An the same will happen to AI. We will remember these days as the golden age for AI, where you weren't required to prompt an AI three times before it answers with a non-ad response.
What it actually is: a salesperson
It took the mass public a long time (15 years?) to realize search engines had shifted from the former to the latter, and that allowed Google to leverage that misplaced trust into huge profits.
Expect commercial AI to be the same, unless it's explicitly set up otherwise (read: Kagi assistant).
This guy couldn't recognise the conflict of interest, and neither will 80%+ of AI users.
Edit: https://www.youtube.com/watch?v=Lzbz0HDVKEs "Louis Theroux visits top gambler's Hilton hotel suite - Gambling in Las Vegas - BBC"
It even supports bidding for the ad space!
Source: https://research.google/blog/mechanism-design-for-large-lang...
But whether it's search or AI-chat, what's annoying is efforts to have it replace that things that exist rather than serving as useful addition. I use ChatGPT X many times a day (or hour) but unless I ask for an AI's opinion, I don't want it.
That's not how that works; "someone is doing this" doesn't prove a rule "no one is doing this" -- quite the opposite
"The exception that proves the rule" is for things like "closed Thursdays" (rule = open on other days), "no parking after 8 PM" (rule = parking allowed before 8 PM), "no refunds on games" (rule = refunds available on other items), etc.
Or more succinctly, as first-year law students learn: Expressio unius est exclusio alterius — to state one thing is to implicitly exclude others.
https://definitions.lsd.law/expressio-unius-est-exclusio-alt...
> "The exception that proves the rule" is a saying whose meaning is contested. Henry Watson Fowler's Modern English Usage identifies five ways in which the phrase has been used…
Personally, I use it in cases like:
- Rule: Don't do X, it's a bad idea.
- Exception: One time, someone with very special circumstances did X, and with a lot of finagling and effort they managed to make it work sort of OK.
Or:
- Rule: This fortress was an impregnable defensive position.
- Exception: In A.D. 1305, the fortress was taken, with great difficulty and many casualties, by an attacking army 100 times larger than the defending force.
Or:
- Rule: This river never overflows its banks.
- Exception: Once in history, on the day of the biggest rainstorm in 1000 years, the river is recorded to have overflowed its banks very slightly for a short time.
The exception proves the rule because the circumstances necessary for the exception to occur were themselves exceptional.
For example, imagine if your skydiving instructor said "if your parachute doesn't open when you jump out of the airplane, you're gonna die", and you replied with "well actually that's not true, Vesna Vulović survived a fall from high altitude." Yeah, okay. The fact that you had to be smarty-pants about it and dig up a random exception really proves the point they were trying to make.
(Also that story is nuts! https://wikipedia.org/wiki/Vesna_Vulovi%C4%87)
People on this website are so fucking pedantic and argumentative over the most obvious or inconsequential minutiae it drives me nuts.
That they were responding to someone else who couldn't see how it worked means it wasn't just pedantic nitpicking
I didn't say it was personal, I said it was irritating.
Also sometimes when doing more complicated purchases that require multiple products, I use it to sift through Amazon.
Especially ChatGPT seems to be optimizing for this use case, like a “search engine that can actually reason” (by lack of a better description). It’s convenient, and saves me a lot of time compared to the mess that Google has become.
(Obviously it’s likely this will happen to AI as well in the future, but right now, it’s pretty good)
Obviously, a better search engine that also doesn't display ads is better. But is it $20/month better? When it's also got daily usage limits? And they're almost certain to start injecting ads as soon as they possibly can without alienating people?
Your phone and internet connections also have usage limits, and you’re also using them in various ways.
I agree that it’s extremely likely that, especially post-IPO, monetization will kill the current user experience, which I already hinted at in my previous comment.
Given the growing distrust of ad supported tech, I could see AI remaining as a paid product.
...seems more like a case against Amazon (search) than for AI, then.
Maybe I'm fortunate enough to live someplace where Geizhals[0] exists, but it's been years since I gave up on Amazon altogether. The bad UX is just user hostile and there are many competitively priced retailers with web shops anyway.
[0]: https://geizhals.eu/
These machines are general purpose technologies used by hundreds of millions of people. ChatGPT alone is used by over 900M people every week at least. You can count the technologies with that scale of users in your hand.
You'll never conceive all the sort of uses it could possibly have, much like nobody could ever conceive all the uses the internet had and will have and it would be misguided to think so. As you see, there's like 2 dozen people here telling OP the thing he thought 'No one' could possibly LLMs use for is in-fact seeing some use.
That's the problem. It moves an incredibly amount of power into a small handful of multinationals.
I don't want to live in a fucking world where an AI watches everywhere I go, reads everything I write, listens to everything I say, and makes decisions that affect me with zero appeal or recourse.
Because that's exactly where we are headed as people.
---
As businesses, we are headed to a world where if you don't pay tribute to the AI syndicates, your business will be undiscoverable.
> I can imagine lots of cases where people with specific needs would find benefit from the “AI clothes buying” experience,
That is kind of the idea of serving the long tail. Everyone is unique, and there are a lot of everyones.
That said, I don't get online clothes shopping. The fit is 80% of the product.
Indeed it is, but when you are p95 (at least for height, but not overweight), you'll soon learn that you do not have any other option: common sizes stop growing in length (at least noticeably) usually at XL or even L, so you are looking for specific fits (long, slim) and those are rarely stocked in stores. Sometimes I'll try a model from one brand and buy a different colour online.
But enter online shopping and 14-30 day return windows.
Still, for formal wear (shirts, jackets, suits), I simply stick with made-to-measure and custom tailoring.
I ended up with a T14 Gen 4 and I'm super happy with it.
isn't that what search engines were built for? we've just forgotten how to build a search engine that's not just an ad factory, so instead we're putting an ad factory into our new search engine?
If I know what I'm looking for, Kagi is much easier.
If I don't know what I'm looking for (I have hobbies that involve learning new techniques, and my method for learning a new technique seems to involve getting inspired by short-form videos, which don't come with a glossary of terms or a dictionary of tool names, so I often don't know what I'm looking for) then I can describe it to claude who can usually come up with a name for the thing, some useful advice about it, and where to start looking.
Last time, as an example, was all about enamalling and cloisonne, which was quite a rabbit-hole. And yes, I could search for beginner guides for the thing. But that is going to land me at a YouTube video which has 5 minutes of "hi welcome to my show, hit the like and subscribe" and then 15 minutes of waffle before finally getting to "the thing you want is called a trivet". I can read way, way, faster than people on YouTube can get to the point, so I prefer talking to claude.
If you have a list of specific criteria, search engines are impossibly bad at finding what you’re looking for, but top LLMs do it with ease.
ChatGPT and Claude have been amazing timesavers in my recent tech acquisitions at work, and I find I am able to find better solutions
Chatgpt found me a lot more choices.
I wanted custom lifts for a shoe. Chatgpt found me a local store that did it, I'd been calling around for years asking to no avail.
Chatgpt is really damn good at niche stuff.
I feel like I see a brand new way of saying “something that people don’t really want” on a near daily basis nowadays
- mining the 95th percentile, leveraging the Pareto Principle
- optimizing and ubiquitizing under-optimized paradigms
- pioneering agentic solutions to aggressively expand product frontiers
- innovating high-risk strategies to serve underserved markets
- digging deep into the inner recesses of my being and extracting what's left of my soul through my nostrils
And so much more.
I don’t want to believe LLMs are the future of shopping either, but it’s wrong to dismiss actual successful users with hot air.
Using several AI models to cut through the multidimensional sea of options.
It's not all grim, thia technology can genuinely be helpful.
I previously would have spent this time opening up 4 tabs on three diff hardware store sites, and an additional tab to pull up the relevant car forums for tips and advice. Which I ended up doing anyways, as well as some YouTube videos because I don't trust the results. But it still saved me a ton of time investigating and weighing out options as a decent aggregator of info.
I have a half dozen facebook marketplace searches going. I used to automate craigslist searches before craigslist became irrelevant. It's nothing complicated but "AI searches for me and notifies me" is better than me remembering to look.
Well obviously you shop for clothes, but nowhere like the way people who like clothes shop for clothes.
Finding clothes is about matching the vision in your head. If you’re the type that just buy clothes whatever, this is not a problem that exists in your world.
You know, everyone used to have specific needs in clothing when I was young. Somehow fast fashion advertised that out of us to solve their own supply chain problems.
For the average consumer, clothes fit them.
There is no generalizability from a niche possible here, since the mass market is already served.
Beyond that, even if we limit it to height alone, there are hundreds of millions of people who are much shorter or much taller than average.
Quality is amazing, fit is incredible, and the price is only 20-30% more than off the rack, but the clothes can last a decade+.
Sometimes the ancient solution (meet another person with a measuring tape) is the best one.
I have one shirt where each button hole has contrast stitching around it, an absolute baller of a shirt.
Even if they cost me $200 today it'd be worth it. They last so long and being able to define your exact own personal style feels great.
Love the idea but difficult problem to fix.
I wouldn't buy a deeply-ingrained AI laptop even if you paid me, and even then I'd install Linux on it in a heartbeat.
Or have shopping items be shown on your twin in a simulated fashion shoot on a doppelganger simulation. It should also show movement, situations and vibes.
"Turning Doug Quaid back into Carl Hauser"
Yet on the other hand, I've got a very extensive page of me wearing and using a bunch of different things. (see link in bio) It'd be interesting(?) to have a hundred(?) fans(?) wear what I wear. Some may be my size, most wouldn't be. I don't know how this world would end up. I presume it's about building a sort of "icon kingdom" or mob of Mr. Andersons. It may be utopian if you find the right community.
It's not just about size and fit, but what people may be looking for is vibe, community and vision. The interplay between fashion and sub-culture is not always so clear. People may want belonging and community, but will that sacrifice individuality and freedom-of-thought? Would you rebel until anarchy or to improvement? What's the focus and vision of your life? Times by 1000 and you're impacting the world through a prayer-like scenario.
This is a good one: https://tallslimtees.com/
I also do some made to order from Son of a Tailor and Proper Cloth, but it’s expensive and laborious and slow and risky.
I'm genuinely curious, whatever you're doing sounds cool, but more details beyond the buzzword pitch you'd tell your manager would be welcome on a hard technical site like hn?
(ftr, I'm skeptical of all applications of machine learning, but I keep experimenting with all the various kinds of it, generally with no good result; last real-world useful [to any extent] ml model I tried was BASnet, but whatever you tinkered out sounds cool and if it actually scrapes and filters clothes the way you describe, that'd be quite cool [perhaps even product worthy…?], cuz there are way too many clothes online to look at all of them manually and then esp. on fast fashion sites, there are oftentimes reviews you want to be wary for that indicate low quality products… anyway, that just sounds impossible to automate in my experience, but feel free to prove otherwise)
Was it a technically impressive effort from the prompter? No. Are the tools created in the session somehow a massive technical achievement? No. But was it a very useful result? Yes. It took the kind of task that would likely never get done otherwise, and turned it into the kind of thing that got done on a whim.
Doesn't mean that your laptop needs "AI buttons" though.
(I mean honestly the project idea[?] they posted sounds like daydreaming some science fiction scenarios, otherwise with all the hype and investment around chat bots, this way of shopping would definitely be mainstream already. If it weren't daydreams, that is. But if my grandma had wheels, she would've been a bicycle, no…?)
What was done is more like using the LLM as a personal assistant that doing long manual labor to find what you might be looking for.
This way of shopping is already a thing. "Hype and investment" goes into how the companies can monetize AI harder (ads! integrated LLM shopping! business development! premium pro max enterprise data policies!), it doesn't really focus much on how the individual can save time and money through non-flashy tasks.
Well, that assumes descriptions are extremely accurate down to the last seam, which is not true. You'd be better off considering reviews and photos, esp. user provided photos, you also need to take into account the model/s in the photos are not necessarily shaped the same as you, so you need to somehow counter that bias in training. This is simply not a task achievable with current ml techniques, however again, feel free to prove otherwise.
(and ftr, I'm of course making a basic assumption that we're past the topic of markov chain/'llm' based chat bots at this point? Those are completely irrelevant to the goal of categorizing clothes based on some characteristics [i.e. the so-called 'vibes'])
My friend is probably 5" shorter than me. A small on him would be too long.
So he's always on the hunt for things that fit him properly in both dimensions.
By using AI to filter out results, you don't see ads, upsells, other products, recommendations, reviews, dark patterns, etc.
Just asking a web search / browser-enabled chatbot, as you are now, is already close to the optimally efficient tool for you. Unfortunately, aggregating results from many disparate retailers into one seller-neutral page filtered down to what you uniquely need today is no longer considered optimally efficient by most web retailers. Just like they erected barriers to stop being indexed by unaffiliated shopping aggregators, most large retailers will try to stop automatic aggregation of their current inventory (or lack thereof).
Sadly, we're now in a post-enshittification world where Amazon's learned removing search features like requiring or excluding terms increases revenue and Google's learned giving you the search result you want first reduces ads served.
As far as this laptop is concerned I feel like it's a repeat of that super expensive chrome book that fizzled out because it was basically nerfed by Google unshockingly. As one of the top posters here if they delivered quality hardware, good Linux and solid Google support and even gapps, this would be an absolute win. instead i can only guess what this is unless I missed any real information on the site it's just a metal Chromebook with extra AI?
I don't know how good AI is at these kinds of tasks, but I can tell you that it's not easy manually, especially in some parts of the world where you might have to factor in shipping/return costs.
I wonder if for the next period websites will really try hard to prevent scraping (already happening, in some industries very pervasive, i.e. its impossible to get accurate quote for power) until they realize they can sell much more to people using agents.
Or everything just going to race to the bottom like a manufacturer or distributor since it's so easy to find everything anything you need. Kinda already happening with saas companies loosing value while infrastructure is soaring.
It is deeply annoying we have to do this.
It's a frustrating existence.