The right way to do this is to create an API or MCP or whatever catering to bots, and implement corresponding terms in your legal agreements (i.e. after approving a bot to interact with us, you're responsible for purchases it makes), along with some limits to prevent runaway shopping when Muse decides to binge.
Less friction = more shopping = more profits.
That used to be a concept companies like Amazon understood well. (As well as companies like Google in the context of web browsing and keeping it user-friendly as opposed to user-hostile).
If the bots mean customers are getting a better deal elsewhere, then you're doing something wrong elsewhere in your business model and need to go address that. Or eventually free market economics will, no matter how big an artificial moat you try to build.
> The right way to do this is to create an API or MCP or whatever catering to bots, and implement corresponding terms in your legal agreements (i.e. after approving a bot to interact with us, you're responsible for purchases it makes), along with some limits to prevent runaway shopping when Muse decides to binge.
From Amazon's perspective, that would be very much the wrong way.
Amazon's concerns here aren't technical. It's entirely about being disintermediated.
Right, agents work on behalf of the user and only cost compute (not time/attention, which is the ultimate human currency). They act on behalf of me not Amazon.
It is my personally optimistic perspective on AI that it will mediate all interactions with 3P who are acting in their own interest and not mine. Ie, “my” agent.
maybe agentic commerce will finally realize the micro-payments based internet that crypto people had been wishing would happen for years? Should there be stablecoin based micropayments built into MCP to charge for agentic interactions to replace the micropayment of an ad into someones eyeballs?
This is (still) never going to happen. A more realistic outcome is a deal between Meta and Amazon to enable access in exchange for invisibly integrating advertisements into the agents' decisions.I don't know exactly what that will look like, but I'd bet we will see a version of it soon.
It can only be this. And it will fail the same reason Alexa failed.
Nobody will blindly trust Amazon to buy products when they know Amazon cannot and will not truly suggest the best product for the consumer.
Alexa was supposed to be the voice gateway to purchasing. But nobody trusts the result Alexa picks because it will not default to the best value to the user - only the best value to Amazon.
Recently it was estimated that Amazon makes $30-80 billion(!) per year off not showing you the economically best results by deliberately disallowing sorting on $ per unit [1] to force sellers into ads. Amongst other tactics (ex. filter by country of origin).
AI agents can sort on $ / unit to easily find you the best deal, rendering Amazon's various forms of paid placement irrelevant.
Extend this to filtering Chinese slop brands (also not able to be filtered) and the significant portion of revenue starts looking like knocking back margins to the 3-5% range. Basically an existential threat.
Amazon cannot allow AI agents and maintain current margins.
The AIs are also providing the product suggestions, and it comes from what data they are given from their searches. I think they can put 2 and 2 together on how to make that happen with the right MCP response bodies.
I asked this in another thread: who benefits from all these bot integrations?
It's not the consumers of the AI, or the consumers of the real businesses? It just seems like everyone things "if they build it" it'll come.
But it's clear Amazon cares about how it's costomers have to view all the items as both upsells and advertising, and a bot is going to either be _really_ into it or act much like ublockorigin does.
But either way, no company making money on costumer direct sales sees a real benefit to allowing a bot to easily bypass all the necessary details.
Whether it's advertisement, liabilities, manufacturer's recommendations; a bot will ignore these things and generate a costumer expectation that suddenly the producer has to live up to.
Then you'll tell me: YEAH, that's why you need a MCP friendly website.
That experience actively drives me towards other retailers.
I want to buy what I want to buy, and I will never buy things that I do not want to buy. If the time ever comes that I am tricked into buying the wrong thing from Amazon, then I'm sending it back on their dime.
If that means that only 3 things fit, then I only want to see exactly 3 things. If that means that 0 things fit, then 0 is also a valid result. In response to these results, I can then elect to soften my constraints at my own volition.
But an Amazon search typically results endless pages of vaguely-related products that represent a jumbled mess of things largely do not fit or only-sorta fit the broad concept of what I intend to purchase.
When I want a 4300 Kelvin D2S HID headlight bulb for my car, then that's all that I want to find when I search for it. I do not want results for LEDs. I do not want selective yellow bulbs. I do not want 5000k or 6500k or 10000k. I do not want an "upgrade" kit, nor D2R bulbs, nor D3S. None of those things are useful to me.
The list of things that I do not want is infinite, and when Amazon presents a seemingly-infinite list of them then that represents the opposite of usefulness.
I only want to peruse listings that match. I require no more and no less than this in order to be satisfied. There is no possible amount of additional noise that can ever dissuade me from seeking the thing that I very specifically desire.
A modern LLM can weed out and separate the bad results from the good ones and give me an easier time of making a purchase decision, but it can only do so when it is not blocked from the start.
That ship has long since sailed for every search product. Companies have long-since learned that intent-based results are much more profitable (why show an empty result page when you can show something for people to buy or click on?), but it is also often genuinely useful when your search is not perfect or what you want doesn't actually exist. I agree that it is annoying when you really truly do only want something very specific and the ranking prioritizes the wrong things (e.g., sponsored results, more profitable items, etc.) over exact matches, but can you honestly and truly say you've never found any value from intent search algorithms?
I very often truly only want something very specific. I'm not apologetic about that at all.
When I want oolong tea bags and I search for "oolong tea bags," then I definitely don't want to be shown bulk green tea instead.
If I want to find any kind of tea instead of specifically oolong -- if, perchance, I wish to peruse the broader market of available tea -- then I'm very capable of removing "oolong" and/or "bags" from my search prompt.
---
I can look around and observe that the ship has sailed, but that doesn't mean I'll ever be happy with the outcome or accepting of any notion that the situation is anything other than hostile.
Can you recognize that it gets between me and my willingness to spend money?
You mean if they falsify the search for both sides they make more money, as ironically spelt out in Google's original paper "The Anatomy of a Large-Scale Hypertextual Web Search Engine":
“We expect that advertising funded search engines will be inherently biased towards the advertisers and away from the needs of the consumers.”
“advertising income often provides an incentive to provide poor quality search results.”
“the better the search engine is, the fewer advertisements will be needed for the consumer to find what they want. This of course erodes the advertising supported business model...”
Note: they even explain why advertising funded search makes things worse for customers and for advertisers. Worse search finds more ads and thus more dollars, resulting in both worsening search results over time AND less value for ad money over time.
Amazon is only an interface to purchase products made by other companies (except for Kindle devices).
Your alternative interfaces exist and have always existed: Go to a physical store, order by phone, order through the OEM website, order through an Amazon competitor, etc.
"you can already build [Dropbox] yourself quite trivially by getting an FTP account, mounting it locally with curlftpfs, and then using SVN or CVS on the mounted filesystem."
For all the complaints, it's nice to see HN hasn't changed that much.
Yes, you're right. It's almost impossible to purchase from a physical store or from competitors to Amazon (who many times are much bigger than Amazon in their country).
On the next episode of Oppressed Hackers: We get a load of a guy who had the nerve to say that it was possible to drink a glass of water.
The point wasn't that there are no alternatives, but that calling Amazon "only an interface to purchase products made by other companies" is reductive to the point of being completely wrong.
It makes total sense for Amazon to block agents. It sucks having to manually check multiple sites for prices and delivery of something you want so is easier just to use Amazon which is usually very good. If people start using agents Amazon is less likely to get the sale and are more vulnerable to competition. Better to kill this trend off asap.
Monopoly is a word that has a real meaning, and not a blanket term for companies you personally don't like and not even a blanket term for big companies.
Amazon has nothing even approaching a monopoly. You can go right now to a store built from bricks and purchase goods.
Sortof? It probably doesn’t have a monopoly in any segment that matters, but it does have many illegal monopolies (as defined by US FCR and the Sherman Act).
Evidence: Amazon has been found guilty of being an illegal monopoly multiple times already.
I don't think it does make sense. I think Amazon is way overestimating brand loyalty, and underestimating how much of their success comes from being a one-stop shop for everything.
I'm using Claude for everything from finances to dinner planning to project planning. If Amazon doesn't let Claude search and buy, I'm not abandoning Claude, I'm abandoning Amazon.
As long as my actions are as simple as possible, I don't care whether Claude is searching 5000 online retailers or only 4999, minus Amazon. I'm certainly not going to say "ok, this project requires buying 13 things for about $500, but I don't see amazon in the list, so let me go manually search for each one on Amazon and see if they might have a better price or something."
This is Walmart / Apple Pay all over again. Yes, near-monopolies, lots of power. But they got there by making things easy for consumers. Making things difficult to force consumer behavior never ends well.
I never let agents make purchases unsupervised. But that wasn't the point; the point is that the convenience of getting a complete list including multiple-vendors when pricing makes sense is way better than going to amazon and searching one by one, therefore Amazon will lose this power play.
The same would be true for someone shopping for one $10 item.
I always check Google before Amazon, because it's impossible to find anything on Amazon. Amazon is a bit like the bargain bin of a DYI shop. If it's immediately visible it's okay, of not, don't bother searching.
Half the time I can't even find the category page. And if I do find it, the filters will just be useless anyway.
In general, preventing consumers from having fully featured AIs (especially local ones) will be worth about 10 trillion dollars to companies. That's my prediction.
Amazon is one example. If my basement AI drives my browser and never clicks an ad, Amazon loses big.
Or, the ad industry in general -- imagine the basement AI filtering out all ads at the analog level so nobody in my house ever sees an ad, and no software update can change that.
Or other optimizations. Right now companies enjoy a power imbalance because they have people who ruthlessly optimize 24/7 to save a buck. I can't do the same, because I have to work, and also prefer to spend my personal time with family -- a foolish thing to do in a capitalist society, at least from a capital perspective. Imagine I have an AI wheeling and dealing on my behalf 24/7 too. Especially in an age of online shopping and product deliveries, my AI will absolutely order from your competitor to save a penny, my AI will absolutely spend 400 hours disputing a medical bill I don't think is right, barely an inconvenience to me.
One of the optimistic aspects of what's happening is that it doesn't seem like the leading model capabilities are outpacing the more open ones by that much- that just as a general principal there are pretty ok models relatively close behind the best ones.
This could mean the long term viability of my open claw bot to shop for me- that openai / google etc won't control AI, and I'll still just be able to use a model via API to do what's actually in my own self-interest.
The ads (in general) will be redirected to where you will see it, regardless of whether or not Amazon allows bots. You don't think you're getting off that easily, do you?
Once AI is good enough the AI will control what I see, plus, I will just spend a lot less time looking at screens altogether.
It's also trivial for Amazon competitors to pop up in this potential AI future. Maybe you only sell door knobs, but all you need to do is set up a mostly text-based API designed for AI consumption and you're in business. If you offer the door knob for 10 cents less, why wouldn't the AI agents buy from you?
Right now making an Amazon competitor will fail mainly because humans are lazy and can't be bother to check other sites. AI might change that.
I still prefer the Amazon shopping experience to anything else, unless I already know the product is available locally. The primary quality problem I've run into is the quality of the products themselves.
That's irrelevant from Amazon's perspective. If an agent is popular enough and has some fingerprint they can identify, they have a strong incentive to block it.
Agents that scrape and digest web data are entirely anathema to brand-reliant and advertising-reliant vendors and marketplaces. Even human assistants could be trusted to develop brand attachment and respond to advertising meant to shape their preferences. And while LLM agents are still manipulable, that's a frontier that nobody really knows how to navigate effectively yet, at least not to the degree of optimization modern vendors and marketplaces (especially Amazon) have acheived.
It's quaint to imagine what AI agents might do for consumers in the naive internet that preceded their emergence (if you ignore their vulnerabilities and flaws), but that's not what they'll be able to achieve as that same internet responds to them -- at least not at Meta scale. It's not very different than how social media companies once offered, and then aggressively withdrew and policed, API's to their graphs and feeds. Those open API's simply aren't compatible with the business objectives that keep these companies running, no matter what cool and profoundly user-friendly things could be built on them.
The same kind of incompatibility applies to AI agents when it comes to e-commerce. It won't be a permitted thing until, and except for where, vendors/marketplaces can assert control over the agents themselves and manipulate them to ends comparable to what they can achieve with human consumer sentiment already.
What's crazy is that the ToS is basically, "allow us, Amazon, to give you algorithmically higher prices- and if you try to circumvent that, you can't buy from us anymore".
I think the arguments for these kinds of near-monopoly companies being true market-driven capitalism seem more and more false.
The idea that Amazon will tell you that you don't even own the right to your own shopping experience tells you everything you need to know. Apple owns your device and Amazon owns the right to show you whatever price/product they want, regarless of your search intent.
I sometimes look around at people proposing technical solutions for these political/legal/economic problems, and think: "Was I that naive X decades ago, or were things really different then?"
I don't care about Meta even one bit so more power to Amazon. At the same time, Amazon is going downhill at an alarming rate. Is anyone else annoyed by the no name brands flooding every product listing there? There is no way to even enunciate the freaking names! How am I supposed to trust them?
The other issue is overpromising delivery times and vastly underperforming once the order is placed. The worst is when I purchase something that I want urgently. Let's say the price is $23. Amazon suggests buying something cheap to make it $25 and get overnight delivery. So I add some Chobani product from Amazon Fresh. The next morning I wake up to see Chobani delivered while the item I want is delayed until the next day. This has happened 3-4 times now so I have become wiser. But what a crappy customer experience from Amazon!
It's possible that it's location specific. We have a Whole Foods nearby so Amazon invariably shows me Amazon Fresh products. I remember once I even added a lime to the cart because my total was $24.99 and it wanted me to make it $25.
I ask this in every thread on muse / OpenClaw / whatever platform. I’ve never gotten a clear answer beyond “yeah me too”:
Does anyone have any meaningful use cases that have actually improved your life in some measurable way?
I’ve tried hard and beyond filling out some tedious forms in about the same time and energy as just doing it myself, I’ve not personally. I’ve tried hard. So I figure it’s probably my lack of creativity given the prevalence of fans. But every time I ask, I get no meaningful responses. So it makes me wonder if all the fans are just product managers on those products?
- Dusted off a 3d printer that wasn't being used enough, largely because of the tax of setting up profiles for materials and each print. That was fun once, I'm too busy now. Was able to get claude to use MCP and configure the slicer just right for the job and filaments.
- Needed to get IPTV/etc set up for football season. If you've done this you know what a pain it is, how EPG's don't always work right, finding the 50 channels you care about out of 80k channels is a nightmare, etc. Pointed claude at my utility server, it got Dispatcharr configured, wrote scripts to parse channel guides from sources not covered in the EPG, scheduled to update daily, deployed.
- I'm considering getting back to 3D printed clay, which is fun but challenging. Based on a thread here on HN, had claude go investigate using FPGAs (which it would program) for a realtime flow modulation system that would remove the variability from slight changes in viscosity within a batch of clay. Not sure I'll pull the trigger but I have a much deeper understanding of the solution.
- Realized I was paying like $30/mo to RocketMoney and Copilot (finance, not MS), for rare usage and basically just occasional "find subscriptions I forgot about" usage. Got claude connected to plaid via MCP, found more than those other two ever did, got it to set up cron to send me email weekly with current subscriptions and any price changes.
That's literally in the past 4-5 days. I find great value there, YMMV of course.
I find it useful, so here are some concrete examples from last few days.
As a baseline which many skeptics will have a hard time accepting for their own usage - I think there is a correlation between utility and trust (even which is not truly earned). If you don't let it capture memories and learn about you, and you don't connect it to external tools like email, it's less useful. But I think of it usually like a ChatGPT w/ extra tools and knowledge.
I just moved, and last night I asked my agent how to test how safe the soil in my (urban) garden was, and it found some warnings from my city government, and suggested a particular test service that is available in my area. Since it knows exactly where I live, it was able to find recent warnings of contamination in my neighborhood, and suggest specific chemicals to test for.
Additionally, I give it (read only) access to my email, and it'll ~2x a day notify me of some important email. I'm working with a contractor, and getting emails from them bumped, or getting emails about AWS usage alerts pulled out of inbox and notify me directly is super helpful for filtering the noise of my crazy email volume of 99% spam.
Also, as others have said, it's easy to take a picture of a flier, and say "add to my calendar".
At work? I've got ADHD, so the idea of making a to-do list is fine but keeping it up to date was a challenge. I've created a set of agentic skills that comes through the last 36 hours of emails, meetings, chats, and gathers up all the work I volunteered to do (or was told to do), triages them across urgency/impact, and sorts them into Do/Defer/Schedule/Delegate piles.
Last, I created an app to give a UI to this data to see and act on it all, including a way to flag a to-do that was found and enter why I thought the status or categorization was wrong. On the next run, a review skill reads those comments, makes changes to the other skills accordingly, and sends me an email with details of the changes, and the commit I can use to roll back the change if I'd like.
I've been running it since April and while there were a lot of updates at first, I haven't had to flag anything for review in months, and I get to feel a whole lot better about managing my to-do pile.
I'm looking at open-sourcing it, but since I created it with work resources on work time, I need to jump through a few hoops first.
For myself, it's been super helpful in at least two ways:
1) Staying on top of my kid's school obligations, homework, etc. School Learning Management Systems are a hot mess, and I use agents to consolidate and collate, to keep track of the homework that has been done, what needs to be done, and even pre-grade homework to give feedback (most of the work is done on digital platforms too). It's been immensely time saving for these kinds of things.
2) Multiple instances of watching things passively: hotel room prices, availability of tickets, finding the best place to sell things online. In general things that require a lot of browsing, collating, and watching. These are the kinds of things that specialized software products could do as well, but with an agent they are just a prompt away, and you get exactly what you want. (More or less.)
As someone who doesn't have kids and was last in school decades ago....
Do those School Learning Management Systems tell you everyday what homework your kid is expected to complete? Do you feel such a granular degree of monitoring and direction has impacted your child's opportunity to build character, intrinsic responsibility and self-management skills?
When I was a kid, teachers gave out assignments, and the student was responsible for keeping track of them, staying on top of progress, and handing in by the deadline. If a problem arose that wasn't an isolated incident or resolvable directly between the student and teacher, the feedback would flow at a much higher level e.g. scheduled parent-teacher interviews or via a phone call.
And what do parents do who don't accept the terms and conditions of said educational platforms? (e.g. due to overly broad privacy clauses or one-sided indemnities and the like)
But nowadays kids are expected to check 2-4 platforms plus their notes and there is no guaranteed time when they know it will be there. It is like the worlds worst paid disfunctional work place and one of my kids schools literally uses teams for one third of their assignments. In word format so sometimes kids overwrite it.
If you don’t accept the terms your kids life is just even harder.
Keeping track of your kids' homework is your concern? Doesn't it only become your concern, briefly, when the school contacts you about your kid not getting their homework done?
As a parent people from prior generations do not get the current situation. I actually agree that this is a useful use of the agentic work tools - and it’s the only case I mentioned above I’ve used it for successfully, if not saving of time.
Homework isn’t assigned in class any more. It’s posted on dozens of inscrutable websites in weird UX’s that are impossible for any single human to understand. Children are mystified, parents confused, teachers at their wits end.
A few times a week children miss some website or queue from some system, teachers can’t get things to share properly, this needs tending, they need to photograph this and upload that to this place as well as that but not the other one.
It’s truly remarkably astounding. I have serious doubts frontier models can navigate it either honestly. But this is in fact where I see the most chance of using these tools.
“Computers at school” is the worst idea ever, and no one who hasn’t experienced the clusterfuck of middle school in 2026 can understand it. It makes the world of the movie Brazil seem utopian.
Btw I did this Muse, giving it the ability to use its cred vault to log into her school information system and read through all the crap and other websites and collate the homework’s and their due dates, adding to my child’s calendar when they’re due and a reminder the day before with a link to the calendar. I had model the workflow formally (I do this with all LLMs for procedures) and set it up to run daily. We shall see!
Overall it was not bad. The cred vault is useful. But I’m not giving meta any information that I wouldn’t otherwise be fine with disseminating and I feel the burden the school puts on us vs the information loss is worth the trade.
Thanks! First meaningful use of an OpenClaw like system.
I'm not the person you're asking, but in general, I would also stay on top of my child's obligations - homework, chores, etc. I don't want the first sign of trouble to be a call from a teacher, or the complete absence of clean clothes for them to wear in the morning.
You don't teach children responsibility by removing yourself completely from that equation
You equally don't teach them responsibility by always being there as the backstop though. I think missing a couple of homeworks is quite a low-stakes thing that can teach them a lot about responsibility, failure, and how to recover from it.
A lot of kids now seem to go into adulthood having never been allowed to fail, and it can really mess them up the first time it inevitably happens.
There is quite a large gap between checking in with your kid and having an AI analyse their homework and assignments and raising an alarm if anything appears to slip. The latter is a solid WHUP-WHUP-WHUP-WHUP-WHUP on the helicopter parent spectrum.
Missing homework once or twice is part of life and a really good life lesson with low risk besides.
Travel research is so much better with an asynchronous agent than with normal LLM chat. You can just let it hack away for an hour looking over options and browsing different websites.
I'm not using a platform but I set up a telegram bot that is backed by AI (DeepSeek flash tokens are almost too cheap to meter) that handles my strength training workouts. It reminds me of the days workout, takes my feedback on each exercise difficulty and adjusts it for the next round, handles injury rehab prescriptions, and so on.
I'm now considering putting a coding harness inside it so I can actually ask it to improve itself on the fly, since I realized that having to switch to my laptop just to give a short feedback is a bit silly, why not right through the bot?
It's working well so far and helps with my tendency to forget my workout program for days at a time since it pings me daily.
I agree with you. I also wouldn't trust it calling on my behalf or doing the majority of what people say it can do. Putting aside it making mistakes and not even being beneficial, I definitely wouldn't want meta of all companies to have the relevant information to do those things.
Just in the last three hours, Muse helped me make progress on my go-to-market plan, drafted me a script for my demo video, and went and listed 50 high value creators in my target market that i can go interact with and ask for product feedback. It could even DM them for me if I wanted to. It's the first time a product like this has actually shocked me... it closes the gap between decision and action. I'm hooked.
I still have to do the work, but everything's all teed up for me.
I use Claude Code often for entirely non-code things: the harness concept is super helpful for going through documents etc.
I think these Claw-type harnesses are basically this for somewhat less technical folks. If eg Muse has been able to make this work smoothly from your phone for completely non-techy people, then that seems like an incredible product.
i just have the claude scheduled job kinda tool running, but at the grocery store i can spot something and think of a meal, then ask claude if any nearby grocery stores have other key elements to the meal on sale.
its pretty handy.
i was trying to get a nice setup for finding what shows are happening around me, but theres just too much stuff, and im not having a good hook time for when i check
Doesnt it then adds work? I buy ingredients for dinner in the store I am at. If they dont have everything, I just do different meal. Saves time and effort compared to visiting multiple stores.
Instinct contacted Oracle support for me and got movement on a ticket that had been stuck.
It's nice as like an interactive note keeper. I tell it thoughts/ideas I have and it stores them in markdown.
I'm having Muse record a local radio show (no online archives), transcribe it, and send me the transcription after the show. Hasn't ran yet, but if it actually works it'll help me keep up with local news.
This is all stuff I could do myself or do with other apps/code, but I wouldn't have time or would rather not
> use cases that have actually improved your life in some measurable way?
"actually" and "measurable" are, to pardon the Claudism, load bearing. If something made my life better in some imperceptible way that makes me happy, why the need to force metric on it? Not having to call Comcast to fake threaten to cancel service made me THREE happier and SEVEN less angsty. That pretty picture made me FOUR smilier. It makes me wonder if we can't just have nice things.
Measurement doesn’t have to be in a metric space, it can be relative. This tree is taller than that bush is a measurement. I am happier today than yesterday because of my agentic harness is a measurement. A metric space is in some ways useful but neither necessary nor superior to a relative measurement.
Everyone is bringing up ads or lock-in, but there's another reason companies (beyond Amazon) are going to be squeamish about allowing AI agents to make purchases:
The return rate for AI-purchased products is probably much higher than when a human reviews the page and makes the decision. There are going to be a lot of returns ranging from "I didn't mean to order this" to "It's not what I thought I ordered" because someone asked their agent to do something and an box showed up at their door 3 days later with something they would not have purchased if they were driving the browser.
There's a difference between automated industrial scale scraping and well-behaved agents acting on the behalf of individuals. Right now there isn't a robust, standard way to distinguish between them, so sites just block known datacenter IPs and throw out the baby with the bathwater.
That's a main advantage of running your own local claw setup using your residential connection -- difficult/impossible to block.
That said, eventually a site blocking all agents would be like blocking all search engines, something that hurts more than it helps as agentic interactions become "the norm". WebMCP or similar support will likely be a basic expectation at some point.
It's worse than that, they will block all agents and specifically make exceptions for maybe 2 search engines per country. Thereby strengthening existing players and weakening all incumbants.
Cloudflare for instance does allow indexing by Google and Bing I think by default, but does challenge other bots. Double check me I am speaking from memory.
With AI's capable of defeating every captcha short of ARC-AGI, I don't know how this cat-and-mouse game will continue. Services (now banks, later Amazon) will require attested hardware, but even that can be put in the hands of a robot.
I wish services would get over trying to block how their customers want to use (or NOT use) their websites.
How are AI's capable of defeating every captcha when I, a human, get sometimes entirely blocked by captchas all the time.
I get captchas when trying to do google searches if I'm not logged in. Or a lot of the time its "please wait a moment. Don't refresh the page" (the google results will then never load until you refresh the page). A few weeks ago I couldn't get past a cloudflare checkbox on a modern iPhone, and I'm not on a VPN or anything either. Yesterday I tried to visit an old reddit thread I had bookmarked and got "You've been blocked by network security," without a chance to even attempt a captcha.
It is unfair. I too (a human) have been caught in Cloudflare loops. Hopefully these are the dying gasps for captchas and sites will give up soon since they are becoming reverse-captchas.
I'm still annoyed that Amazon stopped sending order details in emails so that gmail can't show me when specific items are arriving. More recently 'getsauce.com' started to do the same thing, hiding order details behind an opaque link to their website. Every player in tech wants to force you to interact with their platforms purely on their own terms, to the detriment of the users.
We should blame Google for this, and not Amazon. Google added this data from your email to your advertisement profile and used it to increase their own revenue.
But I'm using Google, which is my choice (and my own fault). Why shouldn't I be allowed to let Google have that data if I find using Gmail to be beneficial to me? Amazon should be satisfied that I am purchasing goods and services through their platform without seeking to sacrifice my experience of their platform for some nebulous gain elsewhere.
> Why shouldn't I be allowed to let Google have that data if I find using Gmail to be beneficial to me?
The main issue is that this is not a conscious decision that google asks you to make. They offer a free mail service and then turn around and harvest your personal mail to increase their revenue, without most consumers realizing it.
They could've been happy with your lock-in to the google ecosystem and to google search, but they wanted to more data and systematically scoured all transactions from users' inboxes for additional data.
Like with physical mail, the digital mail has an envelope and a message. Gmail offers to handle your envelopes and assists you in viewing the message, but then went a step further and data mined all info from the messages.
In the physical world, if a company handles my inbox, it would be forbidden to open the envelope and read the message without a proper consent. The company would just open the inbox, tell me there are two letters from the bank and one letter from the customer, without reading the actual message.
I think it is very clear that the "deal" with Gmail is that they host your email and use the platform to sell you things based on the content of the email (and your Google account profile in general). I don't think this is necessarily a good thing, but I don't think Amazon is defending my best interests by stripping information from emails so that I (and thus Google) cannot see their contents. Of the two companies at play in this scenario I dislike Amazon more. I can understand why you might dislike Google more; certainly they have been working hard on similar walled garden tactics on every service and platform they control. However, in each specific such scenario I am going to come down against the company crippling their product and punishing their users, and in this case it is Amazon.
Right now I am upset at Google for what they are pulling with Android developers and app installations; and I am perpetually mad at them and every other company for abandoning true third party APIs for using their services. RIP XMPP chat. RIP the golden iteration of any walled garden product that I happened to like that big tech then killed with an update, which is impossible to functionally recreate because APIs for services no longer exist.
I'm really enjoying muse+have used it to make some amazon orders. This is a big bummer to me, in future I'll find an alternative to amazon OR find an alternative to muse.
No matter what I'll continue to use some AI agent and some shopping provider.
I’ve been purchasing more and more stuff from Walmart (good interaction with ChatGPT) and other stores.
It’s clear why Amazon is doing this from a business perspective, but this is a stop gap for them. They will need figure out a better plan for people using their agents.
Yeah I experimented with shopping and found it was really tedious jumping through the 2FA and other bot check hoops every third click. Agree it's a huge convenience if it would work though.
Agents don't click through or interact with ads or get influenced by things like lightning deals, snap promotions, "others who bought X bought Y" and such the same way a human does. This is likely a large portion of the reasoning.
This is just going to get worse before it gets better. More large sites are likely going to block or attempt to block agents because they're so reliant on ads for revenue or exploiting human psychology.
It’s funny watching all these large tech firms with so much resource commit the sin of innovation: starting with the tech and not the customer + economics of it all.
I get it. It’s tempting and alluring. But no - reimagining something takes a lot more work than sticky taping an agent on.
Impatiently waiting for all this agent stuff to fek off.
Yeah subscriptions are another big one. So many profits today are generated by bleeding off customers that just can't deal with the hoops they need to jump through to get unsubscribed. Tasking that to an agent would destroy the whole model.
Its a very good point. There are lots of orgs in company's whose whole job is to push shit from all possible points in the website - its probably called "real estate" internally.
These orgs have a lot to lose. But the company? Not so much, they will make it up and more - agents are great to use.
And yes ads are hard to place because agents let you bypass all this and give you a nice filtered view of exactly what you need.
Where do you think the world will head? Maybe Amazon will sell their inventory for agents at a hefty API price?
How do you know they are standing up for users? From where I'm sitting it looks like they just are trying to protect Alexa Shopping's revenue and/or monopoly.
Ha. Amazon's in-app AI agent is called Rufus. It's incredibly annoying, and they love directing you to an 'Ask Rufus' page where you watch an AI agent not-quite-finish repeating your search terms back at you before you realize you're not actually on your search results page, and then spend some additional time trying to find the link to take you to said search results page.
I assume somewhere there's a team hitting their AI query KPI's due to this, but I really wish they'd shut off a few server racks and save everyone the annoyance instead.
This is not them standing up for the small guy, this is them seeing their actual profit center in retail (advertisement and paid promotion) evaporating if human beings never view their listings.
It is a money grab from Amazon I'm sure. They want to make a deal with meta to get paid for access. Maybe also worried about liability I imagine on accidental unauthorized purchases?
Returns cost money and if an agent generates more returns and hitting the margins or seller retention then that's something that's likely going to get increased scrutiny.
The generally quoted return rate is 5% to 15% (and it varies by category). If Muse identified purchases are generating returns of 10% to 30%, then amazon is probably going to crack down on it. (Amazon Sellers pay from 30% to 50% for customer returns from their own pockets. https://medium.com/@myrealprofit/how-much-do-refunds-really-... )
Amazon’s enterprise value is made up entirely of AWS and Amazon’s advertising business. Thus according to Wall Street their e-commerce is a loss-leader whose only purpose is to provide inventory for ads. Agents shopping on Amazon do not show Amazon ads and are thus antithetical to Amazon’s business model and need to be blocked.
Amazon doesn't provide financial breakdowns between Ads & E-commerce fulfillment and inventory costs, so take the Wall Street article with a grain of salt.
The real big question is what happens to display advertising revenue on Amazon, if even a fraction of its traffic moves to these agents.
Amazon ads is a relatively large portion of its retail profits.
Display ads are meaningless to agents. And agentic commerce is a huge threat to Amazon’s profitability.
No wonders, Amazon is actively banning these agents. but for how long is this even viable? Do we simply move our shopping to more agent friendly destinations? Is this an opportunity for Walmart to finally reclaim it’s throne?
> Display ads are meaningless to agents. And agentic commerce is a huge threat to Amazon’s profitability.
Would agents be susceptible to advertising? If humans have issues differentiating ads from real results, and are swayed by TV ads and stuff (they are), then why wouldn't an AI agent.
I kind of assume they would, given how ChatGPT brings up garbage websites sometimes for my questions.
This is inevitably going to get looked at by antitrust regulators, either in this administration (depending on how much Trump likes the leaders involved), or in the next.
Shopify seems to have no issues partnering with Muse
>We are excited to announce we are partnering deeply with Muse to enable agentic checkout with Shop Pay on all Shopify stores, offering people an easy and delightful way to shop and check out with Muse.
There's a lot to discuss here - agentic commerce, IMO, is where things are heading. I'm completely convinced that people _want_ to shop using agents. But I want to understand the dynamics at play here: will underlying inventory providers like Amazon give up control and allow agents to use their website?
I find this question very interesting and I want to predict where things will head. On one hand Amazon has a lot to lose by giving up control on how their website is used. Ideally they'd like to control the full end to end experience. What happens to the work they put in their UI? All those endless A/B experiments to make the buttons look perfect?
But on the other hand - this is where things are heading. They can't stop agents from using their website. Not for long. And they have a lot to gain because overall purchases will increase since. They can accept it and create a plugin and try to offer as much "branding" as possible inside it. Or they can completely go against agents and block it.
Then you have to wonder, what really is the moat here? Amazon's logistics and its relations with partners and customers (return policy etc) and the trust it brings. The UI part dissolves and is no longer important.
Maybe they want only an allowed list of agents access to Amazon? Maybe that's where things are heading? They sign deals with maybe just OpenAI (for example)? Its not like Amazon makes their own agent + model. I would hate it things fragment this much.
In theory, legislation could help. If there were a pro-consumer movement in government.
Something like declaring agents as representatives of consumers and blocking them violates ADA or just shouldn’t be allowed unless they are abusive.
Lacking that, it seems it’ll be an arms race between offense and defense. At some point, computer control probably wins (running my browser, on my computer, is pretty difficult to detect with reasonable f1 score)
Anything is possible I suppose... I don't see a path that leads to Amazon actually further controlling the shopping experience inside a browser (where a local agent might have control that they can't reliably detect), short of forcing users into some kind of hardware-based attestation solution or something, but even that is not foolproof.
> On one hand Amazon has a lot to lose by giving up control
The everyday person is told to get over it and accept that AI is here so adapt or die. Companies start reducing headcounts to be replaced with AI. So now, a company is having to accept that people are using AI in ways they don't like so they should adapt or die as well too, right? I have no tears to shed if Amazon loses control. Also, it'll be one billionaire battling another billionaire at that point, and I'm kind of okay with that chaos. My world is no longer in my control so fuck 'em is my stance
It’s an optimistic view on AI - brings power to the people by mediating all the garbage.
I’ve already used it to handle customer support request calls. I see 30+ minute hold+conversation times and generally the issue gets resolved. That is hugely empowering.
OK according to game theory, if this is correct, why even put up a fight? You have nothing to gain by fighting. Just accept it and play along. Make your damn MCP. You have 100,000 engineers. Why can't connect Amazon to ChatGPT today? I don't know.
> Why can't connect Amazon to ChatGPT today? I don't know.
They just don't want to, because there's no ads. The human gets removed from the loop of visiting the site entirely. And it doesn't say Amazon all over it.
If the user is not browsing the site/app, Amazon cannot collect the analytics about your browsing habits. The AI system does though. There is absolutely no incentive for Amazon to discourage users from using their system. After all, the AI system might not show the user an Amazon item.
Amazon cant hold this off for the long run, nobody wants to scroll through a list of two hundred near identical packages of toilet paper. Agents will intermediate, and they will remove demand for low value items
This is already solved by the existence of hundreds of other stores that actually do some level of filtering and curation of products. I don't know why we want to play some cat and mouse game of getting an LLM to find the best toilet paper on Amazon versus just finding a reliable source of toilet paper to begin with.
Theh will probably consolidate more. Less choices even if they are worse is also easier for llms to do. Just look at how they pick technologies to use. Almost always the same for a given model.
I would hate it if every company comes up with their own agent and only their agent is allowed to access the API. EU do your thing instead of cookie banners plz.
that's what they're going to do, trying to get $20 out of the average person is terrible compared to giving away your product for free and getting an endless supply of metadata
The right way to do this is to create an API or MCP or whatever catering to bots, and implement corresponding terms in your legal agreements (i.e. after approving a bot to interact with us, you're responsible for purchases it makes), along with some limits to prevent runaway shopping when Muse decides to binge.
Less friction = more shopping = more profits.
That used to be a concept companies like Amazon understood well. (As well as companies like Google in the context of web browsing and keeping it user-friendly as opposed to user-hostile).
If the bots mean customers are getting a better deal elsewhere, then you're doing something wrong elsewhere in your business model and need to go address that. Or eventually free market economics will, no matter how big an artificial moat you try to build.
> The right way to do this is to create an API or MCP or whatever catering to bots, and implement corresponding terms in your legal agreements (i.e. after approving a bot to interact with us, you're responsible for purchases it makes), along with some limits to prevent runaway shopping when Muse decides to binge.
From Amazon's perspective, that would be very much the wrong way.
Amazon's concerns here aren't technical. It's entirely about being disintermediated.
But how can they serve product suggestions this way?
Right. Agents represent a direct attack on a significant portion Amazon's revenue model.
Right, agents work on behalf of the user and only cost compute (not time/attention, which is the ultimate human currency). They act on behalf of me not Amazon.
It is my personally optimistic perspective on AI that it will mediate all interactions with 3P who are acting in their own interest and not mine. Ie, “my” agent.
maybe agentic commerce will finally realize the micro-payments based internet that crypto people had been wishing would happen for years? Should there be stablecoin based micropayments built into MCP to charge for agentic interactions to replace the micropayment of an ad into someones eyeballs?
This is (still) never going to happen. A more realistic outcome is a deal between Meta and Amazon to enable access in exchange for invisibly integrating advertisements into the agents' decisions.I don't know exactly what that will look like, but I'd bet we will see a version of it soon.
It can only be this. And it will fail the same reason Alexa failed.
Nobody will blindly trust Amazon to buy products when they know Amazon cannot and will not truly suggest the best product for the consumer.
Alexa was supposed to be the voice gateway to purchasing. But nobody trusts the result Alexa picks because it will not default to the best value to the user - only the best value to Amazon.
very interesting, and im inclined to agree with you. Depends how centralized the agent economy is I suppose where bespoke deals are worth it.
This.
Recently it was estimated that Amazon makes $30-80 billion(!) per year off not showing you the economically best results by deliberately disallowing sorting on $ per unit [1] to force sellers into ads. Amongst other tactics (ex. filter by country of origin).
AI agents can sort on $ / unit to easily find you the best deal, rendering Amazon's various forms of paid placement irrelevant.
Extend this to filtering Chinese slop brands (also not able to be filtered) and the significant portion of revenue starts looking like knocking back margins to the 3-5% range. Basically an existential threat.
Amazon cannot allow AI agents and maintain current margins.
[1] - https://pluralistic.net/2026/09/04/cheating-at-fraud/
The AIs are also providing the product suggestions, and it comes from what data they are given from their searches. I think they can put 2 and 2 together on how to make that happen with the right MCP response bodies.
If Amazon doesn't want to offer agent-first shopping I'm sure someone else will come along and eat their lunch. Adapt or die.
API or MCP that does context injection and forbids stripping it in ToS.
I asked this in another thread: who benefits from all these bot integrations?
It's not the consumers of the AI, or the consumers of the real businesses? It just seems like everyone things "if they build it" it'll come.
But it's clear Amazon cares about how it's costomers have to view all the items as both upsells and advertising, and a bot is going to either be _really_ into it or act much like ublockorigin does.
But either way, no company making money on costumer direct sales sees a real benefit to allowing a bot to easily bypass all the necessary details.
Whether it's advertisement, liabilities, manufacturer's recommendations; a bot will ignore these things and generate a costumer expectation that suddenly the producer has to live up to.
Then you'll tell me: YEAH, that's why you need a MCP friendly website.
And I'll tell you: that's how the mafia works.
I’m actually making this:
https://marketplace.openship.org
It’s an interoperable marketplace that uses MCP-UI to connect to seller websites directly. It brings the storefront and cart directly into the chat.
Nice. That would allow people to create alternative interfaces for humans (like me) who absolutely HATE Amazon's web interface.
They will absolutely put something in to prevent you from doing that, too much of their profit relies on the experience you hate.
That experience actively drives me towards other retailers.
I want to buy what I want to buy, and I will never buy things that I do not want to buy. If the time ever comes that I am tricked into buying the wrong thing from Amazon, then I'm sending it back on their dime.
If that means that only 3 things fit, then I only want to see exactly 3 things. If that means that 0 things fit, then 0 is also a valid result. In response to these results, I can then elect to soften my constraints at my own volition.
But an Amazon search typically results endless pages of vaguely-related products that represent a jumbled mess of things largely do not fit or only-sorta fit the broad concept of what I intend to purchase.
When I want a 4300 Kelvin D2S HID headlight bulb for my car, then that's all that I want to find when I search for it. I do not want results for LEDs. I do not want selective yellow bulbs. I do not want 5000k or 6500k or 10000k. I do not want an "upgrade" kit, nor D2R bulbs, nor D3S. None of those things are useful to me.
The list of things that I do not want is infinite, and when Amazon presents a seemingly-infinite list of them then that represents the opposite of usefulness.
I only want to peruse listings that match. I require no more and no less than this in order to be satisfied. There is no possible amount of additional noise that can ever dissuade me from seeking the thing that I very specifically desire.
A modern LLM can weed out and separate the bad results from the good ones and give me an easier time of making a purchase decision, but it can only do so when it is not blocked from the start.
> that's all that I want to find when I search
That ship has long since sailed for every search product. Companies have long-since learned that intent-based results are much more profitable (why show an empty result page when you can show something for people to buy or click on?), but it is also often genuinely useful when your search is not perfect or what you want doesn't actually exist. I agree that it is annoying when you really truly do only want something very specific and the ranking prioritizes the wrong things (e.g., sponsored results, more profitable items, etc.) over exact matches, but can you honestly and truly say you've never found any value from intent search algorithms?
I very often truly only want something very specific. I'm not apologetic about that at all.
When I want oolong tea bags and I search for "oolong tea bags," then I definitely don't want to be shown bulk green tea instead.
If I want to find any kind of tea instead of specifically oolong -- if, perchance, I wish to peruse the broader market of available tea -- then I'm very capable of removing "oolong" and/or "bags" from my search prompt.
---
I can look around and observe that the ship has sailed, but that doesn't mean I'll ever be happy with the outcome or accepting of any notion that the situation is anything other than hostile.
Can you recognize that it gets between me and my willingness to spend money?
You mean if they falsify the search for both sides they make more money, as ironically spelt out in Google's original paper "The Anatomy of a Large-Scale Hypertextual Web Search Engine":
“We expect that advertising funded search engines will be inherently biased towards the advertisers and away from the needs of the consumers.”
“advertising income often provides an incentive to provide poor quality search results.”
“the better the search engine is, the fewer advertisements will be needed for the consumer to find what they want. This of course erodes the advertising supported business model...”
https://snap.stanford.edu/class/cs224w-readings/Brin98Anatom...
Note: they even explain why advertising funded search makes things worse for customers and for advertisers. Worse search finds more ads and thus more dollars, resulting in both worsening search results over time AND less value for ad money over time.
Amazon is only an interface to purchase products made by other companies (except for Kindle devices).
Your alternative interfaces exist and have always existed: Go to a physical store, order by phone, order through the OEM website, order through an Amazon competitor, etc.
"you can already build [Dropbox] yourself quite trivially by getting an FTP account, mounting it locally with curlftpfs, and then using SVN or CVS on the mounted filesystem."
For all the complaints, it's nice to see HN hasn't changed that much.
Yes, you're right. It's almost impossible to purchase from a physical store or from competitors to Amazon (who many times are much bigger than Amazon in their country).
On the next episode of Oppressed Hackers: We get a load of a guy who had the nerve to say that it was possible to drink a glass of water.
The point wasn't that there are no alternatives, but that calling Amazon "only an interface to purchase products made by other companies" is reductive to the point of being completely wrong.
It makes total sense for Amazon to block agents. It sucks having to manually check multiple sites for prices and delivery of something you want so is easier just to use Amazon which is usually very good. If people start using agents Amazon is less likely to get the sale and are more vulnerable to competition. Better to kill this trend off asap.
Sounds like Amazon is trying to use its monopoly position to give it an competitive advantage over smaller stores.
Monopoly is a word that has a real meaning, and not a blanket term for companies you personally don't like and not even a blanket term for big companies.
Amazon has nothing even approaching a monopoly. You can go right now to a store built from bricks and purchase goods.
> Monopoly is a word that has a real meaning
Correct
> Amazon has nothing even approaching a monopoly
Sortof? It probably doesn’t have a monopoly in any segment that matters, but it does have many illegal monopolies (as defined by US FCR and the Sherman Act).
Evidence: Amazon has been found guilty of being an illegal monopoly multiple times already.
I don't think it does make sense. I think Amazon is way overestimating brand loyalty, and underestimating how much of their success comes from being a one-stop shop for everything.
I'm using Claude for everything from finances to dinner planning to project planning. If Amazon doesn't let Claude search and buy, I'm not abandoning Claude, I'm abandoning Amazon.
As long as my actions are as simple as possible, I don't care whether Claude is searching 5000 online retailers or only 4999, minus Amazon. I'm certainly not going to say "ok, this project requires buying 13 things for about $500, but I don't see amazon in the list, so let me go manually search for each one on Amazon and see if they might have a better price or something."
This is Walmart / Apple Pay all over again. Yes, near-monopolies, lots of power. But they got there by making things easy for consumers. Making things difficult to force consumer behavior never ends well.
Most people don't have the disposable income to ever let an agent casually make a purchase for them.
I never let agents make purchases unsupervised. But that wasn't the point; the point is that the convenience of getting a complete list including multiple-vendors when pricing makes sense is way better than going to amazon and searching one by one, therefore Amazon will lose this power play.
The same would be true for someone shopping for one $10 item.
I always check Google before Amazon, because it's impossible to find anything on Amazon. Amazon is a bit like the bargain bin of a DYI shop. If it's immediately visible it's okay, of not, don't bother searching.
Half the time I can't even find the category page. And if I do find it, the filters will just be useless anyway.
In general, preventing consumers from having fully featured AIs (especially local ones) will be worth about 10 trillion dollars to companies. That's my prediction.
Amazon is one example. If my basement AI drives my browser and never clicks an ad, Amazon loses big.
Or, the ad industry in general -- imagine the basement AI filtering out all ads at the analog level so nobody in my house ever sees an ad, and no software update can change that.
Or other optimizations. Right now companies enjoy a power imbalance because they have people who ruthlessly optimize 24/7 to save a buck. I can't do the same, because I have to work, and also prefer to spend my personal time with family -- a foolish thing to do in a capitalist society, at least from a capital perspective. Imagine I have an AI wheeling and dealing on my behalf 24/7 too. Especially in an age of online shopping and product deliveries, my AI will absolutely order from your competitor to save a penny, my AI will absolutely spend 400 hours disputing a medical bill I don't think is right, barely an inconvenience to me.
One of the optimistic aspects of what's happening is that it doesn't seem like the leading model capabilities are outpacing the more open ones by that much- that just as a general principal there are pretty ok models relatively close behind the best ones.
This could mean the long term viability of my open claw bot to shop for me- that openai / google etc won't control AI, and I'll still just be able to use a model via API to do what's actually in my own self-interest.
The ads (in general) will be redirected to where you will see it, regardless of whether or not Amazon allows bots. You don't think you're getting off that easily, do you?
Once AI is good enough the AI will control what I see, plus, I will just spend a lot less time looking at screens altogether.
It's also trivial for Amazon competitors to pop up in this potential AI future. Maybe you only sell door knobs, but all you need to do is set up a mostly text-based API designed for AI consumption and you're in business. If you offer the door knob for 10 cents less, why wouldn't the AI agents buy from you?
Right now making an Amazon competitor will fail mainly because humans are lazy and can't be bother to check other sites. AI might change that.
> You don't think you're getting off that easily, do you?
Agreed. But that's where Amazon needs some thrashing from users, competitors and regulators to start with.
Amazon is not usually very good anymore, quality has suffered massively.
I still prefer the Amazon shopping experience to anything else, unless I already know the product is available locally. The primary quality problem I've run into is the quality of the products themselves.
Should not stores make shopping with AI easier instead, as the number of ”customers” that are agents will increase and they will shop where they can?
Amazon is supposed to be the cheapest/best delivery store due to it's huge scale.
It's how they killed the competition.
Why is it even Meta's job to announce that an AI agent is being used? Is there a law that actually forces them to make that known?
Otherwise people could just use local agents on their own machine controlling their existing browser/logins as they likely already do, without issues.
That's irrelevant from Amazon's perspective. If an agent is popular enough and has some fingerprint they can identify, they have a strong incentive to block it.
Agents that scrape and digest web data are entirely anathema to brand-reliant and advertising-reliant vendors and marketplaces. Even human assistants could be trusted to develop brand attachment and respond to advertising meant to shape their preferences. And while LLM agents are still manipulable, that's a frontier that nobody really knows how to navigate effectively yet, at least not to the degree of optimization modern vendors and marketplaces (especially Amazon) have acheived.
It's quaint to imagine what AI agents might do for consumers in the naive internet that preceded their emergence (if you ignore their vulnerabilities and flaws), but that's not what they'll be able to achieve as that same internet responds to them -- at least not at Meta scale. It's not very different than how social media companies once offered, and then aggressively withdrew and policed, API's to their graphs and feeds. Those open API's simply aren't compatible with the business objectives that keep these companies running, no matter what cool and profoundly user-friendly things could be built on them.
The same kind of incompatibility applies to AI agents when it comes to e-commerce. It won't be a permitted thing until, and except for where, vendors/marketplaces can assert control over the agents themselves and manipulate them to ends comparable to what they can achieve with human consumer sentiment already.
What's crazy is that the ToS is basically, "allow us, Amazon, to give you algorithmically higher prices- and if you try to circumvent that, you can't buy from us anymore".
I think the arguments for these kinds of near-monopoly companies being true market-driven capitalism seem more and more false.
The idea that Amazon will tell you that you don't even own the right to your own shopping experience tells you everything you need to know. Apple owns your device and Amazon owns the right to show you whatever price/product they want, regarless of your search intent.
I sometimes look around at people proposing technical solutions for these political/legal/economic problems, and think: "Was I that naive X decades ago, or were things really different then?"
And let's not forget that the browser is already considered to be a user agent.
I guess it’s more that Amazon’s server queries for an answer and intentionally obfuscating a bad one comes across as lying.
What do you mean? Declaring the agent is transparency. You should be congratulating Meta for their rare good deed.
I don't care about Meta even one bit so more power to Amazon. At the same time, Amazon is going downhill at an alarming rate. Is anyone else annoyed by the no name brands flooding every product listing there? There is no way to even enunciate the freaking names! How am I supposed to trust them?
The other issue is overpromising delivery times and vastly underperforming once the order is placed. The worst is when I purchase something that I want urgently. Let's say the price is $23. Amazon suggests buying something cheap to make it $25 and get overnight delivery. So I add some Chobani product from Amazon Fresh. The next morning I wake up to see Chobani delivered while the item I want is delayed until the next day. This has happened 3-4 times now so I have become wiser. But what a crappy customer experience from Amazon!
Do Amazon Fresh purchases count towards the $25? I've noticed that only specific prime only thinks count towards it most of the time.
It's possible that it's location specific. We have a Whole Foods nearby so Amazon invariably shows me Amazon Fresh products. I remember once I even added a lime to the cart because my total was $24.99 and it wanted me to make it $25.
Amazon is basically a frontend for AliExpress, with some US products thrown in to keep people clicking.
I ask this in every thread on muse / OpenClaw / whatever platform. I’ve never gotten a clear answer beyond “yeah me too”:
Does anyone have any meaningful use cases that have actually improved your life in some measurable way?
I’ve tried hard and beyond filling out some tedious forms in about the same time and energy as just doing it myself, I’ve not personally. I’ve tried hard. So I figure it’s probably my lack of creativity given the prevalence of fans. But every time I ask, I get no meaningful responses. So it makes me wonder if all the fans are just product managers on those products?
In the past few days:
- Dusted off a 3d printer that wasn't being used enough, largely because of the tax of setting up profiles for materials and each print. That was fun once, I'm too busy now. Was able to get claude to use MCP and configure the slicer just right for the job and filaments.
- Needed to get IPTV/etc set up for football season. If you've done this you know what a pain it is, how EPG's don't always work right, finding the 50 channels you care about out of 80k channels is a nightmare, etc. Pointed claude at my utility server, it got Dispatcharr configured, wrote scripts to parse channel guides from sources not covered in the EPG, scheduled to update daily, deployed.
- I'm considering getting back to 3D printed clay, which is fun but challenging. Based on a thread here on HN, had claude go investigate using FPGAs (which it would program) for a realtime flow modulation system that would remove the variability from slight changes in viscosity within a batch of clay. Not sure I'll pull the trigger but I have a much deeper understanding of the solution.
- Realized I was paying like $30/mo to RocketMoney and Copilot (finance, not MS), for rare usage and basically just occasional "find subscriptions I forgot about" usage. Got claude connected to plaid via MCP, found more than those other two ever did, got it to set up cron to send me email weekly with current subscriptions and any price changes.
That's literally in the past 4-5 days. I find great value there, YMMV of course.
I do all these things but use codex, which lets me use git for settings etc and I end up building software I can re-use. Am I wrong?
Why wrong? Seems reasonable.
> Got claude connected to plaid via MCP
I would like to check this out. My expenses are getting out of hand.
Well, with EXTREME cautions (read the readme): https://github.com/brookstalley/bankmachine
Thanks! I will check it out
I find it useful, so here are some concrete examples from last few days.
As a baseline which many skeptics will have a hard time accepting for their own usage - I think there is a correlation between utility and trust (even which is not truly earned). If you don't let it capture memories and learn about you, and you don't connect it to external tools like email, it's less useful. But I think of it usually like a ChatGPT w/ extra tools and knowledge.
I just moved, and last night I asked my agent how to test how safe the soil in my (urban) garden was, and it found some warnings from my city government, and suggested a particular test service that is available in my area. Since it knows exactly where I live, it was able to find recent warnings of contamination in my neighborhood, and suggest specific chemicals to test for.
Additionally, I give it (read only) access to my email, and it'll ~2x a day notify me of some important email. I'm working with a contractor, and getting emails from them bumped, or getting emails about AWS usage alerts pulled out of inbox and notify me directly is super helpful for filtering the noise of my crazy email volume of 99% spam.
Also, as others have said, it's easy to take a picture of a flier, and say "add to my calendar".
At home? Not really, no.
At work? I've got ADHD, so the idea of making a to-do list is fine but keeping it up to date was a challenge. I've created a set of agentic skills that comes through the last 36 hours of emails, meetings, chats, and gathers up all the work I volunteered to do (or was told to do), triages them across urgency/impact, and sorts them into Do/Defer/Schedule/Delegate piles.
Last, I created an app to give a UI to this data to see and act on it all, including a way to flag a to-do that was found and enter why I thought the status or categorization was wrong. On the next run, a review skill reads those comments, makes changes to the other skills accordingly, and sends me an email with details of the changes, and the commit I can use to roll back the change if I'd like.
I've been running it since April and while there were a lot of updates at first, I haven't had to flag anything for review in months, and I get to feel a whole lot better about managing my to-do pile.
I'm looking at open-sourcing it, but since I created it with work resources on work time, I need to jump through a few hoops first.
For myself, it's been super helpful in at least two ways:
1) Staying on top of my kid's school obligations, homework, etc. School Learning Management Systems are a hot mess, and I use agents to consolidate and collate, to keep track of the homework that has been done, what needs to be done, and even pre-grade homework to give feedback (most of the work is done on digital platforms too). It's been immensely time saving for these kinds of things. 2) Multiple instances of watching things passively: hotel room prices, availability of tickets, finding the best place to sell things online. In general things that require a lot of browsing, collating, and watching. These are the kinds of things that specialized software products could do as well, but with an agent they are just a prompt away, and you get exactly what you want. (More or less.)
As someone who doesn't have kids and was last in school decades ago....
Do those School Learning Management Systems tell you everyday what homework your kid is expected to complete? Do you feel such a granular degree of monitoring and direction has impacted your child's opportunity to build character, intrinsic responsibility and self-management skills?
When I was a kid, teachers gave out assignments, and the student was responsible for keeping track of them, staying on top of progress, and handing in by the deadline. If a problem arose that wasn't an isolated incident or resolvable directly between the student and teacher, the feedback would flow at a much higher level e.g. scheduled parent-teacher interviews or via a phone call.
And what do parents do who don't accept the terms and conditions of said educational platforms? (e.g. due to overly broad privacy clauses or one-sided indemnities and the like)
But nowadays kids are expected to check 2-4 platforms plus their notes and there is no guaranteed time when they know it will be there. It is like the worlds worst paid disfunctional work place and one of my kids schools literally uses teams for one third of their assignments. In word format so sometimes kids overwrite it.
If you don’t accept the terms your kids life is just even harder.
Keeping track of your kids' homework is your concern? Doesn't it only become your concern, briefly, when the school contacts you about your kid not getting their homework done?
As a parent people from prior generations do not get the current situation. I actually agree that this is a useful use of the agentic work tools - and it’s the only case I mentioned above I’ve used it for successfully, if not saving of time.
Homework isn’t assigned in class any more. It’s posted on dozens of inscrutable websites in weird UX’s that are impossible for any single human to understand. Children are mystified, parents confused, teachers at their wits end.
A few times a week children miss some website or queue from some system, teachers can’t get things to share properly, this needs tending, they need to photograph this and upload that to this place as well as that but not the other one.
It’s truly remarkably astounding. I have serious doubts frontier models can navigate it either honestly. But this is in fact where I see the most chance of using these tools.
“Computers at school” is the worst idea ever, and no one who hasn’t experienced the clusterfuck of middle school in 2026 can understand it. It makes the world of the movie Brazil seem utopian.
Btw I did this Muse, giving it the ability to use its cred vault to log into her school information system and read through all the crap and other websites and collate the homework’s and their due dates, adding to my child’s calendar when they’re due and a reminder the day before with a link to the calendar. I had model the workflow formally (I do this with all LLMs for procedures) and set it up to run daily. We shall see!
Overall it was not bad. The cred vault is useful. But I’m not giving meta any information that I wouldn’t otherwise be fine with disseminating and I feel the burden the school puts on us vs the information loss is worth the trade.
Thanks! First meaningful use of an OpenClaw like system.
I'm not the person you're asking, but in general, I would also stay on top of my child's obligations - homework, chores, etc. I don't want the first sign of trouble to be a call from a teacher, or the complete absence of clean clothes for them to wear in the morning.
You don't teach children responsibility by removing yourself completely from that equation
You equally don't teach them responsibility by always being there as the backstop though. I think missing a couple of homeworks is quite a low-stakes thing that can teach them a lot about responsibility, failure, and how to recover from it.
A lot of kids now seem to go into adulthood having never been allowed to fail, and it can really mess them up the first time it inevitably happens.
Oh, sure. I'm not saying I'd remind her about every piece of homework. But I want to know if she's starting to not do it.
There is quite a large gap between checking in with your kid and having an AI analyse their homework and assignments and raising an alarm if anything appears to slip. The latter is a solid WHUP-WHUP-WHUP-WHUP-WHUP on the helicopter parent spectrum.
Missing homework once or twice is part of life and a really good life lesson with low risk besides.
- find piece to buy to repair an appliance
- travel research for places where I need an apartment with requirements one cannot just filter for
- gardening tips based on photos of my plants
- financial planning (similar tips to what I would get for more money, without paying per hour until I know much more)
- take photo of all physical spam and have me removed from the list
and many more…
If the access to this product went away I will survive, but it will be annoying and waste my time.
But aren’t all of these currently fully achievable with normal chat these days?
Travel research is so much better with an asynchronous agent than with normal LLM chat. You can just let it hack away for an hour looking over options and browsing different websites.
I'm not using a platform but I set up a telegram bot that is backed by AI (DeepSeek flash tokens are almost too cheap to meter) that handles my strength training workouts. It reminds me of the days workout, takes my feedback on each exercise difficulty and adjusts it for the next round, handles injury rehab prescriptions, and so on.
I'm now considering putting a coding harness inside it so I can actually ask it to improve itself on the fly, since I realized that having to switch to my laptop just to give a short feedback is a bit silly, why not right through the bot?
It's working well so far and helps with my tendency to forget my workout program for days at a time since it pings me daily.
I agree with you. I also wouldn't trust it calling on my behalf or doing the majority of what people say it can do. Putting aside it making mistakes and not even being beneficial, I definitely wouldn't want meta of all companies to have the relevant information to do those things.
Just in the last three hours, Muse helped me make progress on my go-to-market plan, drafted me a script for my demo video, and went and listed 50 high value creators in my target market that i can go interact with and ask for product feedback. It could even DM them for me if I wanted to. It's the first time a product like this has actually shocked me... it closes the gap between decision and action. I'm hooked.
I still have to do the work, but everything's all teed up for me.
"It even registered a new HN account today to impersonate a human and post this ad!"
I use Claude Code often for entirely non-code things: the harness concept is super helpful for going through documents etc.
I think these Claw-type harnesses are basically this for somewhat less technical folks. If eg Muse has been able to make this work smoothly from your phone for completely non-techy people, then that seems like an incredible product.
I use coding agents to fill out dunkin donuts surveys for me
i just have the claude scheduled job kinda tool running, but at the grocery store i can spot something and think of a meal, then ask claude if any nearby grocery stores have other key elements to the meal on sale.
its pretty handy.
i was trying to get a nice setup for finding what shows are happening around me, but theres just too much stuff, and im not having a good hook time for when i check
Doesnt it then adds work? I buy ingredients for dinner in the store I am at. If they dont have everything, I just do different meal. Saves time and effort compared to visiting multiple stores.
I just used it today to find and reserve a parking spot in SF.
Previously, I used Claude to gather AWS invoices for reimbursement purposes.
Instinct contacted Oracle support for me and got movement on a ticket that had been stuck.
It's nice as like an interactive note keeper. I tell it thoughts/ideas I have and it stores them in markdown.
I'm having Muse record a local radio show (no online archives), transcribe it, and send me the transcription after the show. Hasn't ran yet, but if it actually works it'll help me keep up with local news.
This is all stuff I could do myself or do with other apps/code, but I wouldn't have time or would rather not
> use cases that have actually improved your life in some measurable way?
"actually" and "measurable" are, to pardon the Claudism, load bearing. If something made my life better in some imperceptible way that makes me happy, why the need to force metric on it? Not having to call Comcast to fake threaten to cancel service made me THREE happier and SEVEN less angsty. That pretty picture made me FOUR smilier. It makes me wonder if we can't just have nice things.
Measurement doesn’t have to be in a metric space, it can be relative. This tree is taller than that bush is a measurement. I am happier today than yesterday because of my agentic harness is a measurement. A metric space is in some ways useful but neither necessary nor superior to a relative measurement.
>in some imperceptible way
How do you know it has or has not improved your situation if imperceptible?
I made appointment with my doctor and haircut all using chatgpt. Also bought some stuff on a local website.
I hate having to work through random UI. I just wanna get stuff done.
Everyone is bringing up ads or lock-in, but there's another reason companies (beyond Amazon) are going to be squeamish about allowing AI agents to make purchases:
The return rate for AI-purchased products is probably much higher than when a human reviews the page and makes the decision. There are going to be a lot of returns ranging from "I didn't mean to order this" to "It's not what I thought I ordered" because someone asked their agent to do something and an box showed up at their door 3 days later with something they would not have purchased if they were driving the browser.
There's a difference between automated industrial scale scraping and well-behaved agents acting on the behalf of individuals. Right now there isn't a robust, standard way to distinguish between them, so sites just block known datacenter IPs and throw out the baby with the bathwater.
That's a main advantage of running your own local claw setup using your residential connection -- difficult/impossible to block.
That said, eventually a site blocking all agents would be like blocking all search engines, something that hurts more than it helps as agentic interactions become "the norm". WebMCP or similar support will likely be a basic expectation at some point.
It's worse than that, they will block all agents and specifically make exceptions for maybe 2 search engines per country. Thereby strengthening existing players and weakening all incumbants.
Cloudflare for instance does allow indexing by Google and Bing I think by default, but does challenge other bots. Double check me I am speaking from memory.
> Right now there isn't a robust, standard way to distinguish between them
There is, it’s called an API
> That's a main advantage of running your own local claw setup using your residential connection -- difficult/impossible to block.
https://aws.amazon.com/waf/features/bot-control/
With AI's capable of defeating every captcha short of ARC-AGI, I don't know how this cat-and-mouse game will continue. Services (now banks, later Amazon) will require attested hardware, but even that can be put in the hands of a robot.
I wish services would get over trying to block how their customers want to use (or NOT use) their websites.
How are AI's capable of defeating every captcha when I, a human, get sometimes entirely blocked by captchas all the time.
I get captchas when trying to do google searches if I'm not logged in. Or a lot of the time its "please wait a moment. Don't refresh the page" (the google results will then never load until you refresh the page). A few weeks ago I couldn't get past a cloudflare checkbox on a modern iPhone, and I'm not on a VPN or anything either. Yesterday I tried to visit an old reddit thread I had bookmarked and got "You've been blocked by network security," without a chance to even attempt a captcha.
What are bots doing to get past all of these?
Check your network for unwanted traffic. You could have a residential proxy running on some device. Including almost any LG TV.
https://krebsonsecurity.com/2026/07/lg-to-ban-residential-pr...
It is unfair. I too (a human) have been caught in Cloudflare loops. Hopefully these are the dying gasps for captchas and sites will give up soon since they are becoming reverse-captchas.
I'm still annoyed that Amazon stopped sending order details in emails so that gmail can't show me when specific items are arriving. More recently 'getsauce.com' started to do the same thing, hiding order details behind an opaque link to their website. Every player in tech wants to force you to interact with their platforms purely on their own terms, to the detriment of the users.
Agreed, that really sucks.
We should blame Google for this, and not Amazon. Google added this data from your email to your advertisement profile and used it to increase their own revenue.
But I'm using Google, which is my choice (and my own fault). Why shouldn't I be allowed to let Google have that data if I find using Gmail to be beneficial to me? Amazon should be satisfied that I am purchasing goods and services through their platform without seeking to sacrifice my experience of their platform for some nebulous gain elsewhere.
> Why shouldn't I be allowed to let Google have that data if I find using Gmail to be beneficial to me?
The main issue is that this is not a conscious decision that google asks you to make. They offer a free mail service and then turn around and harvest your personal mail to increase their revenue, without most consumers realizing it.
They could've been happy with your lock-in to the google ecosystem and to google search, but they wanted to more data and systematically scoured all transactions from users' inboxes for additional data.
Like with physical mail, the digital mail has an envelope and a message. Gmail offers to handle your envelopes and assists you in viewing the message, but then went a step further and data mined all info from the messages.
In the physical world, if a company handles my inbox, it would be forbidden to open the envelope and read the message without a proper consent. The company would just open the inbox, tell me there are two letters from the bank and one letter from the customer, without reading the actual message.
I think it is very clear that the "deal" with Gmail is that they host your email and use the platform to sell you things based on the content of the email (and your Google account profile in general). I don't think this is necessarily a good thing, but I don't think Amazon is defending my best interests by stripping information from emails so that I (and thus Google) cannot see their contents. Of the two companies at play in this scenario I dislike Amazon more. I can understand why you might dislike Google more; certainly they have been working hard on similar walled garden tactics on every service and platform they control. However, in each specific such scenario I am going to come down against the company crippling their product and punishing their users, and in this case it is Amazon.
Right now I am upset at Google for what they are pulling with Android developers and app installations; and I am perpetually mad at them and every other company for abandoning true third party APIs for using their services. RIP XMPP chat. RIP the golden iteration of any walled garden product that I happened to like that big tech then killed with an update, which is impossible to functionally recreate because APIs for services no longer exist.
It’s not Amazon’s place to police this at all.
I'm really enjoying muse+have used it to make some amazon orders. This is a big bummer to me, in future I'll find an alternative to amazon OR find an alternative to muse.
No matter what I'll continue to use some AI agent and some shopping provider.
I’ve been purchasing more and more stuff from Walmart (good interaction with ChatGPT) and other stores.
It’s clear why Amazon is doing this from a business perspective, but this is a stop gap for them. They will need figure out a better plan for people using their agents.
Yeah I experimented with shopping and found it was really tedious jumping through the 2FA and other bot check hoops every third click. Agree it's a huge convenience if it would work though.
Agentic commerce will never work for the same reason I can't build a Twitter client. Technically it's possible but the incentives are all wrong.
Agents don't click through or interact with ads or get influenced by things like lightning deals, snap promotions, "others who bought X bought Y" and such the same way a human does. This is likely a large portion of the reasoning.
This is just going to get worse before it gets better. More large sites are likely going to block or attempt to block agents because they're so reliant on ads for revenue or exploiting human psychology.
It’s funny watching all these large tech firms with so much resource commit the sin of innovation: starting with the tech and not the customer + economics of it all.
I get it. It’s tempting and alluring. But no - reimagining something takes a lot more work than sticky taping an agent on. Impatiently waiting for all this agent stuff to fek off.
its not going anywhere. Agents are the future.
Yeah subscriptions are another big one. So many profits today are generated by bleeding off customers that just can't deal with the hoops they need to jump through to get unsubscribed. Tasking that to an agent would destroy the whole model.
You’re deluded if you think it’s as an easy as ‘send agent task, hey Presto - done’.
Of course they will be influencable. It is already possible to poison LLM output. It will happen with agents too.
Its a very good point. There are lots of orgs in company's whose whole job is to push shit from all possible points in the website - its probably called "real estate" internally.
These orgs have a lot to lose. But the company? Not so much, they will make it up and more - agents are great to use.
And yes ads are hard to place because agents let you bypass all this and give you a nice filtered view of exactly what you need.
Where do you think the world will head? Maybe Amazon will sell their inventory for agents at a hefty API price?
I don’t get it.
Codex has full computer use capability. Surely such bans mean nothing even in the very near term.
Thank you Amazon for standing up for everyman Rufus. We don't need AI agent from evil ad giants.
How do you know they are standing up for users? From where I'm sitting it looks like they just are trying to protect Alexa Shopping's revenue and/or monopoly.
I assume it was sarcasm, Rufus is an Amazon AI product.
And there go my Bill & Ted's Excellent Adventure memories. George Carlin's Rufus was a "good 'un."
Ha. Amazon's in-app AI agent is called Rufus. It's incredibly annoying, and they love directing you to an 'Ask Rufus' page where you watch an AI agent not-quite-finish repeating your search terms back at you before you realize you're not actually on your search results page, and then spend some additional time trying to find the link to take you to said search results page.
I assume somewhere there's a team hitting their AI query KPI's due to this, but I really wish they'd shut off a few server racks and save everyone the annoyance instead.
There was internal JIRA ticket to rename Rufus -> Doofus as to dramatically improve performance and sentiment among shoppers on Amazon site.
We should see this improved version soon.
Amazon: famously not evil ad giant.
This is not them standing up for the small guy, this is them seeing their actual profit center in retail (advertisement and paid promotion) evaporating if human beings never view their listings.
It is a money grab from Amazon I'm sure. They want to make a deal with meta to get paid for access. Maybe also worried about liability I imagine on accidental unauthorized purchases?
Returns cost money and if an agent generates more returns and hitting the margins or seller retention then that's something that's likely going to get increased scrutiny.
The generally quoted return rate is 5% to 15% (and it varies by category). If Muse identified purchases are generating returns of 10% to 30%, then amazon is probably going to crack down on it. (Amazon Sellers pay from 30% to 50% for customer returns from their own pockets. https://medium.com/@myrealprofit/how-much-do-refunds-really-... )
This all has been progressively looking more and more like Stanislav Lem's Washing Machine Tragedy. <Emoji of a monkey holding its eyes shut>.
I'm sure Facebook will allow agents to post any day now...
This is why the chrome driver plugin and similar tools are so important. The site shouldn't know or care if its me or my agent working on my behalf.
Amazon is genuinely unusable since forcing their own Rufus agent. A 300LOC userscript is the minimum to get it functioning again.
This is straightforward to counter. Meatbag-as-a-service to hit the Buy button.
And so the cat and mouse game carries on...
That's already a thing:
https://en.wikipedia.org/wiki/Personal_shopper
Amazon’s enterprise value is made up entirely of AWS and Amazon’s advertising business. Thus according to Wall Street their e-commerce is a loss-leader whose only purpose is to provide inventory for ads. Agents shopping on Amazon do not show Amazon ads and are thus antithetical to Amazon’s business model and need to be blocked.
Amazon doesn't provide financial breakdowns between Ads & E-commerce fulfillment and inventory costs, so take the Wall Street article with a grain of salt.
The real big question is what happens to display advertising revenue on Amazon, if even a fraction of its traffic moves to these agents.
Amazon ads is a relatively large portion of its retail profits.
Display ads are meaningless to agents. And agentic commerce is a huge threat to Amazon’s profitability.
No wonders, Amazon is actively banning these agents. but for how long is this even viable? Do we simply move our shopping to more agent friendly destinations? Is this an opportunity for Walmart to finally reclaim it’s throne?
> Display ads are meaningless to agents. And agentic commerce is a huge threat to Amazon’s profitability.
Would agents be susceptible to advertising? If humans have issues differentiating ads from real results, and are swayed by TV ads and stuff (they are), then why wouldn't an AI agent.
I kind of assume they would, given how ChatGPT brings up garbage websites sometimes for my questions.
That’s an interesting thought experiment. How would advertising need to evolve to appeal to agents vs humans?
This is inevitably going to get looked at by antitrust regulators, either in this administration (depending on how much Trump likes the leaders involved), or in the next.
Shopify seems to have no issues partnering with Muse
>We are excited to announce we are partnering deeply with Muse to enable agentic checkout with Shop Pay on all Shopify stores, offering people an easy and delightful way to shop and check out with Muse.
A remarkable display of market power by Amazon.
Probably to protect their advertising business which is very strong.
We’ll see how this plays out.
There's a lot to discuss here - agentic commerce, IMO, is where things are heading. I'm completely convinced that people _want_ to shop using agents. But I want to understand the dynamics at play here: will underlying inventory providers like Amazon give up control and allow agents to use their website?
I find this question very interesting and I want to predict where things will head. On one hand Amazon has a lot to lose by giving up control on how their website is used. Ideally they'd like to control the full end to end experience. What happens to the work they put in their UI? All those endless A/B experiments to make the buttons look perfect?
But on the other hand - this is where things are heading. They can't stop agents from using their website. Not for long. And they have a lot to gain because overall purchases will increase since. They can accept it and create a plugin and try to offer as much "branding" as possible inside it. Or they can completely go against agents and block it.
Then you have to wonder, what really is the moat here? Amazon's logistics and its relations with partners and customers (return policy etc) and the trust it brings. The UI part dissolves and is no longer important.
Maybe they want only an allowed list of agents access to Amazon? Maybe that's where things are heading? They sign deals with maybe just OpenAI (for example)? Its not like Amazon makes their own agent + model. I would hate it things fragment this much.
In theory, legislation could help. If there were a pro-consumer movement in government.
Something like declaring agents as representatives of consumers and blocking them violates ADA or just shouldn’t be allowed unless they are abusive.
Lacking that, it seems it’ll be an arms race between offense and defense. At some point, computer control probably wins (running my browser, on my computer, is pretty difficult to detect with reasonable f1 score)
Anything is possible I suppose... I don't see a path that leads to Amazon actually further controlling the shopping experience inside a browser (where a local agent might have control that they can't reliably detect), short of forcing users into some kind of hardware-based attestation solution or something, but even that is not foolproof.
> On one hand Amazon has a lot to lose by giving up control
The everyday person is told to get over it and accept that AI is here so adapt or die. Companies start reducing headcounts to be replaced with AI. So now, a company is having to accept that people are using AI in ways they don't like so they should adapt or die as well too, right? I have no tears to shed if Amazon loses control. Also, it'll be one billionaire battling another billionaire at that point, and I'm kind of okay with that chaos. My world is no longer in my control so fuck 'em is my stance
It’s an optimistic view on AI - brings power to the people by mediating all the garbage.
I’ve already used it to handle customer support request calls. I see 30+ minute hold+conversation times and generally the issue gets resolved. That is hugely empowering.
OK according to game theory, if this is correct, why even put up a fight? You have nothing to gain by fighting. Just accept it and play along. Make your damn MCP. You have 100,000 engineers. Why can't connect Amazon to ChatGPT today? I don't know.
> Why can't connect Amazon to ChatGPT today? I don't know.
They just don't want to, because there's no ads. The human gets removed from the loop of visiting the site entirely. And it doesn't say Amazon all over it.
If the user is not browsing the site/app, Amazon cannot collect the analytics about your browsing habits. The AI system does though. There is absolutely no incentive for Amazon to discourage users from using their system. After all, the AI system might not show the user an Amazon item.
Amazon cant hold this off for the long run, nobody wants to scroll through a list of two hundred near identical packages of toilet paper. Agents will intermediate, and they will remove demand for low value items
This is already solved by the existence of hundreds of other stores that actually do some level of filtering and curation of products. I don't know why we want to play some cat and mouse game of getting an LLM to find the best toilet paper on Amazon versus just finding a reliable source of toilet paper to begin with.
Theh will probably consolidate more. Less choices even if they are worse is also easier for llms to do. Just look at how they pick technologies to use. Almost always the same for a given model.
Amazon is just trying to get you to use their agent, so they're not anti-agent, just anti-someone-else's
I would hate it if every company comes up with their own agent and only their agent is allowed to access the API. EU do your thing instead of cookie banners plz.
that's what they're going to do, trying to get $20 out of the average person is terrible compared to giving away your product for free and getting an endless supply of metadata