Amazon's Secret Weapon in Chip Design Is Amazon(spectrum.ieee.org)
spectrum.ieee.org
Amazon's Secret Weapon in Chip Design Is Amazon
https://spectrum.ieee.org/amazon-ai
84 comments
All Cloud Providers are working on vertically integrating compute - for example, look at GCP's bet on TPUs and Azure's investment into Cobalt and Maia.
The difference is, Amazon's bet on commodified ML Compute infra largely paid off due to a mix of developer advocacy and the fact that there is a large existing market of users somewhat adept with AWS.
In fact, it could be a case study of how an incumbent can lose the ball - back in the 2014-18 period Tensorflow was THE framework, and Google absolutely could have used it as a killer app to market then new GCP (and they did try), but Amazon was able to outcompete GCP on both Containerization and Cloud ML Compute because of their strong developer advocacy and training programs.
Don't be dismissive about Amazon's historically strong developer advocacy motion. Peak Microsoft, Intel, Cisco, VMWare etc all placed similar bets, and Nvidia has done something similar since the mid-2010s in the ML space. At the end of the day, buyers are somewhat technical.
GTM strategy is just as important as technical and product strategy.
The difference is, Amazon's bet on commodified ML Compute infra largely paid off due to a mix of developer advocacy and the fact that there is a large existing market of users somewhat adept with AWS.
In fact, it could be a case study of how an incumbent can lose the ball - back in the 2014-18 period Tensorflow was THE framework, and Google absolutely could have used it as a killer app to market then new GCP (and they did try), but Amazon was able to outcompete GCP on both Containerization and Cloud ML Compute because of their strong developer advocacy and training programs.
Don't be dismissive about Amazon's historically strong developer advocacy motion. Peak Microsoft, Intel, Cisco, VMWare etc all placed similar bets, and Nvidia has done something similar since the mid-2010s in the ML space. At the end of the day, buyers are somewhat technical.
GTM strategy is just as important as technical and product strategy.
Apple's approach is a prime example - why depend on Intel when they're lagging behind?
By designing their own chips and partnering with a foundry to manufacture them, Apple can create customized solutions that meet their product's specific requirements, distinguishing themselves from other PC manufacturers that use Intel/AMD chips.
Amazon recognizing it as an opportunity to stand out from the competition in the cloud services market by offering its own chips.
By designing their own chips and partnering with a foundry to manufacture them, Apple can create customized solutions that meet their product's specific requirements, distinguishing themselves from other PC manufacturers that use Intel/AMD chips.
Amazon recognizing it as an opportunity to stand out from the competition in the cloud services market by offering its own chips.
> why depend on Intel when they're lagging behind?
Tbf, Apple transitioned to Samsung by the early 2010s for their SoCs. The MacBook on Apple Silicon was a recent transition after the kinks in the iPad Pro (which is laptop specced) were ironed out.
> Amazon recognizing it as an opportunity to stand out from the competition in the cloud services market by offering its own chips
Exactly, and evangelizing earlier than other cloud providers.
Tbf, Apple transitioned to Samsung by the early 2010s for their SoCs. The MacBook on Apple Silicon was a recent transition after the kinks in the iPad Pro (which is laptop specced) were ironed out.
> Amazon recognizing it as an opportunity to stand out from the competition in the cloud services market by offering its own chips
Exactly, and evangelizing earlier than other cloud providers.
I am at a complete loss as to how you get from "Of course Amazon's chip advantage is their own chip consumption and scale in AWS" to "being dismissive of Amazon's developer advocacy".
If they're doing better than the others, good for them. It still blindingly obvious that having the biggest cloud is a huge advantage for chip design and to successfully exploit savings because of that scale, not some sort of super amazing secret just now being revealed by IEEE.
If they're doing better than the others, good for them. It still blindingly obvious that having the biggest cloud is a huge advantage for chip design and to successfully exploit savings because of that scale, not some sort of super amazing secret just now being revealed by IEEE.
Both Google and MS had advantages that Amazon did not have in the mid-2010s in the ML space.
Google had the advantage of owning the entire ML and Infra stack (TensorFlow, K8s, BERT, CNCF) and Microsoft had an inbuilt advantage in research communities thanks to MS Research's outsized impact in fundamental ML research.
At the time, the Annapurna Labs acquisition was seen as a massive coin-toss because IBM went down a similar path a decade before and failed.
Google had the advantage of owning the entire ML and Infra stack (TensorFlow, K8s, BERT, CNCF) and Microsoft had an inbuilt advantage in research communities thanks to MS Research's outsized impact in fundamental ML research.
At the time, the Annapurna Labs acquisition was seen as a massive coin-toss because IBM went down a similar path a decade before and failed.
"Microsoft had an inbuilt advantage in research communities thanks to MS Research's outsized impact in fundamental ML research"
I thought for a few minutes and I could not come up with an example of an ML technology that originated at MS Research and then spread outside MSFT. Care to give some examples? Thanks!
I thought for a few minutes and I could not come up with an example of an ML technology that originated at MS Research and then spread outside MSFT. Care to give some examples? Thanks!
> Care to give some examples
In the 2010s they were the leader in NLP and the precursor of LLMs like GPT3/3.5/4/4o
Machine Translation with Human Parity (2018) - https://arxiv.org/abs/1803.05567
MT-DNN (2019) - https://arxiv.org/abs/1901.11504
MASS (2019) - https://arxiv.org/abs/1905.02450
VALL-E (2023) - https://arxiv.org/abs/2301.02111
VALL-E 2 (2024) - https://arxiv.org/abs/2406.05370
While OpenAI was the first to monetize an LLM at scale via ChatGPT, it's still the early stages of this field, and there is a lot of innovation that can still be leveraged, especially in non-English language modeling, machine translation, text-to-speech, etc.
It's in this segment that Microsoft Research shines moreso than even Google Research let alone other organizations because of their strong NLP background in Chinese (Microsoft Research Asia), South Asian languages (Microsoft Research India), Arabic (Microsoft Research's older work during the Iraq War), etc.
In the 2010s they were the leader in NLP and the precursor of LLMs like GPT3/3.5/4/4o
Machine Translation with Human Parity (2018) - https://arxiv.org/abs/1803.05567
MT-DNN (2019) - https://arxiv.org/abs/1901.11504
MASS (2019) - https://arxiv.org/abs/1905.02450
VALL-E (2023) - https://arxiv.org/abs/2301.02111
VALL-E 2 (2024) - https://arxiv.org/abs/2406.05370
While OpenAI was the first to monetize an LLM at scale via ChatGPT, it's still the early stages of this field, and there is a lot of innovation that can still be leveraged, especially in non-English language modeling, machine translation, text-to-speech, etc.
It's in this segment that Microsoft Research shines moreso than even Google Research let alone other organizations because of their strong NLP background in Chinese (Microsoft Research Asia), South Asian languages (Microsoft Research India), Arabic (Microsoft Research's older work during the Iraq War), etc.
Annapurna had some pretty ok consumer chips before their buy out
Ditto for PA Semi, who were bought by Apple.
SiFive changed to being an IP provider, is no longer selling chips.
It feels sometimes like the PC world is in collapse. No one wants to sell chips any more, it's all for cloud or for appliance-ized phone-lile systems. When someone does come along and starts making new chips, they get bought.
I wish the hyperscalers had some interest in keeping a competitive market alive, in supporting independent competition. I respect the desire to have a strong in house team and Graviton 3+ have really been excellent (Nitro is also super interesting deep tech I wish we saw in the world; alas AMD's SeaMicro acquisition took one similar out in the wild super-fabric offering off the table rather than promoted it). But man, the consolidation in chip making has been brutal & it doesn't feel like there's enough folks getting started making cores to keep things healthy. There's some disruption from below with RISC-V starting, but it's been slow & is very down market still (Tenstorrent being the notable exception). The broad ecosystem feels like it needs help, needs new vitality, is ossofying, in large part because of these acquisitions.
Ditto for PA Semi, who were bought by Apple.
SiFive changed to being an IP provider, is no longer selling chips.
It feels sometimes like the PC world is in collapse. No one wants to sell chips any more, it's all for cloud or for appliance-ized phone-lile systems. When someone does come along and starts making new chips, they get bought.
I wish the hyperscalers had some interest in keeping a competitive market alive, in supporting independent competition. I respect the desire to have a strong in house team and Graviton 3+ have really been excellent (Nitro is also super interesting deep tech I wish we saw in the world; alas AMD's SeaMicro acquisition took one similar out in the wild super-fabric offering off the table rather than promoted it). But man, the consolidation in chip making has been brutal & it doesn't feel like there's enough folks getting started making cores to keep things healthy. There's some disruption from below with RISC-V starting, but it's been slow & is very down market still (Tenstorrent being the notable exception). The broad ecosystem feels like it needs help, needs new vitality, is ossofying, in large part because of these acquisitions.
> It feels sometimes like the PC world is in collapse. No one wants to sell chips any more, it's all for cloud or for appliance-ized phone-lile systems. When someone does come along and starts making new chips, they get bought
Alternatively, smartphones are the PC (personal computer) of choice.
SoC and mobile chips are themselves near desktop level performance. Heck, Apple's A18 can outcompete the M1 in certain benchmarks, and outcompetes a Kaby Lake Intel i5 (2017-19 period) in most aspects.
We are reaching a point where commodity mobile processors have mid-2010s desktop chip level performance but at a fraction of the cost, which opens up plenty of opportunities.
Alternatively, smartphones are the PC (personal computer) of choice.
SoC and mobile chips are themselves near desktop level performance. Heck, Apple's A18 can outcompete the M1 in certain benchmarks, and outcompetes a Kaby Lake Intel i5 (2017-19 period) in most aspects.
We are reaching a point where commodity mobile processors have mid-2010s desktop chip level performance but at a fraction of the cost, which opens up plenty of opportunities.
I agree there's a lot interesting and great about phone chips.
And there has been some progress getting some phones running real Linuxes, with upstream kernels & more regular userspaces. There's some cheating too, using hybrid Android drivers bent to be Linux-y.
But man it is so irritating to me that there's such tight controls on these chips in every way. So many seem to only be available for large devices makers. There seems to be very narrow segmentation. Ideally one would hope a small SBC with say a Snapdragon 7s Gen 3 (SM7675-A), with a single X4 core, would be so interesting as a low cost small board. But instead of the most popular chips on the planet - cellphone chips - being everywhere, there's a whole market of extra special extra old-core embedded chips - the Allwinner and Rockchips and Broadcoms - providing an alternate. One would hope removing display, touchscreen, battery, and case could lower costs, but it's just not done. https://www.anandtech.com/show/21316/qualcomm-intros-snapdra...
And phones, man, so many are locked locked locked down. Many of the Qualcomm Snapdragon X Elite laptops are locked down - laptops - which can by design install no OS but windows. Highly restricted use seems normal in the ARM world, makes these so much less able to be enjoyed.
The other major downside to phones is they have enormously limited io. A single USB 3 port isn't the worst, but it's still a very narrow straw for something like a modern flash drive to squeeze it's data through. Hopefully, again, USB4 improves this, but ideally I'd love some PCIe connectivity and especially multi-port designs (Lenovo has a couple gaming phones with >1 port in some markets, way cool).
I agree with the excitement for mobile. But it's also a dark segment, a pinnacle of consumerdom & regression to the mean versus that brief great amazing age of Personal Computer compatible, which has spawned systems both small and massive and mighty, that we have been able to extend & use however we might imagine. Phones returned us to an age of control, where our species is impotent at using the tech we have all around us, phones are infernal devices trapping us, are the spiritual foe of human spirit.
And there has been some progress getting some phones running real Linuxes, with upstream kernels & more regular userspaces. There's some cheating too, using hybrid Android drivers bent to be Linux-y.
But man it is so irritating to me that there's such tight controls on these chips in every way. So many seem to only be available for large devices makers. There seems to be very narrow segmentation. Ideally one would hope a small SBC with say a Snapdragon 7s Gen 3 (SM7675-A), with a single X4 core, would be so interesting as a low cost small board. But instead of the most popular chips on the planet - cellphone chips - being everywhere, there's a whole market of extra special extra old-core embedded chips - the Allwinner and Rockchips and Broadcoms - providing an alternate. One would hope removing display, touchscreen, battery, and case could lower costs, but it's just not done. https://www.anandtech.com/show/21316/qualcomm-intros-snapdra...
And phones, man, so many are locked locked locked down. Many of the Qualcomm Snapdragon X Elite laptops are locked down - laptops - which can by design install no OS but windows. Highly restricted use seems normal in the ARM world, makes these so much less able to be enjoyed.
The other major downside to phones is they have enormously limited io. A single USB 3 port isn't the worst, but it's still a very narrow straw for something like a modern flash drive to squeeze it's data through. Hopefully, again, USB4 improves this, but ideally I'd love some PCIe connectivity and especially multi-port designs (Lenovo has a couple gaming phones with >1 port in some markets, way cool).
I agree with the excitement for mobile. But it's also a dark segment, a pinnacle of consumerdom & regression to the mean versus that brief great amazing age of Personal Computer compatible, which has spawned systems both small and massive and mighty, that we have been able to extend & use however we might imagine. Phones returned us to an age of control, where our species is impotent at using the tech we have all around us, phones are infernal devices trapping us, are the spiritual foe of human spirit.
Amazon is not a secret weapon, poison pill maybe?
This is a company that is actively shooting itself in the foot on all fronts and by the time they realize it’s gone too far they won’t be able to turn the ship fast enough.
This is a company that is actively shooting itself in the foot on all fronts and by the time they realize it’s gone too far they won’t be able to turn the ship fast enough.
People desperately want Amazon to be failing but it just isn't. They made 30 billion in profit in 2023. The rumors of Amazon's demise are greatly exaggerated.
It’s not that people are wanting them to fail, just that they are being so anti everything whilst their competitors amass.
I’ve got friends who’ve moved from AWS-centric roles to Azure-centric, prime is a dead product walking, twitch is unprofitable, Alexa is dead, and their original business is squeezing users, drivers, and employees. A strong viable Amazon.com alternative and they are done.
I’ve got friends who’ve moved from AWS-centric roles to Azure-centric, prime is a dead product walking, twitch is unprofitable, Alexa is dead, and their original business is squeezing users, drivers, and employees. A strong viable Amazon.com alternative and they are done.
Sure, a store where you can buy literally everything and have it delivered in days, and they're done.
WalMart is like 90% of the way there. They'll bring it right this second, for a fee, or tomorrow for free, for anything in its store. 1 day shipping on anything in its warehouses. The only thing they're missing is the random do-dads I seem to need quite often.
I don't think they have the best name reputation. At least people my age, WalMart is synonymous with cheap and poor quality. But that was before the age of dropshipping and Amazon, where now Walmart's offerings are probably actually better quality than a lot of stuff. IMO they probably should have leveraged Jet there instead, but it is what it is.
I don't think they have the best name reputation. At least people my age, WalMart is synonymous with cheap and poor quality. But that was before the age of dropshipping and Amazon, where now Walmart's offerings are probably actually better quality than a lot of stuff. IMO they probably should have leveraged Jet there instead, but it is what it is.
"Delivered in days", my experience is that most deliveries get a "we couldn't find you" notice in my letter box, get bounced to a collection point, and that the collection point is often further away than the physical shop selling the same item. And two days after scheduled delivery.
Unless I want to buy something weird, in which case there's also a good chance delivery fails entirely for some reason and the order gets cancelled on the other side.
Unless I want to buy something weird, in which case there's also a good chance delivery fails entirely for some reason and the order gets cancelled on the other side.
My experience living in an apartment building is that delivered packages are just stolen, so I either have them sent to an Amazon Locker or collection point (a local Whole Foods has a counter for this).
Our apartment building has a package room with a camera and my packages have never been stolen from it, and the delivery instructions in my Amazon profile say to use that package room including how to get there and that Amazon can tap into the room. Often times delivery drivers don't do that and just leave the package on a shelf out in the open next to the mailboxes. I cannot even count the number of times someone has ripped into the package to see if it contained something interesting only to leave it behind when it's boring like pencils or cleaning supplies.
Amazon refuses to take any responsibility for packages stolen and has in the past denied refund requests without a police report on the incident (which I have filed multiple for just to get a refund though the package thefts are never investigated so I'm not sure what filing a police report is meant to do here).
Their delivery is inconsistent so I just don't bother. Though that does mean that I don't order as often from them as perhaps they'd like - if I can pick it up locally after work then I'll do that rather than order it and wait for it to be present and unopened only to be disappointed yet again at the Amazon driver who didn't read the delivery instructions which state multiple times not to leave the package next to the mailboxes because it will be destroyed or stolen.
Our apartment building has a package room with a camera and my packages have never been stolen from it, and the delivery instructions in my Amazon profile say to use that package room including how to get there and that Amazon can tap into the room. Often times delivery drivers don't do that and just leave the package on a shelf out in the open next to the mailboxes. I cannot even count the number of times someone has ripped into the package to see if it contained something interesting only to leave it behind when it's boring like pencils or cleaning supplies.
Amazon refuses to take any responsibility for packages stolen and has in the past denied refund requests without a police report on the incident (which I have filed multiple for just to get a refund though the package thefts are never investigated so I'm not sure what filing a police report is meant to do here).
Their delivery is inconsistent so I just don't bother. Though that does mean that I don't order as often from them as perhaps they'd like - if I can pick it up locally after work then I'll do that rather than order it and wait for it to be present and unopened only to be disappointed yet again at the Amazon driver who didn't read the delivery instructions which state multiple times not to leave the package next to the mailboxes because it will be destroyed or stolen.
Your experience is unusual.
In 2023 I placed exactly 100 orders and this year I'm probably going to hit the same number. I've had very few problems over the years (my first order was March 29, 2000). During that time, I've lived in a bunch of different cities across three states.
In 2023 I placed exactly 100 orders and this year I'm probably going to hit the same number. I've had very few problems over the years (my first order was March 29, 2000). During that time, I've lived in a bunch of different cities across three states.
I wonder what country you're in?
*in hours, for certain product/location combos.
Yep.
I remember one move I made to a new city. I'd downsized a bunch of my stuff and got settled in pretty quickly. Everything was right with the world except my Harmony remote didn't survive the move and my guitar hanger got left behind.
So I looked at Amazon to at least get a baseline price for these two things, and they offered to deliver both of them to my door in a couple of hours.
I sanity-checked the prices and they were fine.
"What is this wizardry," I thought to myself, when I had both items at my door in a couple of hours.
I remember one move I made to a new city. I'd downsized a bunch of my stuff and got settled in pretty quickly. Everything was right with the world except my Harmony remote didn't survive the move and my guitar hanger got left behind.
So I looked at Amazon to at least get a baseline price for these two things, and they offered to deliver both of them to my door in a couple of hours.
I sanity-checked the prices and they were fine.
"What is this wizardry," I thought to myself, when I had both items at my door in a couple of hours.
How'd that work out for the Sears Catalog?
I'm talking specifically about the catalog here - the rest of the recent (last 20 years or so) dysfunction of Sears is of a different kind.
The Sears Catalog was the Amazon of the early 20th century. So we must ask ourselves "Why did Sears discontinue its catalog service in 1993?" One of the answers to that question is the rise of retailers like Walmart where people could walk into a local (or at least closer to them than a Sears location) store and buy all sorts of things that they otherwise would have had to order. The catalog was inconvenient with other alternatives available.
The concern with Amazon is with its delivery ability - sure, for now, their unsustainable model that burns out drivers and pays them a pittance is working. Should that slip where they cannot deliver same-day/next day/day after reliably then that's an opportunity for other retailers to do to them what Walmart did to Sears.
You can even see it in this discussion - Walmart's online component directly competes and with many, many more local retails locations than Amazon and can often either have the items ready for pickup the same day or even deliver the same day.
So yes, there's a precedent for this, and if Amazon is focusing more on AWS and the buckets and buckets of money there but starts neglecting the retail part then that leaves a massive opening for competitors.
The Sears Catalog was the Amazon of the early 20th century. So we must ask ourselves "Why did Sears discontinue its catalog service in 1993?" One of the answers to that question is the rise of retailers like Walmart where people could walk into a local (or at least closer to them than a Sears location) store and buy all sorts of things that they otherwise would have had to order. The catalog was inconvenient with other alternatives available.
The concern with Amazon is with its delivery ability - sure, for now, their unsustainable model that burns out drivers and pays them a pittance is working. Should that slip where they cannot deliver same-day/next day/day after reliably then that's an opportunity for other retailers to do to them what Walmart did to Sears.
You can even see it in this discussion - Walmart's online component directly competes and with many, many more local retails locations than Amazon and can often either have the items ready for pickup the same day or even deliver the same day.
So yes, there's a precedent for this, and if Amazon is focusing more on AWS and the buckets and buckets of money there but starts neglecting the retail part then that leaves a massive opening for competitors.
Sears was destroyed by standard corporate raider tactics. I don't think that they serve as the example that you may think they do.
I remember my last time walking into a Sears. I figured I'd give it a shot to see if they had anything different than HD/Lowes in the category I was looking at (shop vac attachments). A clerk pointed me to a computer saying "If we don't have it in the store you can order it from our catalog". The computer catalog was literally just searching Amazon listings. That basically confirmed to me that Sears was in a death spiral.
Yep. They were already dead by that point (again, due to standard corporate raider tactics) -- a zombie just awaiting its coup de grâce.
There was a time before that when they had a better web presence for selling regular hardgoods than Amazon (which was still mostly known as just an online book store), and the Sears Parts website was the very first place to look online for manuals, diagrams, parts, and standard accessories for any random household thing (including shop vac attachments).
They also co-founded the Prodigy network back in the 80s when home computers were still very novel and people weren't broadly sure if the concept would ever catch on.
There was a time before that when they had a better web presence for selling regular hardgoods than Amazon (which was still mostly known as just an online book store), and the Sears Parts website was the very first place to look online for manuals, diagrams, parts, and standard accessories for any random household thing (including shop vac attachments).
They also co-founded the Prodigy network back in the 80s when home computers were still very novel and people weren't broadly sure if the concept would ever catch on.
I used to work at Starbucks in their corporate office building in South Downtown Seattle. The building used to be a Sears warehouse ago [1]. The walls next to the elevator in the ground floor still had printed some sort of "Sears customer contract". It had clauses like "you can return the merchandise if you are not satisfied with no questions asked. We will refund you for any shipping charges when returning merchandise". Or "if you cannot find it in our warehouse you can have delivered from our catalog to your home". These slogans reminded me of Amazon and their customer obsession.
Sears could have been Amazon if they kept their customer obsession.
[1] https://en.wikipedia.org/wiki/Sears,_Roebuck_%26_Company_Mai...
and
https://www.waymarking.com/waymarks/wm4RCM_Sears_Tower_Seatt...
Sears could have been Amazon if they kept their customer obsession.
[1] https://en.wikipedia.org/wiki/Sears,_Roebuck_%26_Company_Mai...
and
https://www.waymarking.com/waymarks/wm4RCM_Sears_Tower_Seatt...
You've just described a whole slew of online stores, their logistics aren't on par yet, but they will be.
Logistics is definitely Amazon's biggest moat. Second biggest would probably be their customer service.
Which, to be honest, has been downhill - amazon used to be super focused on customer satisfaction, but these days pretty much every interaction I've had with amazon makes me sad - pages of promoted products before I hit the product I'm actually searching for, account execs that don't seem to understand that they're their to fix problems not create them, and honestly AWS is lagging -- too many teams fighting for promotion by creating new services, nobody supporting or maintaining the plumbing.
I make most of my money supporting companies on AWS, but I've stopped recommending it to new shops largely for the same reasons - if you're not big enough for them to knock on your door, they won't answer when you knock on theirs.
I make most of my money supporting companies on AWS, but I've stopped recommending it to new shops largely for the same reasons - if you're not big enough for them to knock on your door, they won't answer when you knock on theirs.
What customer service? It's impossible to get a human
Where are you based? I just talked to a rep on the phone last week about a missing package (EU/DE). I've also had pretty frictionless interactions with CS over chat.
Customer service is just a bot, and anyone can have one of those
What I mean by their customer service moat is that they can afford to give full refunds at the slightest inconvenience to the customer.
I've never looked at the market place terms vendors agree to, but I can imagine return claw backs are a thing. Very similar to credit card merchant agreements for charge backs. A very lax return policy on Amazon doesn't have to be expensive to Amazon.
I'm not just talking about money. They can set whatever terms they like with vendors because selling on Amazon is a very attractive business for independent dropshippers etc.
How about you describe what your thinking here is, because as of now all your doing is trying to counter my comments by saying "no". You're adding nothing to why you think the moat exists and no other online vendor can compete with it.
I don't have the inclination to hunt for evidence for you. I'm not saying other companies can't do it. But it will be hard because of the capital/scaling/network effect requirements that Amazon has today.
This is my opinion and I'm not trying to convince you of it. Have a nice day.
This is my opinion and I'm not trying to convince you of it. Have a nice day.
I mean, they probably don't need their feet to continue making money and selling things online. IBM still makes money after all
Yes, yes, unified billing and security generally make up for how much Amazon's technical offerings suck donkey balls and those network effects are why we let them get away with pushing such god awful product. We all know how this works. We don't have to like it.
> Amazon's technical offerings suck donkey balls and those network effects are why we let them get away with pushing such god awful product.
What are we comparing to here?
Google Cloud? Azure? Linode? Openstack? OVH? Oracle Cloud? Hetzner?
They all suck too, just in different ways.
What are we comparing to here?
Google Cloud? Azure? Linode? Openstack? OVH? Oracle Cloud? Hetzner?
They all suck too, just in different ways.
Individually developed domain-specific products. Even ones that long predate the hyperscaler offerings are often much better... except for billing and security and integration. Not through any fault of the developers, but integrating those things takes work and the ability to reduce that work is a killer advantage of the hyperscalers. But the hyperscalers know this so they invest less in making the core product any good, so it isn't. Reduction in billing/security/integration suck -> increase in technical suck. That's my whole point.
I know we'll never know the answer, but I'd love to see a break down of AWS billings from expensive misconfiguration of cloud settings. After the recent Accident Forgiveness post from an AWS competitor it does seem like an area that AWS happily accepts. Small bootstrapped start up hires a 3rd party "expert" suggested directly by AWS to be experts. The 3rd party doesn't actually know everything and causes a 3000x increase in monthly cloud expense. Small startup uses their paid AWS support which resolves the expensive issue in under 10 minutes. The 3000x increased bill is still due. </rant> That's a story as old as time (of cloud), but it's gotta be a nice way to earn bonus for product managers
I mean, are we so unwise as to think those things are mutually exclusive?
I can think of a dozen quotes related to the decline of a thing being "impossible", and then sudden.
Enron for example had it's best year, right before it suddenly stopped existing.
"The candle that burns twice as bright, burns half as long", and "The night is darkest before the dawn" both relate to things appearing one way because the end is more likely.
Amazon is mostly inertia, they are user hostile to a fault, which is a far removal from what made them huge in the first place. Normally when there has been such a reversal in a companies principles: it's the beginning of the end.
I'm not desperate for it, I couldn't care less. But what we will witness now is the power of capital to beget capital, don't be fooled into thinking Amazon is innovative.
Don't anthropomorphise the lawnmower.
I can think of a dozen quotes related to the decline of a thing being "impossible", and then sudden.
Enron for example had it's best year, right before it suddenly stopped existing.
"The candle that burns twice as bright, burns half as long", and "The night is darkest before the dawn" both relate to things appearing one way because the end is more likely.
Amazon is mostly inertia, they are user hostile to a fault, which is a far removal from what made them huge in the first place. Normally when there has been such a reversal in a companies principles: it's the beginning of the end.
I'm not desperate for it, I couldn't care less. But what we will witness now is the power of capital to beget capital, don't be fooled into thinking Amazon is innovative.
Don't anthropomorphise the lawnmower.
Nothing in my comment suggest that Amazon dying is "impossible," but don't try to will into existence this idea that Amazon is a wind gust away from it all tumbling down. People are trying to say Amazon is dying and their business is in the mud but the financials don't support that idea.
and nothing I wrote suggests that financials and health of the business are linked.
There are countless examples of companies that seem to have a lot of money then die suddenly.
There are plenty of zombie companies (IBM) too, hard to know what this will be, but I wouldn't take financials as an indicator of anything really.
There are countless examples of companies that seem to have a lot of money then die suddenly.
There are plenty of zombie companies (IBM) too, hard to know what this will be, but I wouldn't take financials as an indicator of anything really.
> and nothing I wrote suggests that financials and health of the business are linked.
For profit businesses exist to make money. That is all. Any other claim is fairy dust and bullshit. IBM may not be making waves anywhere but they are still pulling at least 10 billion in profit so they are doing a hell of a lot better than most other companies.
For profit businesses exist to make money. That is all. Any other claim is fairy dust and bullshit. IBM may not be making waves anywhere but they are still pulling at least 10 billion in profit so they are doing a hell of a lot better than most other companies.
The US system has got you good.
The intended purpose of a corporation is to do something. Profit shouldn't be a goal in of itself, it's a consequence of successfully doing something.
Revenue is good, as it employs people and keeps society churning, but this slavish devotion to the idea that companies only exist to make profits is alien; it's the personification of greed.
Many companies exist because the founder had a passion for actually doing something, and they can make money doing it. If they chased only profit, their life would be miserable.
The intended purpose of a corporation is to do something. Profit shouldn't be a goal in of itself, it's a consequence of successfully doing something.
Revenue is good, as it employs people and keeps society churning, but this slavish devotion to the idea that companies only exist to make profits is alien; it's the personification of greed.
Many companies exist because the founder had a passion for actually doing something, and they can make money doing it. If they chased only profit, their life would be miserable.
You are making a lot of mental leaps to things I didn't say. You only stay in business if money in so I hope the something you are doing brings in enough to stay in business or else you won't be doing anything
> If they chased only profit, their life would be miserable.
Get some hobbies and don't make work your life.
> If they chased only profit, their life would be miserable.
Get some hobbies and don't make work your life.
Good points. I mean this is how we get old or get bankrupt. First slowly and then suddenly.
One point that is not mentioned is that Amazon is making boatload of money by advertising. Their ad revenue is 50 billion dollars and rapidly rising. This make them look invincible and total shit at the same time.
One point that is not mentioned is that Amazon is making boatload of money by advertising. Their ad revenue is 50 billion dollars and rapidly rising. This make them look invincible and total shit at the same time.
Amen.
I have had so many issues with orders from Amazon recently that they told me I’m on a watch list for returns and could have my account terminated.
I’ve been a customer for about 20 years with this account.
I’ll be shopping at other retailers now.
I have had so many issues with orders from Amazon recently that they told me I’m on a watch list for returns and could have my account terminated.
I’ve been a customer for about 20 years with this account.
I’ll be shopping at other retailers now.
If you had asked me five years ago if I'd be regularly ordering from walmart.com instead of amazon.com, I'd have thought it unlikely. But here I am -- I don't have to worry about counterfeits as long as it's sold & shipped by Walmart, and I can get same-day delivery (usually within an hour or two) for the cost of a tip. Their inventory is different (many more consumer staples, at better prices; many fewer random long-tail products), but it's replaced maybe 1/4 of my Amazon purchases. I also order from target.com once in a while; I never ever used to.
Amazon has not lost, but it is definitely losing its unique edge.
Amazon has not lost, but it is definitely losing its unique edge.
Out of sheer curiosity because it is such a different experience to mine, what are you shopping for online that you have so many returns?
I bought and returned:
- Meta Quest headset / it arrived in a shipping box that was dented, with the quests product box damaged, and when I went to try it the system wouldn’t launch apps
- a usb memory card reader that advertised support for older Sony memory sticks but wouldn’t read them
- an electric kettle that burned my hand to pour water out of at a normal angle for use and despite high reviews on the product page
Those are three examples in the last couple of months…
Those are three examples in the last couple of months…
https://news.ycombinator.com/item?id=41304337
My first use of Walmart.com and I won’t be going back.
Worth noting that the comment you reply to calls out buying only "sold & shipped by Walmart" items, as opposed to those from other vendors like your case.
Fair. When I bought the item I wrote about, I had no idea it wasn't direct from Walmart. IMO, as bad as Amazon does, they make who you are buying from very clear.
Remember ~70% of their profit is AWS. The consumer business gives them massive revenues but costs dramatically more to operate.
Amazon seems to be solving this problem. I've noticed a lot of things I've ordered recently are non-returnable
Ive been caught out by this, very annoying, but if you open a support chat you can often get a refund
Have you actually been returning an unusual amount of stuff?
They look at certain categories and punish you for their fraud. I got on the naughty list after returning 4 chromebooks.
Each had an issue as a result of Amazon’s shady practices. One was returned, sold as new. The replacement was defective.
So I ordered another model. First was damaged in transit. Second was a return that another OEMs charger.
I said “fuck this” and went to Best Buy, paid approximately the same price and had an actual, new, working Chromebook for my son.
Each had an issue as a result of Amazon’s shady practices. One was returned, sold as new. The replacement was defective.
So I ordered another model. First was damaged in transit. Second was a return that another OEMs charger.
I said “fuck this” and went to Best Buy, paid approximately the same price and had an actual, new, working Chromebook for my son.
I am on a watch-list too because of a recent problem with a broken package with only one item out of two inside. It was clearly a too light packaging for a heavy item with sharp corners.
The only other problem I had was a warranty claim 5 years ago.
My account is not closed (but they "reserve the right to close it" anytime). I am glad I have no kindle or DRM stuff...
I tried to get an explanation, but just got a robot email.
I will never buy there anymore. Even if some stuff are hard to find elsewhere
My account is not closed (but they "reserve the right to close it" anytime). I am glad I have no kindle or DRM stuff...
I tried to get an explanation, but just got a robot email.
I will never buy there anymore. Even if some stuff are hard to find elsewhere
My bet: Statistically, and adjusted for Amazon's relentless enshittification of the products they sell, he has not done that. But since Amazon's optimal strategy - short-term - is to just keep dialing up their shit, and dialing down their tolerance for people returning their shit...
>they told me I’m on a watch list for returns and could have my account terminated
Do they actually use such direct language? I received a passive-aggressive message from Amazon implying I was returning things too frequently (which was unfair IMHO; I had just returned three items that I had purchased as part of a bundle, but which might have made my return stats for the year look bad). But they haven't escalated it yet for me with more direct language.
Do they actually use such direct language? I received a passive-aggressive message from Amazon implying I was returning things too frequently (which was unfair IMHO; I had just returned three items that I had purchased as part of a bundle, but which might have made my return stats for the year look bad). But they haven't escalated it yet for me with more direct language.
Does designing custom chips to meet specific needs amount to self-sabotage?
It appears to be a reasonable strategy, as seen with Apple's successful introduction of the M1 chip, which set them apart from other PC makers.
It appears to be a reasonable strategy, as seen with Apple's successful introduction of the M1 chip, which set them apart from other PC makers.
How is the company actively shooting itself?
AWS is so anti-customer with respect to GPUs right now.
They have the highest prices of any cloud. What happened to “your margin is my opportunity”?
And, as far as I know, customers are unable to allocate a VM with fewer than eight A100, H100, or H200 GPUs. (Please tell me how if I’m wrong.)
So, customers are incentivized to use other cloud products for GPUs in the short term.
They seem to be heavily invested in their own chips in the medium term.
They have the highest prices of any cloud. What happened to “your margin is my opportunity”?
And, as far as I know, customers are unable to allocate a VM with fewer than eight A100, H100, or H200 GPUs. (Please tell me how if I’m wrong.)
So, customers are incentivized to use other cloud products for GPUs in the short term.
They seem to be heavily invested in their own chips in the medium term.
They typically come physically in 8s as an SXM5 tray.
It takes more effort to virtualize SXM5 than individual PCIe cards, so you find availability in 8s in VMs too regardless of provider
It takes more effort to virtualize SXM5 than individual PCIe cards, so you find availability in 8s in VMs too regardless of provider
They seem to be sold out of an extremely popular product and seem to be making good money doing so (per AWS’s results). How’s that “anti-customer?”
Also, in case you didn’t get the memo GPUs are hard to come by most places these days.
Also, in case you didn’t get the memo GPUs are hard to come by most places these days.
Meanwhile, I had a hard time last week getting a machine with 8 gpus from azure.
Every customer of a 3rd party cloud provider is going to have a hard time getting Nvidia GPU compute from 3rd party cloud providers - not only is the amount of GPUs available limited, but Nvidia themselves are trying to spin up their own cloud provider offering, and unsurprisingly don't want to help a competitor.
> but Nvidia themselves are trying to spin up their own cloud provider offering, and unsurprisingly don't want to help a competitor
I thought Jensen recently said he does not want to offer their own cloud offering. He instead wants to focus on creating ready made solutions for cloud vendors to purchase & re-sell services with.
I thought Jensen recently said he does not want to offer their own cloud offering. He instead wants to focus on creating ready made solutions for cloud vendors to purchase & re-sell services with.
Fair point! I think that's a recentish pivot though (past 2-3 years). I vaguely remember that in the late 2010s they were building and testing DGX Cloud as it's own standalone offering, but I might be wrong and confusing some other offerings they were working on.
They're trying to sell you their managed product, i.e. bedrock.
I am fully expecting a RISC-V Graviton (they might change the name) within two years.
OK, I know this is a bit off topic, but Prime (as a shopper) and Amazon quality of products is definitely tanking in my opinion. Has anyone had good experience with Walmart+? I've considered switching, but I'm skeptical.
Edit: Misspellling ‿
Edit: Misspellling ‿
Why are you viewing them as mutually exclusive things? Drop Prime (aka Amazon Sunk Cost Fallacy), and just accept that your orders will take a few days longer with their Slow Spiteful Shipping. I don't think the the Walmart sunk cost program even affects the shipping time, which mostly ends up being quicker than Amazon's Spiteful Shipping. Same thing with Target and every other large vendor regardless of whether they're pushing a subscription gimmick or not. Hitting a minimum order amount is really not hard once you've got items you buy regularly from each place.
I've been through 4 different calls with Amazon's outsourced to India Customer Support, taken high resolution photographs from 20 different angles for their incorrectly shipped product to me which shows they tagged products incorrectly.
Will probably take them to Small Claims Court. Worst Customer Support ever.
Such a scam.
My wife is happy with them though.
Will probably take them to Small Claims Court. Worst Customer Support ever.
Such a scam.
My wife is happy with them though.
More like "Amazon's Blindlingly Obvious Weapon In Chip Design Is AWS".