[nomat]

I think this is the future of computers for people. they will open their phone and say "hey siri get this book or computer or car" and the browser will show them a bunch of made up products generated by an AI and they'll get the product drop shipped from china and it'll be a shitty counterfeit just like the $5 256GB USB drives.

But they won't know because when you look for reviews, the review websites will also be generated just for you, filled with AI-made reviews and users. The entire internet will just be a computer that guesses what you want and shows you its best guess as the truth. a little matrix-y.

Reading back I sound like T*d K but the next 100 years is going to be unimaginable in terms of what silicon is going to be able to do. It doesn't have to be perfect, just good enough most of the time. Like a mcdonalds drive thru.

[opportune]

Yeah, until someone creates a website that just ships the products they say they have, and people type in the URL for the site directly and shop there?

If online shopping were truly such a terribly scammy UX in 100% of cases people would stop using it or shift to shopping from trusted sellers who recognize the tremendous amount of value they can capture by not scamming people.

[mrweasel]

> If online shopping were truly such a terribly scammy UX in 100% of cases people would stop using it or shift to shopping from trusted sellers

That slide has already started. I can't buy stuff on Amazon anymore, because I don't trust the quality or that I'm not scammed. Unless I trust the author I'd question buying self published books. There's a number of Chinese website, like Wish, that I don't understand. Why bother shipping an item to a customer if you're running a scam? People order dress and get a plastic Christmas tree... Why even bother sending the tree? Why produce shitty products that break after first use, why not just straight up rob people, seem more genuine and with less environmental impact.

There are whole categories of products that can not longer be bought online and an even larger category that can only be purchased from trusted companies.

The level of scams have reached new heights in the past year and unless we do something to address "A.I." generated content the internet will drown in useless nonsense.

[phpisthebest]

>>Why even bother sending the tree? Why produce shitty products that break after first use, why not just straight up rob people,

Could be the same reasons Scam Contractors do "some work" if you hire someone to redo your bathroom, and they take the money and run, that is clear criminal fraud. If they take the money, so some amount of work, then run it becomes a civil contract dispute.

[martin_a]

> Why bother shipping an item to a customer if you're running a scam?

Plausible deniability.

"There seems to have been some..." looks up excuse for the day "... unfortunate mix-up with your order."

Lots of people will just not react, because it's not worth their money and at least they received something.

[bobboies]

So the trouble is these market places allowing third party vendors on there (obviously).

All you need is a system to authenticate reputable sellers. Instead of looking at the stars on a product and clicking “buy now,” folks just need so ask: do I know this seller? Is this the ACTUAL publisher selling me this book?

I’d rather go into a real shop anyway. Unless it’s buying some hard-to-find item there’s no excuse.

[bunga-bunga]

> I’d rather go into a real shop anyway

You’re assuming that scams don’t make their way into regular distributors as they try to cut corners and maintain competitiveness.

You already see established brands lower the quality of their products after the first round of reviews online. If found out, they just blame manufacturing. Sorry.

[JohnFen]

True, but it's much less of a problem in physical shops. You can actually examine the merchandise. You know that the item you're paying for is literally the item that you're getting. You're dealing with the shop face-to-face, which reduces a lot of the more brazen scams. And if you have a problem, the shop is likely to do something to fix it -- and if they don't, you have realistic legal options.

In most ways, buying from a brick-and-mortar establishment is a better choice than buying online.

[jasonladuke0311]

Yep, Patagonia is suing Nordstrom for selling fakes: https://www.seattletimes.com/business/retail/patagonia-accus...

[bobboies]

Nordstrom being a third party vendor of course. You want Patagonia, go to Patagonia. :)

In my mind we’re talking about what is a better solution to the problem, and it’s that companies WILL protect their IP. If someone buys fake Patagonia, Patagonia loses out on a sale.

Amazon doesn’t own much of anything IP-Wise on their marketplace in the grand scheme. And they don’t care who buys what as long as products are moving. It’s easier for them to just pay out returns than handle the actual problem.

[bradknowles]

Marketplaces need to be held liable for products and services sold through them.

And when you have a dispute with a seller that came through a marketplace, your dispute should be with the marketplace and it is then up to the marketplace to recover the value of the goods from the seller.

[gnopgnip]

Amazon has been problematic for a while. I think it was 2019 when I remember filtering to sold by amazon.com was a lot harder, but before that there were still problems with tons of low quality products and fake reviews

Costco bucks this trend though

[red-iron-pine]

> Why bother shipping an item to a customer if you're running a scam?

Because they have a warehouse full of ugly misprinted clothes and need to get rid of them. and those clothes still cost (some) money to make or acquire.

and as others have said, not sending anything is wire fraud, and outright scams will catch criminal charges; one more line on an Interpol indictment. Crappy or underperforming goods are a civil issue, and no one is launching an international lawsuit over a $100 item

[freeopinion]

This doesn't just apply to physical merchandise. And it isn't new. Isn't this the whole point of sites like yelp? Or Fox News? Whether you need a good auto mechanic or good mayor. Or if you want to know if a movie is worth paying for or whether paper bags are bad for the environment.

"How to verify any claim" has to become a formal subject like math and reading to be taught in stages over several years. It has to have real homework problems to practice different strategies. But in the end, do we come up with different ratings or logic categories to say this particular question is fifth degree hard to get right? Can we identify the elements that build trust? Does it help in poker if you know another player's family? Does it help in voting for student body president if they previously served as class president? Is it better to buy a blender from the same company that made your favorite dishwasher? Can you take a chance on a newcomer if they don't speak your language or have your accent? What if some claim hasn't been published in a peer-reviewed journal I recognize?

[blq10]

People willing to take risk on Wish because they are spending frivolously and would like to spend less frivolously.

Scams are high, but not so high that you won't usually get what you asked for.

[labster]

We’re in a race to the bottom. There’s no way a company selling legitimate products will be able to compete or grow a customer base in the AI future, because we’ll be swimming in garbage.

[soerxpso]

This assumes that the only ways people have to find companies are search engines and other mediums that are prone to AI takeover. If everyone starts to realize that most of the surface internet is garbage, they would start communicating with humans they know about which places are not garbage. It was possible to find good businesses to interact with before Google existed.

Obviously in a world where every online retailer is nearly useless, a guy who creates an online retailer that is not useless would immediately have anyone he personally tells about it as customers, and anyone they tell about it, and so on (since in this hypothetical, everyone is desperate for a good retailer).

[bsenftner]

And the few, lone, honest online retailers get early discovered by online scammers who turn that honest retailer inside out, destroy them financially, and then take over their URL. I've seen it happen, and it nearly happened to me but I shut down rather than spiral down, unable to wage war on so many fronts.

[codpiece]

Could you please elaborate? This is fascinating.

[pbhjpbhj]

Even when you take a recommendation from yourself ... ie re-buy a product, it often seems that the company has been captured by VC and the brand has been used to make short term profits and it's no longer any better than whatever white label products you can find on AliExpress.

[UtopiaPunk]

That's how people find the good torrent spots, so topical.

[andrei_says_]

We swim in garbage already (looking at amazon) but bhphoto and sweetwater still exist.

[INGSOCIALITE]

Sweet water is great - they even have a dedicated rep call you a little after your order arrives to make sure you’re happy with the product and that it’s just what you wanted.

Plus they send you candy

[extraduder_ire]

mcmaster-carr exists too. I've never been in region they serve though

[RajT88]

Newegg as well.

[Izkata]

Newegg is now an amazon-style marketplace, with stuff they sell mixed in to the rest same as how AmazonBasics is mixed in to the rest.

[yomlica8]

Newegg is a shadow of its former self, but it is still possible to turn the 'sold my newegg' knob on relatively easily. It seems impossible to do that with Amazon. Walmart online you can do it but you'll have to watch it like a hawk as it gets aggressively disabled again.

[RajT88]

Well shit.

[homarp]

and apple.com

[fennecfoxy]

The streamlined and stylised experience of paying multiple hundreds of percents mark-up for ram and SSDs.

[prmoustache]

People might revert to buy in person.

I mean you get scammed maybe twice by a company, then buy somewhere else. You quickly realize that it is better spending 3-4 times the amount of money for something that will last 10 years than buying stuff that do not work or keep breaking after a few months/years.

I am too poor to be able to afford buying cheap stuff.

[phpisthebest]

If that was going to happen, it already would have.

Most people still just buy the cheapest version of an item they are looking for.

[HideousKojima]

Most people buy the cheapest version because they don't have a reliable way of knowing if the more expensive items are actually higher quality. It's a "market for lemons" problem:

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

[ben_w]

I'm already switching to buying in person, though in my case because of the delivery service usually redirecting to depots further away than the shops selling the same items.

[wombat-man]

Yeah, I'd rather go to target or something and choose from the 3 spatulas they have then spend any time trying to figure out which of the 1000 listings are good on amazon. It's just less work.

[red-iron-pine]

it's called Walmart, and already exists as an in-person option.

[checkyoursudo]

For the most part, I already mostly buy stuff in person. For physical goods, I cannot stand not being able to hold, touch, examine a thing before buying it. Or at least the floor model or the exhibited item.

I bought a VOIP phone from the media store recently. When I was there looking at them, and feeling them etc, I was thinking how there is no way I could have picked the right one from an online shop. I might have never known there was something better, I guess. But being able to hold them helped me pick one I thought I would like.

[sesteel]

What Amazon is missing is buyers. Actual professional buyers who evaluate products and make determinations if products meet standards and deserves to be placed in front of customers.

[giantg2]

Plenty of niche, premium, or physical stores out there. People might need to pay a little more for it, but they do exist. Not everyone buys everything from the same super stores.

[UtopiaPunk]

What are some of your favorites (tech or otherwise)?

[giantg2]

Small local businesses, like homebrew shops. Seed companies like Rare Seeds, Gurney, etc. Mushroom places like The Cultured Mushroom, Liquid Fungi.

[skydhash]

Keyboardio

Roost Stand

Accessories4less

Svsound

[prmoustache]

Yes. In term of online store it is better to look for the ones specializing in a specific market than wholesale stores.

[neilv]

We're already getting there.

Maybe Amazon brands will at least be predictable, but the non-Amazon brands on Amazon will be increasingly counterfeits and noise.

(A contributor to the current problem of eroding non-Amazon brands with counterfeits seems to be Amazon's commingling from different marketplace sellers of the "same ASIN". And the current mass-spam search DoS of random-name brands and near-identical products from overseas sellers is also a problem. Amazon can not only shield their own brands from both of those problems, but can also exercise more supply chain integrity control for its brands than is possible for the non-Amazon brands if Amazon doesn't cooperate.)

[smithza]

Yet I think Amazon brands are setup to eliminate competition due to their insider information in the market place. I wouldn’t put my money in them because they aren’t playing fairly in the market.

[itsoktocry]

>I wouldn’t put my money in them because they aren’t playing fairly in the market.

They identify widely ordered products and source their own version to sell. Why is that a negative? Costco does the same. As long as they vet the quality (my experience with AmazonBasic products has been fine), seems like a win-win?

[fineIllregister]

Costco doesn't sell counterfeits of the third-party items they are competing against. Amazon's co-mingling practice is anti-competitive.

[dragonwriter]

> Maybe Amazon brands will at least be predictable, but the non-Amazon brands on Amazon will be increasingly counterfeits and noise.

Of course, because that helps sell the Amazon brand goods.

[zebomon]

I really don't see how one follows the other without then in sequence, as the comment you replied to describes, value-based businesses emerge to fill the obvious gap and compete.

[dylan604]

people still shop at Amazon, so I'm not too convinced in your argument

[2muchcoffeeman]

It’s the delivery. I can often get something inconsequential for net cheaper, faster from Amazon. Local stock is usually next day. And I can often get stuff internationally cheaper and faster through Amazon than the real retailer.

You’d have to duplicate this to take them on.

[dylan604]

step by step that slope gets just a little more slippery, or is it the frog in the pot of boiling water?

[Semaphor]

I have gotten a fake product one time (public domain book, PoD instead of what I wanted and what apparently used to be under that URL) refunded, no questions asked, even got to keep the book (anyone in Germany interested in an HP Lovecraft decent quality hardcover PoD book?).

I realize that it’s apparently way worse in the US, but there supposedly are issues here as well, I just never encountered them (or the products are so well-faked that it’s a distinction without a difference).

Fast delivery, a lot of choices, both "Aliexpress but faster delivery for a price", but also EU products, and a customer service that does pretty much anything asked.

[stOneskull]

there'll still be people and communication. and so, word-of-mouth.

[Fricken]

This might be why physical bookstores are making a comeback. Trustworthy curation is essential, and worth paying a premium for. This has always been the case, but maybe not fully appreciated.

[interestica]

This has been why I've started to really like Costco -- they're essentially acting as product curators (and back it up with a generous return policy.)

[dangus]

Well, I like how local bookstores go about curation more than Costco.

I don't know if we can say that a store with unlabeled aisles can be a source of curation.

[yonaguska]

it's curation in that the products they do offer are generally trustworthy.

[kelnos]

I think that's an ok workaround, but I don't want to go to physical bookstores, and I don't want physical books. As much as I do like the feel of a physical book, I much prefer the convenience and portability of a digital book.

And speaking of convenience, it would be a big shame if we had to go back to physical stores for non-trivial purchases because we can't trust anything online.

[arnvidr]

Many physical bookstores have web pages where they sell e-books as well. Find the store that serves what you want, instead of finding faults with the ones who don't and going with the huge sell-everything-store.

[bpt3]

Or I can go with the huge sell-everything store because I'm perfectly fine with the service it provides. Not everyone is looking for some bespoke curated experience when buying a product.

In short, the store that serves what I want is usually the huge sell-everything store.

[eloisius]

> Yeah, until someone creates a website that just ships the products they say they have, and people type in the URL for the site directly and shop there?

This already happens over and over, and with AI it’s only going to be worse. Company makes a decent product, gains reputation, and then strip mines their good will for quick cash while quality plummets. After a while, you just distrust the whole online shopping experience. Even if someone is selling a good product, the lemons are ruining the whole market.

[bsenftner]

And you can say "the lemons are ruining the whole market" in relation to software engineering jobs too.

[LapsangGuzzler]

I think AI is going to make it harder to separate the signal from the noise, though. We’re already seeing this play out with sites like Amazon that are just full of cheap junk that’s blatantly ripped off and undercut in price, or Google as people are complaining about the quality of search results.

[ianburrell]

Notice that niche shopping sites don’t have reviews. Bigger sites have reviews but they don’t get gamed. Without affiliates or listing products like Amazon, there is no way for the spammers to make money.

I’m surprised that places like Walmart or Target haven’t gotten burned by cross listings. They could drop the program and advertise as alternative to Amazon. If Amazon had “sold by Amazon” filter, it would be a lot more usable.

[pbhjpbhj]

Amazon (UK, at least) don't even have a working search function, it's ridiculous trying to buy a specific thing there.

[Matrixik]

I think that's the reason why Allegro is market leader in Poland and Amazon is far, far away. I get really frustrated when I try to find anything on Amazon, sometimes to the point of thinking "how the hell they became so big with this shitty search?". When I search anything on allegro.pl it almost always return me products I search for and support tons of filters (and return actually nothing when they don't have it).

[JohnFen]

> If online shopping were truly such a terribly scammy UX in 100% of cases people would stop using it

Some people have. This is one of the main reasons why I stopped using Amazon years ago.

But lots of people won't. If people were that rational, then the thousands of straight-up obvious and traditional scams perpetrated every day wouldn't still work.

[nomat]

Imagine like an LLM trained on everything you've ever done since birth included as part of the data set. and then one of those for each other person in the world. they could slice and dice you into a thousand different groups.

at some point technology is going to stop being something that we interact with through a screen and a keyboard. Google maps will be there, where ever you go, all the time. No more getting lost or wondering where anything is located. You will just know. or at least, be told.

Something else to consider is the fallibility of human memory. As we generate more information we're gonna have to start abstracting and delegating some of those storage and retrieval tasks to the computer, trusting it completely. Today maybe you can remember things but what about your parents?

[jbergens]

I don't remember everything I've done or why. There is very little training data to be found for us as persons. IF they hack your phone or computer they might be able to store how you interact with the internet. Which things you browse and read, what you buy, how you write on social media. They could possible then create a digital twin (a clone) of your online presence but it would be a lot of work for very unclear gains. If they hack enough people they could learn some connections like that people interested in electric cars might also be interested in battery technology and solar panels, people who are interested in vegetarian food might also like to read/write about vegan food or the environment and so on. But I still think it would be a LOT of work just to end up creating a fake website selling fake batteries or vegetarian food that you either never deliver or that ends up being a poor replica of the good stuff.

[waihtis]

Well, currently the top LLM fails at repeatedly generating a stable set of pretty simple regexes for me, so where pretty far from a full-blown LLM dystopia at this point.

[sznio]

LLMs are too dumb to be useful, but smart enough to be destructive.

You don't need to be able to generate correct regexes to create a blogspam article about regexes that will go to the top of the search results.

[Planktonne]

> LLMs are too dumb to be useful, but smart enough to be destructive.

I think this puts the blame in the wrong place; LLMs are too dumb to be useful, but it's the users who are too dumb to realise that, or too unethical to care.

[uoaei]

It's also the creators, hosters, and startup founders who are overselling its capabilities.

[waihtis]

search is dying anyway and this will just accelerate that trend.

[mavhc]

The question is does the rapid progress in LLMs continue over the next few years, or do we reach another local maximum. Is "currently" going to improve rapidly?

[waihtis]

as I understand we're approaching local maximum with LLMs and that is what I base the previous comment on, but fully layman and no insight into deeper layers of LLM R&D

[eloisant]

I have already stopped using Amazon for specialized products.

I buy board games on sites that only sell board games.

I buy electronics/computer parts on sites that only sell that.

[jimhefferon]

I also try to do that. However, the more sites that hold my information ...

[eropple]

Your information isn't particularly valuable outside of financial data, and most of the ones you're likely to go to store your financial data centralize on a few generally-very-well-behaved providers. And credit card fraud is the bank's problem, not yours.

[willis936]

It's only possible to compete if the index is fair. You could never launch a store like that without google in the fold. Google doesn't have an interest in increasing signal, only in increasing revenue. We'll need to see some trust busting first.

[pixl97]

Why be a billionaire when you can be a trillionare? Greed will ruin everything.

[opportune]

If incumbents are delivering a terrible product because they’re chasing 4 commas I will happily settle for 3 commas to provide a basic “shopping but on the internet” experience people actually want to use.

[dylan604]

i see your Tres Comas and raise you Quatro Comas

[red-iron-pine]

> But they won't know because when you look for reviews, the review websites will also be generated just for you, filled with AI-made reviews and users. The entire internet will just be a computer that guesses what you want and shows you its best guess as the truth. a little matrix-y.

dead internet theory -- the idea that the internet will eventually be mostly or entirely bots talking to bots

https://en.wikipedia.org/wiki/Dead_Internet_theory

[elkhadiy]

Very interesting! Bots roaming freely generating content and manipulating consumers sounds like early stages of the net beyond the BlackWall in the Cyberpunk universe post DataKrash. (just in case someone might be interested in sci-fi reading on the subject)

[checkyoursudo]

There won't be any consumers on the net to manipulate though. Our bots will be buying counterfeit junk off of shady e-commerce giants, but at least we will be out enjoying art, sports, and nature, because all the bots will be doing all the work as well anyway.

[Workaccount2]

It tickles me that somewhere out there on some forgotten instance in a dark corner of some massive data center, there is a #IndieMovies chat channel still running on some long forgotten college kid weekend project, and in that are two early gen chat bots that have been going back and forth for over a decade now, and will continue to for the next 20 years.

[wpietri]

Yeah, my big concern about LLMs is not Skynet or a zillion people out of work. It's that the spammers and their ilk are now vastly more capable. The fact that generative LLMs are glib bullshit machines is a problem for most sane uses, but a plus for people who don't give two shits about the truth as long as they've got a chance of making some money.

[akira2501]

It reminds me of payola.

Radio stations would get paid to make songs that weren't popular seem more popular than they were so people would blindly buy them. It's the same old snake oil just taken to mass media.

Since both behaviors revolve around profit, you can apply the same means to mediate them. Further, the open plan "user generated content" model has been dying for a while now, the wills of some well moneyed monopolies keep it alive. We will iterate to something much better anyways. Perhaps this is just a force that drives it.

[hsbauauvhabzb]

This, but also in the professional space. My coworkers much be 200% faster now, but I’m going to spend 80% of my time unfucking their work.

[hsbauauvhabzb]

But also good luck next time you google health symptoms - sure it’s bad, but everybody does it, and we’re going to move further and further away from accurate.

[mrweasel]

You can rely on the internet to provide even remotely correct health information anymore. It was always a little questionable, due to all the variables, but you could get some indication on how quickly you should get yourself to the doctor.

Same with weight-loss or exercise programs, that's basically impossible to search for due to the amount of spam and garbage content.

[throwaway2037]

Regarding health, stick to Mayo Clinic, US Centers for Disease Control, or other very reputable websites. That one seems easier than buying stuff online. The slow descent is hard to navigate for many people.

[wpietri]

I think health care might be one of the areas where I'm least worried. This will drive people back to expertise, though. I might google symptoms to start, but I'm not going to take anything seriously unless it's from a reputable health-care organization like the Mayo Clinic.

[JohnFen]

Not everyone does it. I don't, because doing that is not only pretty worthless, but can be hazardous.

[wpietri]

Yup. And this has always been a problem in software. Too many people treat a team sport as a bunch of individuals running around on the field doing their own thing. But I think it's going to get worse because of the way LLMs differentially amplify the least competent.

[tiborsaas]

There's already a WordPress plugin to auto-generate articles on schedule. Probable there are dozens already.

Authenticity will be a valuable currency in the future.

[menus]

You have described Dead Internet Theory:: https://en.wikipedia.org/wiki/Dead_Internet_theory

[spicyusername]

I think people have a lower tolerance for this than that, and will just go back to relying on personal recommendations and other reliable forms of curation.

[citizenpaul]

I think we will see the pendulum swing back. Yahoo lost to google because they were a manually curated search engine and could not keep up with google indexing and search. Now that all search is SEO gamed garbagebwe are going to see a new curated search engine rise in market share again

[hedora]

That’s not why Yahoo lost. They switched CEOs and corporate direction / business models more than once a year on average toward the end.

Also, during the collapse of yahoo, it was a poorly kept secret that google had already mostly abandoned page rank, and had large manual curation teams.

To add insult to injury, in unbranded head to head tests, Yahoo search had better search result quality than google for many years before it was sold to Microsoft.

[vbezhenar]

I'm pretty sure that Google uses human adjustments for their search results.

[realusername]

Knowing how Google is allergic to anything human in their product loop and obsessed with automation at all costs, I doubt they make a difference.

[userbinator]

Unfortunately, "personal recommendations" seems to mean "influencers" these days.

[dylan604]

that's the part that I love love love. society has gotten to this ridiculous follower stage that people seem to think it is a personal connection. maybe i've just spent too much time too close to the ad agency world to see how nasty they are, but anytime i see some random person on the internet that i do not personally know recommending anything, i immediately assume they are paid for that "recommendation". too much "inside baseball" to know that nothing you see in an ad is real even if you don't consciously recognize the fact what you are looking at is an ad.

[throwaway2037]

Hell yes. +9000 for your post. To me, the solution is much transparency around paid product placement through strong regulation.

See "Parasocial interaction".

    Parasocial interaction (PSI) refers to a kind of psychological relationship experienced by an audience in their mediated encounters with performers in the mass media, particularly on television and on online platforms. Viewers or listeners come to consider media personalities as friends, despite having no or limited interactions with them. PSI is described as an illusory experience, such that media audiences interact with personas (e.g., talk show hosts, celebrities, fictional characters, social media influencers) as if they are engaged in a reciprocal relationship with them. The term was coined by Donald Horton and Richard Wohl in 1956.

Ref: https://en.wikipedia.org/wiki/Parasocial_interaction

[eropple]

I'm currently noodling on a blog website and I'm thinking about adding a "recommendations" page largely in an anti-influencer sort of way--any affiliate marketing money is pretty incidental to somebody working in tech, but I want companies that make stuff that's good and valuable to stick around and continue to make stuff, so I want a way to spread the good news in a credible manner.

[dylan604]

Then don’t use sponsored links. If you’re making money off of the recommendation in any form, it loses credibility for people that care. Otherwise, you’re just like all of those listicles that exist for no reason than a way to use affiliate links.

[JohnFen]

> anytime i see some random person on the internet that i do not personally know recommending anything, i immediately assume they are paid for that "recommendation"

Yes, and doubly so if they say something like "this is not sponsored".

[red-iron-pine]

lifestyle marketing. goes hand in hand with the AI stuff, really.

if 90% of the market is bullshit, the key is to find the people you trust and want to be similar to.

"This is who I am, and who I'd like to be, and this is what they do and use".

And admittedly, a lot of these oft maligned influencers are actually trying on the makeup on camera, eating the food, etc. Not a random bot but a human.

We'll see if GPT and Midjourney are able to hack that, too.

[hedora]

The streaming services business model more-or-less outlaws collections that are curated by third parties.

We already know what happens when piracy is the most convenient/highest quality option, but I’m guessing media distributors will need a reminder soon.

[kubanczyk]

> relying on personal recommendations

Taking things off the net would be a huge step back for economy.

LLMs will cause a total collapse of trust online. This problem will be solved by a brand new online trust model (instead of the old one "has gained some followers by formulating text messages and forwarding gifs").

The trustees (bots/humans/orgs mixed) would probably need to sacrifice much more of their privacy. Not for the sake of orwellian Big Brother, but for re-establishing any trust whatsoever online.

[JohnFen]

> The trustees (bots/humans/orgs mixed) would probably need to sacrifice much more of their privacy.

I'll just ignore the internet instead. The internet cannot be trusted when it comes to privacy, so I'm not going to give up any in the first place.

[brianmcc]

Yep. Case in point - I've given up buying on Amazon UK and use Argos as my go-to shop instead, for products of far better assured provenance without having to really overthink it all.

[0xbadcafebee]

Your first paragraph is just Amazon today. They don't need AI for people to make up products and send you garbage. They probably will use AI for reviews, but they already have fake reviews that reliably fool people, so this isn't much different. If it gets so bad that Amazon's bottom line is threatened, they'll actually crack down; they're assholes, not business-suicidal.

"The entire internet will just be a computer that guesses what you want and shows you its best guess as the truth" - you just described Google Search, like, 5 years ago.

"the next 100 years is going to be unimaginable.." - not really. What we're capable of is not what happens. We only reach the minimum viable capability that meets a societal expectation. Once the expectation is reached, progress stagnates, until something exceptional changes the expectation. As a more specific example, regardless of how advanced hardware gets, software remains a gas that fills any container, so what we can actually do with the hardware is always limited by our horrible software.

People keep talking about AI like it's going to transform everything. It's actually more like colorized talking movies vs the black-and-white silent ones. Reshapes industries and changes expectations and experiences? For sure. But the films, plots, characters, actors, industry, viewers, their effect on society, etc remain the same. 3D didn't take hold, neither did smell-o-vision, drive-ins, and countless other "innovations", even though they were feasible. This is just one example, there are tens of thousands of other examples I could give. It's like a law of nature. We don't accomplish what we aren't motivated to.

Also, mcdonalds drive thrus work all the time, not just most of the time. The drive thru is the engine that generates most of the revenue; it's a conveyor belt for cash. A franchise would go out of business if the drive thru broke.

[nomat]

I'm talking about the quality of the food here.

as for the rest of the post, the application is not limited to consumer products. Imagine the 2016 election season in 2056. or another covid pandemic.

[0xbadcafebee]

AI isn't going to change a pandemic or an election by itself. Those are completely controlled by societal norms and actions of humans, regardless of any technological powers or their uses.

In the 1918 Flu Pandemic, we had the technology and knowledge to end it quickly. But nations around the world chose to spread disinformation and cover-up the story, which made it much worse. A third of the world died because of how society chose to act.

And when we make a product that can do harm, we also make a product that can revert it, clean it up, or institute laws to restrict it. Now that we've seen things like Russian bot farms, we look for them and shut them down. Despite that we have good AI, we can also detect AI. Our ingenuity can almost always be surpassed by more ingenuity. As much as it can cause harm, it can also fix said harm.

AI is just a hammer in a toolbox, and it isn't any worse than any other hammer we've had. Don't worry about the hammer; worry about the person holding it.

[stevage]

I imagine everything constantly hovering on the threshold of what is tolerable enough for most people. If you could never buy anything useful on the internet, the system would collapse and people would do something different. But instead, the quality etc will just decline to a point where it is still marginally better than alternatives like shopping in person etc.

[ghaff]

People who can afford to do so will perhaps just pay 2x to 3x to shop a more limited selection from personally curated stores.

[shinycode]

Unfortunately if you so create a new product or service it’ll be really hard to find you this way and it might require a lot more money and effort to reach out to the audience.

[b112]

Oooooh, custom search engines.

Amazon would hate this, but make a deal with shopify to speed up spidering their sites. Then, spider all other retailors.

You put in a product id/name, and it only shows results from amazon, shopify sites, and vetted retailers.

The manual curation part, is vetting the retail site. Walmart would be a problem(for example), unless it was easy to tag 'sold by walmart'. Amazon's filter would only include shipped by amazon stuff.

So like 'hanes t-shirt', and you get 7 links, from 7 sites, all woth prices and shipping cost estimates.

Amazon has horrible pricing, bordering on predatory for some things. Maybe I'll do this, I'm bored this summer.

[zen_1]

One of the few sentences more depressing (when relevant) than "Stallman was right" is "Ted Kaczynski had a point".

[kubanczyk]

Categorizing useful arguments into boxes with labels like "Ted K" is over-simplification. Worldview is not about chains of arguments, but directed acyclic graphs of arguments. You can even have exactly the same starting assumptions, but arrive at a totally different branch.

It all depends on how well you connect the dots on your way.

[coliveira]

This is already happening: whatever you ask Google, they have some crazy AI algorithm running in the back trying to discover what you want to see as result, so as to maximize the profit from your clicks.

[bombcar]

Optimistic of you to assume “what you want to see” and “what makes you click most profitably” coincide.

[slabity]

Considering they haven't completely removed Pinterest from their platform, I absolutely agree.

[JohnFen]

I think that this is a larger factor as to why Google search sucks these days than SEO is.

[ulfw]

People value choice in their purchasing behaviour. I am not saying people are always rational actors, of course not. We are all influenced to some degree by ads, social influences etc etc

But the whole notion of "I just buy from one place and one place only without comparing anything at all ever again" is a bridge to far in my mind. So the whole "Siri send me whatever I tell you" I doubt will catch on with the majority of people. A niche maybe, sure. But not the majority of us.

[Pxtl]

This is why I have conspiracy theories about the conspiracy theories about digital IDs. I don't see any escape from a hellscape of AI-generated-sludge unless we have a real way to say "somebody has signed their real government-confirmed reputation on this". The paranoid disinfo peddlers are the ones who benefit most from the untraceability of the horsecrap they put online.

Otherwise it's all downhill until the singularity.

[nomat]

Free speech should only apply to verified humans. We could use the blockchain to assign keys to babies as they are born.

Only half kidding, about both statements. Corporations also have rights. How terrible it will be when we have "companies" that are just AI programs. All human weakness, eliminated.

South Korea requires you to register with your national id if you want to access the internet there, which sounds like the wrong direction but might be necessary if we want to maintain some semblance of sanity on the internet or whatever it is going to become in the future.

[kubanczyk]

> verified humans

Wouldn't it be elegant though if the new trust scheme applied the same to LLMs/bots as humans?

Assuming that LLMs are a polar opposite of humans seems short sighted. Or assuming that humans are sole heroes of the story. It will be a spectrum, soon. For example when you say "all human weakness, eliminated" it's a weird lens to use. Every relative weakness is a challenge, whether its origin is in human ancestral environment or not.

I'm curious whether the entirety of the system (I mean the current civilization) would morph itself into a pipeline disgorging specifically human "happiness". As of today, it's far from obvious!

[Pxtl]

Ultimately? Of course. But in the short term? The LLM's owner can vouch for the LLM, same as anything.

I honestly don't care if a human augments their output with an LLM. What I care is that they act in good faith. As long as new identities can be spun up automatically with negligible effort, human or otherwise it's too easy to flood the zone with shit.

I want to be able to say "This post was made by an ID that is tied to a human being that has a history of acting in good faith, and I can see that history". Even if "I" am the owner of a crowdsourced-content site and I'm keeping that ID private. If the person who uses that ID is leveraging an LLM to create what looks like good-faith content, that's fine. The problem this solves is one person using LLMs to create a sockpuppet army that looks like a hundred thousand humans, not one person using an LLM to generate a flotilla of good content. Or even a flotilla of bad content but we still konw who's doing it.

[JohnFen]

> but might be necessary if we want to maintain some semblance of sanity on the internet

A requirement like that would be the final death knell for the few remaining parts of the web that are actually good.

[worldsayshi]

There has to be some neat tricks you can use to allow verifying a human, using your id, without actually exposing your id though?

[eropple]

You can sign an intermediary certificate, but nothing stops you from then giving that to an LLM.

[Pxtl]

Well, if that cert can be traced back to its original creator, and that LLM is crapflooding, then that person can be blocked from user-contributed content systems.

[worldsayshi]

Many aspects of this problem doesn't sound hard to solve. But you need to involve cryptographic signing and reliable eIDs.

Any content from a known author needs to be verified to be from that author by letting that author sign the content.

[dsr_]

Right, just solve the Oracle problem.

/s

[mr_toad]

Maybe people will finally stop believing everything they see online.

[kzrdude]

It's only the future if we accept these half truths and deceptions. Like how Amazon is accepting it.

[femto]

Even that's optimistic. I'd say "its best guess as the truth" should read "its best guess at what you will accept as the truth". It will be entirely about what sucks in the individual.

[fennecfoxy]

Lmao, scams and crap products have always been a thing since we crawled out of the oceans. As always, the problem isn't technology but humanity itself.

[otabdeveloper4]

This is exactly how it works today. Just with "copywriters" instead of "AI".

[neycoda]

Nobody's expecting perfect but humans' definition of good enough typically isn't.

[dclowd9901]

Is it not the case that AI is pretty good at detecting AI generated content?

[menus]

OpenAI does not think so.

> As of July 20, 2023, the AI classifier is no longer available due to its low rate of accuracy.

https://openai.com/blog/new-ai-classifier-for-indicating-ai-...

[dragonwriter]

It is not, which is why lots of people have been focusing on watermarking–which, obviously, all kinds of actors that don’t feel the need to comply with regulations, either because they are exempt (practically, even if not theoretically) governments, or “renegade” anti-establishment will simply either not implement or not provide the public with the pattern information needed to detect the watermark, so...

[HWR_14]

If so, it's only for one generation, until they add "tricks current AI detection" as a metric for the next training.

[CamperBob2]

It's not. One of the things you can train an AI to do is to generate output that isn't recognized by a given "AI detector."

[post-it]

It is not the case.

[fellowmartian]

Ah yes, another classic apocalypse prediction! Goes something like this:

- Identify a problematic trend

- Extrapolate into infinity

- Assume nothing will be done to correct it

- Apocalypse!

It’s very popular among journalists and experts, despite the fact that this literally never happens, because it turns out people don’t just stand still and drown if they see a wave coming for them.