platitude that is said often, but there is no evidence to support it. There is no law of the universe that says whenever a job gets automated, a new type of work pops out of the aether to employ displaced people. It is wishful thinking at best.


Well, if you look at history it keeps happening. Imagine if you said in 1900 that almost all farm jobs would be eliminated in 100 years. The future would be almost unimaginable, but we've turned out okay.

However I think the problem is we're at a point where things are churning so rapidly that many are being left behind. Your 50-year-old truck driver is not in a position to learn a new skill every 5-10 years to stay afloat, or uproot his life and move to a new city where they need workers.

Furthermore, there's conceivably a point when we won't need very many unskilled workers at all -- when we finally run out of things for average people to do -- and humans don't tend to treat other humans very well when they aren't needed. However, that's further off than the churn economy we have right now.


It doesn't "keep happening", it's happened exactly twice, in the two industrial revolutions, in each case the new type of work that needed millions of human hands was obvious. There is no guarantee that it will happen again, especially given that new industrial inventions require no human intervention at all.

To fall back on an example that is not my own, imagine a conversation in 1900 between two horses:

Horse 1: I'm worried about these new automobile things, they might take our jobs

Horse 2: Don't worry, every technological development that displaced horse work in the past has led to more newer and better jobs for horses, there will always be work for us. Just look at trains, we don't have to pull big loads anymore, now we just draw carts and support riders, who knows what great jobs are just around the corner.

The world's horse population peaked in 1915, and is now down by 70% in the US.


In many of those past cases, the new way to do something involved new machines, which required new or expanded manufacturing to build, new or expanded supply chains to provide the materials to make the new machine, and created new jobs to operate and maintain the new machines.

Much of the new present day automation does not involve new machines. It often instead involves just interfacing a computer to an existing machine and writing software to let that computer replace the human operator of the machine. Those computers and interfaces are commodity items, largely built in automated factories.

This seems to me to be a fundamental difference.


Look at how retail has changed in the past 30 years. I usually point out video rental stores as an example of technological disruption. Rental stores and stores that offered rentals as part of their mix made up something like 45,000 locations, and maybe about 125k jobs. They were replaced by maybe 20k jobs or less, with redbox and netflix combined only accounting for 8kish jobs.

Retail changed from a career that offered full time jobs to one making use of a just in time workforce, and has suffered tremendous consolidation over that period.Like entre chains and concept stores have declined; something like Media Play is dead and no longer able to be revived. Even Gamestop suffers, as the amount of physical releases declines.

I think retail shows it can be pretty dire indeed.


Is there evidence to say the contrary?

I imagine instead of Amazon paying 11 dollars an hour, a small business has the opportunity to pay 11 dollars an hour.

Before Amazon's size and HR department sucked up local labor.


I don't need evidence to the contrary, what I said is the null hypothesis. If there's proof that every job lost to automation will necessarily mean one additional job opening in a different field, you are welcome to provide it. In absence of such evidence, I'm going to stick with the default belief: a job lost means a job lost, period.

And what is stopping that small business from starting up right now? Considering how terribly Amazon treats their workers, people should be lining up around the block to apply.

Dollars to donuts, the answer is the small business can't afford to start up because Amazon uses its monopolistic power to crush it, and that power will only grow as they continue to automate.


Where are the new opportunities here? Does the number of people employed designing and manufacturing these machines come close to the number of people put out of work by their creation?

Let's not pretend that Amazon is some benevolent force here. It's the same company that sets work targets high enough to force employees to have to pee in bottles and penalises sick and pregnant workers for taking time off or breaks. [1]

[1] https://www.businessinsider.com/amazon-warehouse-workers-hav...


Yes, now those workers won't have such horrible conditions.

Or at least we are closer.


>Yes, now those workers won't have such horrible conditions.

Instead of working in "horrible" conditions, these workers are now simply unemployed. Do you think they see that as an improvement of their position in life?


Video link of the packing machine that Amazon is using:

https://www.youtube.com/watch?v=9rP1wjEsbak


I suppose the hope is that such eliminations come gradually enough that displaced workers are able to find other accommodations without too much difficulty, maybe with a little extra up-skill that gradually increases the necessary skill level for the average job. I think the doom & gloom "automation is killing jobs" narrative presupposes a sudden shift from manual to automation, which really hasn't been the norm.


This story started me wondering whether hardware, in this case Amazon order packing automatons, or wetware has a smaller net lifetime carbon footprint, assuming both are as green as possible. (Ignoring, of course global over-population and/or mass over-consumption of essentially disposable goods.). These days I am having a hard time imagining that technology could ever have the negative carbon footprint required to mitigate climate change.


This seems to simply wrap each item in a cardboard shell, is that protective? Also whenever someone complains about receiving a tiny item in a large box, the answer is usually that Amazon does so to optimize the filling of its trucks so that all the boxes are densly packed (no room for them to fall), wouldn't having each item its own custom sized box hinder that optimization?


"The new machines, known as the CartonWrap from Italian firm CMC Srl, pack much faster than humans."

Here is the product link: https://www.cmcmachinery.com/portfolio-item/ecommerce1-cmc-c...


https://youtu.be/0eqtqKDbV1Q

Here's the YouTube link.


This is going to be a shock to all of those people who were willing to give tax incentives for Amazon to build distribution warehouses in their city in exchange for "jobs."

Though, this is only the beginning. A few thousand jobs now, the majority of them later.


How do you know overall jobs will decrease?

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.


Although having a very low unemployment rate is better than a high unemployment rate, it doesn't help paint the whole picture of the American workforce. The other metric that ought to be mentioned along with unemployment is the labor participation rate, which has been trending worse for a while now [1]. Considering the two together, the economy isn't as rosy as it looks. [2]

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.

[1] https://www.investopedia.com/ask/answers/061515/what-key-dif... [2] http://www.startribune.com/even-with-unemployment-at-50-year...


Curiously, what does the homelessness & (true, ie not unemployment) jobless rate look like in that graph? Of course though, they're tough metrics.

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.


> 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.

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.

[1] https://azizonomics.files.wordpress.com/2013/01/farmjobs.jpg


So I take it your view is we should not plan / discuss how to handle it? Take it on good faith that it'll be fine?


My view is that the loss of driving jobs can be handled normally, via existing free market and education institutions -- that it's no different from the loss of other job categories over the past 50 years.

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.


That's fair. It would be interesting to have data that compares all of the job classes out there and the number of workers out there, against the worker education.

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.


Just a quick point: they don't necessarily need to move upward. They can move laterally. They can work in package delivery, in restaurants, at Starbucks, in health care, in tourism, in solar panel and home battery installation, in environment cleanup sites, in converting homes to greater energy efficiency, and so on. Honestly a lot of the new jobs 20 years from now we have a hard time guessing what they'll even be (if you magically know, you'll make billions investing), which makes "planning" for it a bit handwavy anyways. (Remember, nobody "planned" for web developers to be an exploding new career.)


Well that goes back to the original point though. Which is, that no education jobs will theoretically dwindle. If we have machines capable of handling all, repeat all no education manual labor jobs; things like picking fruit, delivering boxes, etc. Where will those people go?

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.


Looking at the state of my apartment I could really use a fulltime maid/cook. I can't afford that right now, but if the difference between the average income for the low income group and the high income group goes up enough, that would be a possibility (assuming computer programmers continue to get paid a high income, whereas maids do not).


They seem to be trying to convert these lost employees into "delivery companies"[0]. As much as I dislike the Uberfication of the world, seems like a clever way of shifting human capital from warehouse to delivery.

[0] https://arstechnica.com/information-technology/2019/05/amazo...


Most of those tax breaks have specific empployment figures associated with them.


Bu there's no penalty if this target are not met.


Aren’t the tax breaks often tied directly to the number of jobs? I believe that’s how the incentive structure for Amazon’s new NoVA headquarters works for instance.

If the company doesn’t create the job, the tax break is reduced by a certain percentage per job.


> Aren’t the tax breaks often tied directly to the number of jobs?

Usually, no, though the public justification for the tax breaks is usually in terms of expected jobs.


This is false, there are almost always specific job number requirements that must be met to be eligible for tax credits or the credits are tightly tied to hiring. The Joliet, IL Amazon warehouse incentive deal is a typical example. Amazon can credit employee withholding taxes against corporate income tax. If there are no employees hired, then there is no withholding tax, and no tax credit.


Or the penalty is "cheap" aka worth it.


All of these situations are unique, but most of the time the benefits are awarded on a per job basis. So if the plan is to hire 2000 people, the reward will end up being divyied out depending on how many of the 2000 people are hired.


It shouldn't be a shock. There's overwhelming data supporting the idea that tax incentives for low paying jobs never end up paying for themselves.


They really need to wrap the books with a piece of butcher paper or something. I keep getting books that slid around and got scuffed to heck by contact with the boxes. Since I have been buying more pricey collection quality books, I have been sending them back :|


Isn't that what the bubble wrap is for? Although they don't fit well in the smaller standard book boxes that Amazon uses...


Well, I didn't see any bubble wrap in that video :) My recent return and its replacement had zero protection between the book and the cardboard.


I've bought two books in a reecent box and they weren't padded either.

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.


Is it odd that the headline and article imply Amazon created the tech, then halfway through it's clear they are implementing a solution many other competitors have?

Seems like marketing copy regarding Amazon, possibly being behind in implementation of a common solution.


If we taxed robots, this would be easier for more people to embrace as a good thing.


Is it me or does it feel like the door to economic advancement is closing faster and faster every day?

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.


amazon is on a downhill decline for me. their scale and automation of both humans and machine yields a disconnected service. a large percentage of books i receive from them, which is substantial, are damaged upon arrival. but what do they care? they just initiate a replacement with no questions asked, but now i have to repack the material and go to ups or an amazon locker. i am now working for them for free! my time and effort is explicitly being used as quality control, and this is undoubtedly factored into their statistics.


In related news, Amazon has announced it will give $10000 to any employee that wants to start a delivery business. Seems like a reasonable move, from packer to delivery person.


Now I will tell my small children that they won’t have a box filling job waiting for them when they graduate middle school.


Is anyone complaining about this? These are mindless, terrible, Taylorized jobs.

If machines can do it, they should.


I will start using Amazon again if they replace their dehumanized workers with robots


Bezos goads competitors to raising wages while developing machines to eliminate wages.


As mentioned in the article, Amazon's competitors such as Walmart are also doing trials with the same machine.


Workers of the world, rejoice! Amazon will now be exploiting fewer workers.


more people loosing their jobs will be unable to buy products, no matter how cheap, from amazon, hu?


I wonder if this means I’ll stop receiving boxes the size of a small house containing a single small item, or if it’ll be even worse.


That's usually due to delivery/packing requirements, not a human unable to find the smallest possible box.


Can you elaborate? I routinely get smaller boxes, but I also routinely get boxes that are comically oversized and contain an item that would easily fit in one of the smaller sizes. What requirements are making that happen?


In a lot of cases it's optimization for a truck loadout. The loading setup is usually automated to maximize the number of packages in a truck while minimizing empty space, so you can easily end up needing a larger box to fit a specific gap in the truck.

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.


These items aren’t fragile and usually come with totally inadequate padding, so that’s not it. The truck loadout thing seems likely though.


Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact

Search: