There has been so much Amazon news piling up over the holiday break that I figured it would be best to take a short break from my essays on Amazon, the tech industry, and late capitalism. So instead we'll focus on… the immediate fallout of late capitalism on the tech industry, specifically Amazon. There will just be a lot more links (and a lot fewer Hungarian philosophers).
The biggest story is that Amazon's 2023 layoffs have begun, and they're nearly twice as large as was initially projected: approximately 18,000 people will be let go from the company. 18,000 people is hard to wrap your head around, so I tried to come up with some comparable figures:
A typical freshman class at my alma mater, Michigan State University, a very large public university, has about 9000 students. So twice that number of people will be losing their jobs in this round of layoffs at Amazon.
There were also about 9000 homicides in the United States in 2022. (This was one of the higher murder rates in the last twenty years.) So more than twice as many people as were murdered in the entire US will lose a job at Amazon.
The numbers fluctuate, but between the end of 2020 and the end of 2022, there were only about 36,000 coal miners in the United States. So about half as many coal miners as still exist in the entire US will lose a job at Amazon.
Here's one non-US-centric data point: the historical city of Venice, Italy (the islands, not the mainland), is home to about 55,000 people. If everyone who lived in Venice worked at Amazon, a third of them would be moving away from the city.
This last figure is a little disingenuous, because of course so many more people work at Amazon than live in Venice. At its peak a year ago, 1.62 million people worked at Amazon. (Five years ago, it was 560,000.)
1.62 million people is about the population of the city proper of Barcelona, Spain, which conveniently has a population density of 16,000 people per square kilometer. So you can also imagine a typical square kilometer of downtown Barcelona became suddenly empty: that's what's happening to Amazon. (See also Brad Stone at Bloomberg: "Amazon Kept Inventing, and Now It’s Going to Start Cutting.")
Some Amazon employees who want to leave the company might find that easier soon. The FTC proposed a ban on noncompete contracts, specifically using its fact sheet to single out Amazon's use of the widespread practice.
At CES, Amazon focused on Alexa and the internet of things — perhaps ironically, two of the groups that will probably be hardest hit by layoffs on the hardware side.
Amazon lost its NLRB appeal challenging the Amazon Labor Union's election win in Staten Island in 2022; the company can still contest this ruling to the NLRB governing board, and the case may well end up in court before Amazon negotiates a contract with the union.
The big news in AI this week was Microsoft's reported $10 billion investment in OpenAI, valuing the startup at $29 billion. This story was eventually leaked just about everywhere, which always makes me very suspicious: "Big blue-chip company considers investing in hot new thing" is catnip to the tech press, but 1) nothing has actually happened yet, and 2) someone was very interested in making just about every term in this deal well-known to as many people as possible, for their own reasons. I suspect that has something to do with all of the OpenAI rivals who now want similar multiples to their valuations. Everyone knew the venture market was cooling off, and here comes AI to heat it all up again? I for one will wait and see what actually happens before I hop on the hype train.
I did like Tom Krazit's observation at Mostly Cloudy: "This is the 'EMBRACE' part." And for a more bullish reading on the growing interest in AI inside and outside of the big tech companies, Ben Thompson at Stratechery has a strong essay called "AI and the Big Five," with this take on Amazon:
This issue of marginal costs is, I suspect, an under-appreciated challenge in terms of developing compelling AI products. While cloud services have always had costs, the discrete nature of AI generation may make it challenging to fund the sort of iteration necessary to achieve product-market fit; I don’t think it’s an accident that ChatGPT, the biggest breakout product to-date, was both free to end users and provided by a company in OpenAI that both built its own model and has a sweetheart deal from Microsoft for compute capacity. If AWS had to sell GPUs for cheap that could spur more use in the long run.
That noted, these costs should come down over time: models will become more efficient even as chips become faster and more efficient in their own right, and there should be returns to scale for cloud services once there are sufficient products in the market maximizing utilization of their investments.
Thompson might be taking too limited a view of AWS's role in AI as a pure cloud provider of processing power: AWS also layers in plenty of services, including AI, that compete with ChatGPT and other services, and will certainly offer more. It is useful to note that the company has a lot of roles to play (and a lot of chances to take slices of the pie) in any coming AI boom.
Maybe there's a distinction to be drawn between AI as "using GPUs to analyze and create documents" and AI as… whatever the hell you would call this:
“Buyers in a traditional manner do the same thing over and over again,” Ackerman said. “They get a call, they get a sales pitch, they buy an amount of product, they usually buy the wrong amount because they’re humans, and lo and behold, people buy the products, and it’s a cycle. When you have actions that can be predicted over and over again, you don’t need people doing that. And frankly, computers, or algorithms, or machine learning, are smarter than people.”
Understanding this, Amazon’s leaders decided to attempt to automate traditional vendor manager responsibilities including forecasting, pricing and purchasing. People inside Amazon began calling this initiative Project Yoda. Instead of having vendor managers do the work, Amazon would use the Force.
Automation at scale can (and will) apply to way more things than processing text, images, and voice.
Amazon seems eager to invest in sports, with The Information reporting that Andy Jassy is considering building a stand-alone sports app to compliment Prime Video. At the same time, however, the ratings for Thursday Night Football saw a steep drop after shifting to Prime, going from 16.2 million people in 2021 (Fox, the NFL Network, and Amazon) to just 9.6 million on Amazon alone based on Nielsen's external estimates. (Amazon's own internal numbers showed a somewhat higher figure, 11.3 million viewers.)
Amazon reportedly predicted there would be a year-to-year drop — it only makes sense in a shift from broadcast to streaming, behind a paywall at that — but not that steep a drop. Business Insider reported that Amazon was scrambling to make up the shortfall to advertisers, who were somewhat comforted by the fact that Amazon's viewership is on average seven years younger than the past viewership for Thursday night football.
Live sports have traditionally turned out to be a good bet for video, but just for that reason, it's an expensive business. It may never be the case that pro football pays for itself on Prime, but it gives the service must-have status for a huge number of fans, and a degree of legitimacy for everything else the network shows. If Amazon gets serious about sports, you can imagine them investing in sports news, fantasy sports, maybe even sports betting: lucrative packages that could be particularly attractive to a global market beyond Americans' love of pro football.
Or Jassy just wants to make good on the sunk costs of having already bought an expensive football package. Either way!
Witnesses say a makeshift barrier around the deceased worker using large cardboard bins was used to block off the area on the outbound shipping dock where the incident occurred, and workers criticized the response and lack of transparency about the incident…
“No one should have been told to work alongside a dead body, particularly after witnessing it. Day shift comes in at 7am or 7.30am, and we were never informed until we arrived to where it had occurred. No warnings before walking into the building. No on-site counselor. Simply a flyer put out days later informing us of how to receive mental health counseling.”
This Daily Mail story about how Netflix's Wednesday could end up at Amazon is almost definitely oversold — the article itself walks it back in an appended paragraph — but it is a little strange that Amazon, which owns MGM, which owns The Addams Family and its characters, didn't wind up with the streaming rights to a new show based on those characters.
The unusual Wednesday situation is also a stark reminder that Amazon, through MGM, is now in business with many of its rivals.
MGM Television produces The Handmaid’s Tale for Hulu, Last Light for NBCU’s Peacock, Fargo for FX as well as Vikings: Valhalla, which is also for Netflix, and development projects for networks and streamers including Starz and Paramount+. On the unscripted side, it makes The Voice for NBC, Shark Tank for ABC and Survivor for CBS as well as a swathe of reality series for cable networks.
Maybe that's good for Amazon. For a long time, what distinguished Amazon from many of its tech peers is that it didn't much care how you consumed its products as long as, at some point, you bought them. The long saga of Prime Video slowly making its way to Apple TV is an outlier in this respect, although for a while it seemed closer to the way things were trending.
After all, it's not always easy or profitable to become an omnibus portal for all of your own content, as HBO Max has discovered (and was recently lampooned by The Boys).
I should be back with more and better next week. Take care until then.