> When Amazon can buy stuff on your own website thats out of stock for wholesale prices without your knowledge, it's time to get your shit together. Your shop software is at least misconfigured.
I really wish the article had dug into that more, because it made very little sense.
Odd story. If you strip out the "Amazon" and "AI" stuff, the core seems to be that there's a tech company offering a service called Buy For Me which crawls various merchants who operate their own storefronts, lists the products they find for Buy For Me's users, and have a button the users can press which...buys the product.
Which is a little odd, and the value is questionable, but fundamentally seems...fine? You're a merchant, you're selling pencils on the internet, people are buying your pencils from you. And historically the way this might have been built would be something like a desktop application that users install, and which then goes and loads websites, displays them, fills in payment info, etc. Which of course is exactly what the web browser does already.
And all of the complaints about how it should be opt-in also feel odd. If you install WooCommerce and put a storefront up on the public internet, you've pretty obviously opted in to "selling your products on the internet". You don't need to tell Firefox that it's okay for people to use it to buy your stuff!
Of course, this isn't a desktop app, it's agentic AI run by Amazon, which certainly makes it feel different, but I'm not entirely sure how different it should make our analysis.
But also, the story raises a bunch of interesting questions and then doesn't answer any of them:
> Chua also received at least several orders for products that were either out of stock or no longer existed on her website.
How exactly did this happen? The story is that the orders are being placed through the normal storefront, right? So how exactly?
Or:
> Gorin sells wholesale through a password-protected section of her website, where retailers must submit resale or exemption certificates so orders are properly exempted from sales tax. She said she was still able to complete a “Buy for Me” purchase of a product pulled from her wholesale site despite never opting into the program — a scenario that could expose her business to tax liability if individual shoppers were able to place tax-exempt orders. Gorin also worries that surfacing wholesale pricing could undermine profit margins, allow competitors to undercut her prices or bypass minimum order requirements designed to keep wholesale sales viable.
That's just begging for an explanation. Is Amazon is somehow using stolen credentials to obtain price information? Or is Goren mistaken and the info isn't password protected at all? (And if not, why not?)
I'd also be interested in unpacking a bit more the legal and contractual implications of agreements like Mochi Kids has signed. The brand apparently doesn't allow its products on Amazon, and doesn't allow partners like Mochi Kids to sell on Amazon, but...Michi Kids isn't? Mechanically someone is buying the products at retail and effectively relisting them. Which...I dunno, feels legal? Is any agreement actually being violated here? Does the brand have a course of action? Does Mochi Kids have an actual legal obligation to opt out? Does Amazon have a legal obligation to let vendors opt out? Is Amazon legally buying anything from Mochi Kids, or is the customer the person using Amazon? Given the payment info being used is the customer's, I'm not sure Amazon has a commercial relationship with the brand or the vendor?
And so on. It feels like too much of the story is being carried by it being about Amazon and AI, which means the author felt fine just glossing over the details.
"Fundamentally fine"? How do you think Amazon would react to someone scraping their marketplace and posting the inventory under their own service? Unless the answer is "they'd be perfectly happy to have to opt out individually in each case", that's a double standard. The only reason they wouldn't actually need to care about this is because they have comparatively inexhaustive resources to be able to shut this sort of thing down with a sledgehammer without having to risk meaningful consequences rather than what they're telling sellers to do, which is to ask nicely to stop being included, and that's a sign of an unhealthy ecosystem where competition is non-existent.
The fact that Amazon even allows vendors to request de-listing (and the fact that Amazon does it promptly) would suggest that Amazon's lawyers have recommended that they do this (and that it's likely for a good reason).
We, as non-lawyers, may never know. But they obviously know something... Enough to spook them.
I have bought items from AliExpress which somehow arrived from an Amazon warehouse, unexpectedly quickly. The price was right and I have no other complaints, so I think it does happen.
Any fulfilled can use FBA (fulfilled by Amazon) for their own inventory and their own retail front end. Amazon charges storage and shipping fees and the items do not need to be for sale on Amazon.com.
>Which is a little odd, and the value is questionable, but fundamentally seems...fine?
1. Brand management is strict for a reason. You don't want some third party pretending to represent you and suddenly they become malicious or simply get hacked and have their customers (and indirectly, your customers) assosiswte you with frustration and danger. Or even something completely inconsequential in the grand scheme of things but a PR nightmare. Like having Nintendo products in a list next to some Magnum condoms.
2. There's subtle issues with making things "too convenient" to buy. Trackers and affiliates get frustrated, so it might make you less money in the long term. You might have related items to tempt buyers to buy more so spending goes down. Less users accounts (be it an email list, curation algorithms, or following on social media) weaken outreach for future holiday deals.
And those are 2 points when not considering a trillion dollar tech giant and the Ai concerns.
> Which is a little odd, and the value is questionable, but fundamentally seems...fine? You're a merchant, you're selling pencils on the internet, people are buying your pencils from you.
Why not apply your exact "its fine" standards to Amazon too ? Standards go BOTH ways, after all.
> In November, Amazon sent a cease-and-desist letter to Perplexity over its new Comet browser, which lets users ask an AI agent to find and buy items on Amazon. In a statement, Amazon said third-party shopping agents should “operate openly and respect service provider decisions” on whether or not to participate.
These people want Amazon to "respect service provider decisions" - just like Amazon demands of other people.
Yeah, seems like there’s a big fight brewing over losing the status quo of having control of how humans interact with stuff, as agents come along and make it so we don’t need to wind our way through the digital ad mazes they’ve constructed to do things.
But the incumbents who don’t want to allow this seem destined to lose, this is a tsunami coming, where this is just obviously how things will be done in the future once performance is good enough, and any group who tries to force customers into the old way is just not going to succeed for one reason or another. This is just how the market is shifting.
Yeah, this is structurally no different than hiring an assistant to shop for you.
The complaint isn't a moral one. It's fundamentally a trademark dispute. Manufacturers of goods want control about how their products are presented to consumers. Hermès doesn't want their stuff on the shelves at TJ-Max because it "dilutes their brand" or whatever.
Unfortunately trademark law doesn't speak to AI Agents, which is why there's a tech angle here. This is likely going to need to be solved with legislation.
> this is structurally no different than hiring an assistant to shop for you
In my opinion it's fundamentally not, because when you hire an assistant, you're hiring them with the intent to have them buy the product from the merchant.
Here, it would be like if you went to your local Safeway or other supermarket and there was a man standing at one of those sample carts who said "Hey, what you think of these papayas?" They're good, you look at them and decide you want two. "Great, I'll go in the back and get it." They disappear and come back with the papayas.
What's different:
1. You probably don't know where the papaya came from. Your intent in buying papayas didn't start with a clear understanding of the whole transaction.
2. You didn't interact with the merchant. If you want support, you have to go through the supermarket.
3. Whether you can file a credit card dispute is questionable. You likely won't win a dispute saying "I bought these and they're bad." You paid for a personal shopper, not a product. They substantially complied with their end of the transaction. You can't reliably dispute your instacart order saying "The papayas were disgusting." Instacart didn't sell you papayas, they sold you shopping services.
4. The merchant didn't sell to your email, they sold to some Amazon email. Good luck getting tracking details or getting customer support to talk to you directly. Good luck with returns.
5. Either Amazon is giving out your real credit card number (!) or using a virtual card. If it's the former, they've just invented credit card fraud as a service: you really going to trust Amazon's AI to hand out your card details safely? If it's the latter, you're probably going to get billed separately from the merchant charging you, which means Amazon is a middleman for refunds and payment issues.
In November I ordered a nozzle that I needed, which I knew had been discontinued. I ordered from a small seller, thinking they might still have some in stock. Turns out, they never even charged my card (probably because they don't have one and never will). I have been unable to get in touch with them about the order. I suspect this is very common, especially with drop shipping.
If Amazon charged me up front but they were not charged, that's outrageous. They don't even have a way for me to prove I didn't get my item (how could they?). Or will they mysteriously charge me at some point in the future? Who knows!
To add to this. Having a personal shopper is not new. Net-a-Porter for example do it. But you are paying for the personal shopper and the brands have a closer connection to their customers.
> the brands have a closer connection to their customers
That's... not a thing though. No such thing as "brand rights" [1] beyond stuff like trademark, which clearly doesn't apply here. In particular there's no inherent recognition of a manufacturers ability to control what happens to downstream goods. Stuff is stuff, if you sell stuff the people you sell it to can sell it too.
[1] Nor do we really want there to be? I mean, I get that this seems bad because ZOMG AMZN, but in general do we actually want to be handing more market control to manufacturers vs. middlemen and consumers?
As the source article covers, some manufacturers routinely ensure this kind of closer connection through contractual promises from authorized retailers. (Obviously any individual person who buys a product can still resell it, but for things like clothes consumers widely understand this to be a separate "second-hand market".) Amazon invests a lot of effort themselves in the consumer experience, they understand very well that stuff isn't just stuff and it matters how you sell it.
> No such thing as "brand rights" [1] beyond stuff like trademark, which clearly doesn't apply here.
I don't disagree with you on a personal opinion side, but the more expensive brands have a snobbery about who they sell to. To me it seems less about quality and more about "I'm rich" app style of fashion.
It's not bad because ZOMG AMZN, it's bad because *Amazon is a monopoly*, and thus anything they do to take more control should be treated with extreme suspicion.
Again, no such thing. There's no antitrust regulatory framework that recognizes the ability of "small" brands to constrain their downstream markets in ways "big" ones can't.
People are getting bent out of shape here, again, based on the specific player. But seriously what do you really think the solution is supposed to look like? I just don't see a fix here that won't make things worse, and I absolutely don't see one available under current law.
Did I say this was a legal argument? I don't see that anywhere.
And there's absolutely zero chance the current administration is going to take any positive antitrust steps unless the target just happens to be one that seriously pisses off Trump.
"Monopolies shouldn't be allowed to control everything" is a practical, economic, and moral argument before it is a legal one. If there is no legal framework to protect small brands from a company like Amazon coming in and doing these things, then perhaps there should be. (It's possible, though unlikely, that there's no practical way to do so without sufficient negative side effects that it harms more than it helps: I haven't sat down and tried to work out the second- and third-order effects.)
In case it's not abundantly clear, one very likely endgame of this for Amazon is picking the products within this subset that do the best, ripping them off itself (either fully legally, for simple manufactured goods, or questionably or outright illegally for things one buys because of the design—like shirts with particular art on them), and selling those under the cost the original creators need to be profitable. Those creators then go out of business. Then Amazon can, if they wish, raise the prices to whatever the market will bear.
The creators lose. The consumers lose. Even the wholesalers and manufacturers likely lose, if they're still involved, because Amazon is going to be paying them less for the same product due to economies of scale.
> Did I say this was a legal argument? I don't see that anywhere.
Ahem, I said that, in the comment to which you responded. Forgive me for making assumptions about the context of discussion.
But that said, I still don't see where you're going with this. No fix for what you want exists that wouldn't also outlaw stuff like fashion consultants, custom PC builders and thrift shops.
Of course it wouldn't, if those businesses weren't also monopolies.
It really is frustrating sometimes dealing with people on HN who assume that there can only ever be one set of rules for how businesses can deal with each other: that no matter how dominant a given company gets, you can never make them abide by a preset more-restrictive ruleset, or design specific rules for them that prevent them from abusing that dominance to hurt other people or businesses.
Antitrust law is specifically designed to do exactly that. It has been essentially abandoned over the past 3-4 decades in the US, in favor of Gordon Gekko's motto of "greed is good", with the Chicago School's "principles" essentially being "if it's more efficient™ for the economy, that's better; monopolies are more efficient™, so we should just let them do whatever they want," but what I describe is (more or less) what it is supposed to do.
If I were a merchant and I was bothered by this, I’d start figuring out how to exploit it. Ask your developers to code in the ability to detect the AI buyer (email address is a dead giveaway for now) and give them higher prices. Oh, you have an automated buy-bot? I smell opportunity.
I mean, the S in CFS stands for "syndrome", which is "a set of medical signs and symptoms which are correlated with each others [...] When a syndrome is paired with a definite cause this becomes a disease." (From wikipedia.)
So I mean, yeah, that literally does mean "we don't know what this is, and we don't know what's causing it, so we're dumping everything that looks like it in a bucket while we do more research". But that doesn't mean it's not a real thing; it means that we don't know what it is or what's causing it (and that it may well not be a single thing at all).
That's pretty different than saying "it's not a thing at all".
I don't think that's quite right? There's been a fair amount of evidence pointing at possible issues, but there's no clear answer due to poor (or just different) study design, small sample sizes, different criteria across studies, different sample groups, etc.
> ...it is difficult to establish the role of mitochondria in the pathomechanisms of ME/CFS/SEID due to inconsistencies across the studies. Future well-designed studies using the same ME/CFS/SEID diagnostic criteria and analysis methods are required to determine possible mitochondrial involvement in the pathomechanisms of ME/CFS/SEID. [...] There is consistent genomic research suggesting that ME/CFS/SEID is not a primary mitochondrial disorder, however, mitochondrial decline might occur due to secondary effects of other disrupted pathways. [...] As population samples were small, these results should be interpreted cautiously.
I wouldn't summarise that as "no evidence". It's more like "ME/CFS doesn't seem to be a genetic disorder causing defective mitochondria, and the mitochondria look the same, but they seem to function differently for some reason even if we lack enough data to figure out why yet". Note that, eg, of the 19 studies reviews, 5 tried to check for differences in mitochondrial respiration between ME/CFS patients and healthy controls, and 4 of the 5 found notable differences; one study was able to reliably detect if a cell sample came from a ME/CFS patient or a healthy control based on measuring mitochondrial respiration.
I don't know that's enough to fully reject the null hypothesis just yet, but it's certainly not clear we can accept it either.
>5 tried to check for differences in mitochondrial respiration between ME/CFS patients and healthy controls, and 4 of the 5 found notable differences
But did they look at the same thing? I also don't think that includes all the studies that failed to show mitochondrial differences, and failures to replicate previous studies.
There is the recent Ryback study (currently a preprint) which failed to replicate Fluge and Mella's result, and showed no difference from controls. There is the Tomas study which showed no difference in the ATP profile test from controls. Also, a 2019 Tomas study showed no difference in respiration between patients and controls.
> A strapping newborn baby boy is understood to have set a New Zealand record, weighing in at a whopping 6.85kg (15lb 1oz) - the equivalent of nearly seven 1kg blocks of cheese.
I mean, the kilogram is an SI unit, but uh, I do not know if clarity has been added here.
Fundamentally both "gambling" and "insurance" are about trading a stream of small guaranteed payments for a chance at receiving a large payment. More generally, it's about trading risk. In insurance the risk is transferred to the insurance company; in gambling the risk is trasnferred to the gambler.
They're precisely the same transaction, just packaged in different ways, and with tons of overlap. What differs is the participants.
Individuals don't tend to want or need a lot of risk, so a transaction where risk is transferred from an individual to a large company (like an insurance company) is generally good, and one where risk is transferred from a large company (like a casino) to an individual is generally bad.
The problem is that Google forces actual good cooks to make their recipes look like worthless blogspam, but a good original recipe is not actually worthless blogspam, even when disguised in the way Google requires.
When it looks and acts like the spam sites, then what difference is there really? If I have to scroll 4 pages to find the ingredients and then scroll around like crazy to find the instructions (then scroll back and forth while cooking/baking) then it does not matter how good the recipe is, the page killed it for me.
I'd argue that most web users have a higher tolerance for ads than HN users, so they put up with the scrolling. And if it results in a tasty recipe, then they'll do it next time too, since that's seemingly the (tolerable) price to be paid for good food.
But lots of recipe sites now have a "jump to recipe" link at the top, so they've realised the junk is annoying for some fraction of their users. Although page junk is a pain, shortcuts for low-tolerance users seems like a good compromise.
Look it's not OK to milk humans like this. It's manipulative and rapey. Just because the NPC meme is true does not mean you get to hack their programming for a buck and call yourself a good community member and businessman.
Because websites aren't free to build or run. No one is obligated to put ads on their site, sure. They're also not obligated to work for many hours to provide you with free content or pay $X/no to serve it to you.
But they can also have a separate job that doesn't ruin the internet and produce out of generosity, like some of us, free content that is not span ridden.
Also web hosting doesn't cost much when your website is well made with some frugality in mind.
And there are also better, cleaner ways to make money on the internet: getting rid of the ads and spam and having the content accessible to paid members.
While it is admirable that you are willing to produce content out of your own generosity, it seems a little optimistic to assume that everyone making content on the internet is both willing and able to share it for free.
I am somewhat curious to hear more about the better and cleaner ways to make money on the internet, but I have a suspicion that in some circumstances (such as recipes) they may put you at a competitive disadvantage. I certainly have no desire to pay to access recipes I find via Google searches.
We need to find a metric for anti-profitability. I think that index could yield much higher quality results.
Detect sales/commercial language and structure,* and specifically target that for removal from results as if sales-oriented sites were hardcore porn and the child safety filter is turned on.
*Buy and cart buttons/functions, tables containing prices with descriptions but don't look like long-form reviews (which would be it's own filterable tag), etc, and domains trying to obfuscate are blacklisted permanently.
Really just removing all sites with ads would be a huge improvement. Regular old websites trying to sell you something are usually not nearly as bad as those that want to monetize you while pretending to be free.
The thing is, even if you don't put ads on a page or tool, Google will sometimes not index it because it doesn't think there's 'enough' content, no matter how little sense that makes. At least half the issues with recipe sites and company sites come from them trying to get a site that doesn't need reems of text content indexed by a search engine that seems to blindly value the quantity of content and time spent on the page over all else.
The people who have bad content are the ones to get money, while those who have good content are not. Logical result is that people with good content stop producing that content while the people with bad content continue producing it and being rewarded for it.
Look I hate these SEO-laden pages just as much as the next guy, but I think the binary classification of "good content" and "bad content" lacks nuance. I would refer to it instead as "bad packaging" of (often) good content. As much as I loathe having to hunt for the "jump to recipe" button on my phone each time I open one of these pages, I also appreciate being able to freely view recipes which I enjoy and cook regularly.
I just stopped looking for receipts online if I can avoid it. It became literally faster and easier to search in old school cook book. And there was period when I considered those completely outdated.
I sympathise with the author and agree that there are some pretty terrible stories and numbers floating around, but the unstated premise of the article seems to be:
The game industry is monolithic, it has recently undergone a sweeping change, but no further change is possible, therefore it is safe to make a linear projection based on what's happened over the last six months out to infinity. And I'm skeptical that those are good assumptions to make.
> Epic Games in September laid off over 800 people, almost 15% of the entire company. Epic is one of the most successful and profitable game companies that exist. [...] For years, every other game company has tried to copy Fortnite, and mostly failed at the attempt. This is not enough to ensure job security
Okay, yeah, that's interesting, and no doubt traumatic for the people who were fired, their colleagues we have so far avoided it, and for devs working in the broader industry, and I am sympathetic. But: Sometimes companies overhire, sometimes corporate priorities shift, sometimes a company decides to reorient towards a leaning production model. Sometimes it works; sometimes it doesn't. There isn't any industry where overall profitability is, alone, enough to ensure job security. (And I'd also suggest that putting these layoffs in the context of the broader tech industry and the end of zero interest rates might also yield some useful insights...)
In any case, if Epic fires a bunch of devs, and profitability drops, then that was a mistake and they will try to staff back up. And if it doesn't (or rises) then that suggests that game development is actually more profitable than previously expected and individual game devs are more productive than previous realised, which will of course be cold comfort for the devs laid off, but suggests that overall employment and compensation across the broader industry will be tracking upwards not downwards, which is good news for game devs as a whole in the medium to long term.
To be clear: I do not want for one moment to defend the big studios (who appear, by and large, to have C-suites full of pod people who delight in human misery), or to minimise the very real pain suffered by gave devs, but the idea that an entire industry can somehow run off a cliff in a way which is permenantly non-recoverable is...well, let's say it's a bold claim that needs extraordinary support.
> Games in 2024 and 2025 will be a few labors of love...
Yeah, plausible. But what do you think is going to happen in 2026 and 2027?
We're talking about a hundred billion dollar entertainment industry. Video games have come a long, long way from the odd circumstances in 1983 which notably only affected the nascent industry in America, not Japan or Europe.
I really wish the article had dug into that more, because it made very little sense.