Amazon is on the cusp of using more robots than humans in its warehouses | Hacker News Reader

Amazon is on the cusp of using more robots than humans in its warehouses

wsj.com

65 pointsjbredeche63 comments

Gift link: https://www.wsj.com/tech/amazon-warehouse-robots-automation-...

Amazon is also on the cusp of becoming the largest company by revenue.

The most interesting stat to me is:

> The number of packages that Amazon ships itself per employee each year has also steadily increased since at least 2015 to about 3,870 from about 175...

That's impressive growth, to be sure.

But shipping about 0.5 packages per employee per day is not impressive to start...

It's no wonder Amazon didn't make money for a very long time.

It says "ships themselves". I interpreted that as meaning through their own delivery service versus UPS/USPS, and I thought those third party shipments were still picked up at Amazon warehouses. Meaning this could have much more to do with Amazon building out their delivery service than with warehouse efficiency.

I'm also under the impression that most of the Amazon delivery drivers are contractors, and wouldn't count in "per employee" metrics.

If Jassy had his way, Amazon's sole employee would be him and he'd be rent seeking everything down the line with robots and AI.

He has no vision beyond what is best for him personally. Other people aren't humans.

Have you worked with him?

I had the privilege of working with him for a little over a year.

He was respectful, analytical, and only thought about what is best for Amazon and Amazon's customers.

I don't think his own aggrandizement or what's best for him enters into the calculus of his day-to-day work one itoa.

He's an untiring as well. I realized I had gone far beyond my peter principle when I was working with him.

I have worked with him.

He's a robber baron, not a captain of industry.

Don't have to look much past his "gobble gobble" email to see that, nor his announcement of RTO the day after an all hands where remote work was touted as the future.

> He's a robber baron, not a captain of industry.

Arguably these are synonyms :P

"his announcement of RTO"

At least he walks the walk. He was in the office 100% of the time.

I don't give assholes credit for being unapologetic assholes

This mindset is more common than not among executives.

I think there's a non-zero chance of 95% unemployment within the next 20 years.

> I think there's a non-zero chance of 95% unemployment within the next 20 years.

I can't imagine this would last long with the social instability that would follow.

That also seems like a gross overestimate for many industries. Many supply chains are nowhere near able to be automated to the extent this ratio would imply.

20 years is far too long a time horizon to make a sensible prediction, to even say if the other post is an overestimate or an underestimate by either time or percentage.

20 years ago, the state of the art in AI was so bad that it predates reCAPTCHA being a functional business model for how to get labelled training data for an OCR to read scans of books. In cars, the DARPA Grand Challenge (2004) had no cars able to complete the route, and the winner only got 11.78 km (of 240 km) before getting stuck, and the subsequent Grand Challenge had yet to happen because that was October 8, 2005.

I can't sensibly forecast past 2032, because that's when various things currently under exponential growth reach absolute levels that I think suggest the exponential assumption may stop working. But also, I don't expect wild AI-induced economic disruption* before then, because one of those exponential growths is renewable energy getting to near-parity with current global electricity supply by about then, and the fact that the current global electricity supply is "only" 250 watts/capita, which means** there's just not enough electricity to run enough computers to disrupt that many jobs.

But if the question is "when will the AI be good enough to do ${task}?", rather than "how fast can we deploy the AI when this happens?", I have seen both (1) people claim we're decades away from AI doing things mere months before AI did those things, and also (2) Musk repeatedly saying FSD is 0.5-2 years away every year for the last decade, and when he finally soft-launched his robotaxi it was 6.5 years behind Waymo's.

* by which I mean international >20% unemployment that's not just due to a war or a global financial crisis or a pandemic

** absent algorithmic breakthroughs; I don't expect enough of those to matter, but I do expect some. I am also accounting for computers getting more energy efficient, because that's still happening and we're a long way from the theoretical limits.

> non-zero chance of 95% unemployment within the next 20 years

With that said, there is probably also a non-zero chance that something like that wouldn't be negative but something positive instead. I suppose that'd be the ideal scenario.

U.S. productivity has increased massively over the last half century and yet much of the increased productivity has ended up in the hands of a select few.

This is being accelerated by the bill passed the senate literally today.

Is there any reason to believe we will drastically reverse the trend?

>much of the increased productivity has ended up in the hands of a select few.

I do agree that wealth is not distributed particularly fairly but I don't understand the idea that it's mostly going to a select few. There are almost 20 million more jobs in the US now than there were 20 years ago, that's a lot of productivity right there.

Amazon employs 1.4 million people. Bezos has spent maybe the equivalent to a couple of weeks worth of Amazon payroll so far in his life. His yacht was the equivalent of a few days, his wedding maybe one hour. That money didn't come from Amazon either, it came from people wanting to own Amazon.

Leaving aside whether that's fair or not, surely it means that most of the productivity and money is going to the employees?

What am I missing here? How was this better before, what has changed recently that makes this such a big talking point now?

What you’re missing, in your relentless focus on mostly abstract economic measurements, is that the measures of individual success and happiness are going down. People on average, are less able to afford to buy homes, raise children and pay for health care.

I assure you I have no relentless focus on anything but trying to understand what's going on! There are many problems for sure but I don't understand how they can be solved with faulty explanations!

Many people seem convinced that the problems in society are caused by billionaires hoarding money, when they clearly aren't hoarding any relevant amounts of money. The problems with housing, healthcare and such are very obviously caused by other things (Home mortgages being a huge cornerstone of the entire economy for one, millions are placed in a situation where cheaper housing would be a catastrophe - including banks).

Maybe our parents, who were able to buy a house for like 5 years of a worker's salary, were just the luckiest generation? It sure hasn't been that easy to get a home before or after at any time in history.

Right now is one of the runner ups though, but we're comparing the situation with the best ever.

> There are almost 20 million more jobs in the US now than there were 20 years ago

There are 40 million more people in the USA than 20 years ago.

>> There are almost 20 million more jobs in the US now than there were 20 years ago >There are 40 million more people in the USA than 20 years ago.

Makes sense, 20 million more workers should mean about 40 million more people in total, accounting for children and elderly.

That would be bad for anyone selling stuff

“Executive” is another word for “sociopath” so this is unsurprising.

Is amazon a net benefit to humanity? Is extreme disintermediation in general a beneficial thing, where there is only the consumers left, paying rent-like tithes into one big pocket, without an economic ecosystem in between?

Looking just at their physical-item distribution model, I'd imagine it's far more carbon-efficient than physically shopping.

Case 1: I drive to the store to get a widget. There is a low chance Store-1 doesn't have it, and I have to keep shopping. Regardless, I am driving about 30 minutes total, conservatively.

Case 2: I order a widget from Amazon. The Amazon driver organizes their stops to minimize travel time, and it's undoubtedly less than 30 minutes driving from the previous stop.

Both cases require transport of goods from factory to the final storage location (shopfront or Amazon warehouse), so the difference there is negligible.

I broadly agree, but the assumption that people drive for 30 minutes to on average to buy an item sounds like a stretch. When I don't buy online I simply walk or ride a bike to the store. On the other hand, I live in an urban setting.

From what I've seen of American cities, I can easily believe there are many in those cities who have a 30 minute drive to a store. And while it's only one specific category of product, I keep hearing about "food deserts" in the USA, where poorer communities are only served by dollar stores that don't have a good range of stuff, so I can easily believe that there's worse than what I've seen in all product categories.

I'm in Berlin, so I'm a 25 minutes walk from three supermarkets a bike shop and a pharmacy, despite being in one of the remote corners of the city. Our previous place was much more central so it had a dozen or so supermarkets, a building supply store, and a shopping mall in that distance. Both new and old place have excellent public transport, I have literally seen someone transport a kitchen sink on one of the trams.

Yeah, that tracks. Even while living in the far end of the Toronto urban fabric we have something like nine grocery stores within 25 minutes of walking distance. And we are not living in an urbanist's wet dream, either.

The sort of extreme car-centric urbanism often seen in the US is the exception around the world, not the norm.

Because most people ride bikes and walk to the store despite all of the huge parking lots filled with cars.

Most people in most of the cities around the world get around by walking, transit or cycling. Car-centric development is a pretty recent and relatively localized phenomenon.

As is Amazon

92% of the people in the US have access to at least one vehicle

https://www.fool.com/money/research/car-ownership-statistics...

And this is worldwide

https://www.visualcapitalist.com/vehicles-per-capita-by-coun...

The average driver in the US drives 40 miles per day

https://www.kbb.com/car-advice/average-miles-driven-per-year...

No most people don’t walk to the store

None of the statistics you presented support your last statement. Perhaps try traveling abroad some day.

You would be surprised how much traveling I’ve done…

I literally just got back from a trip to London and Paris a couple of weeks ago.

https://news.ycombinator.com/item?id=44412439

Let’s just say I’m Platinum Medallion on Delta (also from racking up miles on partner airlines Virgin Atlantic and Air France) , for 3 years in a row…

Are you trying to really say that cars aren’t popular outside of the US?

Did I miss the statistic you cited where most people in first world countries walk and ride bikes every where?

I was just in central Madrid and was taken aback by how few cars there were. I think most roads were bus/taxi/delivery only.

I’m from LA so every European city seems wild with tons of people walking (and biking in places like Amsterdam and Barcelona).

That said, I live in West Hollywood so I actually walk to the grocery stores, restaurants, services, gyms etc except when it’s too hot or too rainy (not that it rains much)

Just a quick Google search shows that Spain has 643 cars per 1000 people and the US has 842.

Yes I see many more people riding bikes and walking in London where I just left in June than DC when we went in May. But going to the big urban areas doesn’t tell you about the rest of the country. For instance in the UK it’s 603 cars per 1000 people.

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

And yeah I chose a place where I can go to the gym, a convenience store, and a bar without ever walking outside. But I’m not going to say that’s common like the person I replied to seems to think.

Maybe, but when I order from Amazon it's usually an item or three. When I go to the grocery I'm getting dozens of items.

It would be interesting to quantify the carbon emissions from those two systems (they're not fully independent variable) I wouldn't be surprised if my intuition is wrong. Or yours

Of course, the productivity gains are largely drained by the shareholders, so we only collectively get part of expected benefits of a centrally-planned economy.

Centrally-planned economy was a disaster.

Capitalism might not be perfect, but communism did not function at all. You can have the best of both worlds in socio-capitalist system.

Let Amazon automate everything. Does anyone here really want to do the manual labour in their warehouses?

Amazon is a centrally-planned economy. You do not need to fix prices to achieve this.

Yes, he is a capitalist. This is the logical result of capitalism.

Counting robots is like counting circuits in a piece of electronics: It's Meaningless.

However what would be interesting is if they ever have a situation where the number of humanoid robot employees outnumbers human employees. Once someone is able to run an entire large corporation with no humans at all it will be a big milestone I guess.

But there's no real driving force for this to happen, because the humanoid form factor is not necessarily ideal for most kinds of industrial applications. And a lot of the "Automation" will be AI Agents, which have no form factor at all, being purely knowledge based.

It is an interesting contrast to "lights out manufacturing" --- there are a lot of CNC companies where day shift runs short jobs, and then loads up pallets of materials and stages empty pallets for finished parts, loads the programs and sets quantities and so forth, then presses "Start", and turns out the lights to leave --- there might be a skeleton crew to deal with any issues which arise, often not.

What a meaningless metric. Number of discrete robots is not a measurement of productivity, labor, task throughput or anything like that.

Also the regular yada yada that non-robot automation offset people, non-automated tools also reduce number of people needed.

There are hundreds of computers in every car. We all know that this number has little correlation with human replacement.

Robot is a term with such a wide range of capabilities that a simple count is meaningless. Just like a tiny 4 bit micro that monitors a single sensor isn't comparable to a computer running a large LLM.

It's not supposed to be scientifically meaningful it's a reprinted press release that is supposed to spin a narrative that makes Amazon workers feel insecure about their jobs and investors feel confident that those wages will shortly be converted into dividends.

You can tell what the Wall Street Journal's true feelings are about automation by reading an article about retirement and dividing the number of times they use the term "AI" by the number of times they use the phrase "demographic/retirement crisis" e.g. in https://www.wsj.com/finance/investing/blackrock-larry-fink-a...

"Robots don't take bathroom breaks, so you shouldn't either. If you're lucky, we'll let you work for a little while more until we can replace you with something cheaper. Now, back to work!"

I wonder if there are more shelves than humans?

Depending how you define 'robot', it's sorta surprising this hasn't already been true for years.

I’m sure that when more than half of us are unemployed that we’ll just sit quietly and know our place.

What’s that you say? People with nothing left to lose tend to put the heads of Capitalists on pikes for all to see? Well maybe you should have thought of that before pillaging the commons and treating people like slaves? But “fiduciary duty to shareholder value” you say? You made that shit up in the 1950s and have been pushing it like it were true ever since. Saying a thing doesn’t make it true.

How's that UBI coming? Oh, totally off the table you say? Well, the people will surely find something better to do than rising up.

The cotton and wool mills may not have been great for the Luddites, but everyone else benefitted. Using terms someone else used in a sibling thread, they were a "net benefit for humanity".

I don't think the parent is making the argument that this level of automation can't be a net benefit for humanity, just that if it results in nearly the entire middle class and down being laid off at the same time with no recompense, that's an almost perfect storm for revolution, violent or not.

As long as people working in a warehouse identify with the middle class, inequality will continue. There is very little class conscience in the US, from what I can see. They sooner identify as "poor" than as "working class".

The middle class is on the chopping block too. The lower class will honestly probably last longer than the median of the middle class, like paralegals, etc.

Standing in the same bread line creates an awful lot of solidarity.

Social carers for the elderly might be the last people employed.

Thinking "there will always be something else for everybody to do" because "there's always been something else for everybody to do" is like the dinosaurs thinking "we've always been fine after a meteor strike".

> net benefit for humanity

Some people might argue that wealth concentrated in the top 1% is a net benefit if you look at it as one big pool of resources. But will the remaining 99% actually see any of that benefit? Or will the 1% simply tighten their grip, keeping the rest dependent on their “generosity”?

The people that make that argument are making it in bad faith. They should be flogged.

No worries! It's gonna trickle down to us and we'll be happy owning nothing /s

that depends on how important it is to people to be able to continuously overbuy cheap, poorly-made, low-quality disposable textile goods.

"Rising tides lift all boats" — one of the problems with the mills, was that the people who got richer from them didn't help keep the impoverished Luddites from drowning. This is why Marx et al invented Communism. Can we, perhaps, avoid doing it all so badly this time?

Oh don't worry. Americans at least will be kept so busy in the mines (for which they yearn) and fighting wars that they won't have the energy to rebel.

Reply on news.ycombinator.com