Huh, I shopped for clothes using AI today.
Not super relevant to the Googlebook ad, but in case the perspective is interesting to you: I'm quite tall (194cm) but not very wide, so I usually struggle with buying clothes online. I used AI to scrape a bunch of clothing stores to see whether they sold a men's shirt with an LT or slim fit size, in stock, and matching a particular vibe.
This just shows how bad search engines have become. About 15 years ago you could type fully worded questions into Google and would be pointed to the exact sentence of a website that answers your question. I happened so slowly, we were all frogs in boiling water.
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.
The “nice” thing with AI is that the nudging can be so subtle you don’t even realize you’re being influenced for money.
What people think commercial AI is: a friend
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 reminds me of an old video about a guy that got invited to stay in the penthouse suites of casinos. In the video, he has a 'friend' who organises these trips for him (the friend works for the casinos).
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"
Google is the first company which developed, demonstrated and earned an award for inserting advertisements in AI generated text.
It even supports bidding for the ad space!
Source: https://research.google/blog/mechanism-design-for-large-lang...
No you couldnt
You're right but I think AIs can be better than Google at it's height.
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.
This is kinda the exception that proves the rule. I can imagine lots of cases where people with specific needs would find benefit from the “AI clothes buying” experience, but I will bet you anything that any searches you try to do will lead you to the same half-dozen giant mail-order clothing vendors that everyone already knows about.
> exception that proves the rule
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.
You're confusing "The Exception That Proves the Rule" (in English, as used colloquially) with "exceptio probat regulam in casibus non exceptis" (in Latin, which has a use similar to what you're describing.) While the law attempts to be precise, common usage embraces ambiguity.
> exceptio probat regulam in casibus non exceptis
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...
RIP modus ponens!
They really mean the same. What changed was the meaning of the word "proves" in English. When the saying was coined it meant "tests", not "confirms". People kept saying the...saying even though they were using it backwards.
Which is amusing because “proving grounds” is still using the old definition of “proves” :)
https://en.wikipedia.org/wiki/Exception_that_proves_the_rule:
> "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.
But we all knew what they meant and here you are being tedious about it
I didn't really know what they meant by it. Sounds like "the fact that you do this proves that nobody does it".
I believe the phrase is used to mean something like "the fact that you found something that is obviously an exception proves that the rule normally applies."
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.
In this example the “exception” that proves the rule though was not a smarty-pants special circumstance. Using AI for shopping is just one of its many normal usages and if anything proves it is used by normal people doing normal things. It’s not like the rare example that happens once in a hundred years.
Ok but I’m not sure the relevance here? Everyone has unique needs, if they want to get specific enough. The promise of AI here is that anyone can get as absurdly specific as they want, instead of accepting whatever advertising bucket they’d be traditionally sorted into.
Fair enough, I can buy that. I feel like in most cases where I've heard it it wasn't nearly so clear cut, so that logic wasn't obvious and it sounded like nonsense
(Also that story is nuts! https://wikipedia.org/wiki/Vesna_Vulovi%C4%87)
This exactly.
People on this website are so fucking pedantic and argumentative over the most obvious or inconsequential minutiae it drives me nuts.
Comparing the tone of your message with its parent, do you not see the irony?
Yeah, better to be passive aggressive and faux-intellectual so as not to throw off the usual HN vibe.
Okay, I'll be directly aggressive. They were helpful and added thoughtful discussion, while your response added nothing but bitterness
That they were responding to someone else who couldn't see how it worked means it wasn't just pedantic nitpicking
I concur. I found this discussion on the phrase very enlightening and aided both my understanding of the idiom and how I should use it in the future.
I'm not really 100% certain this is the correct or only meaning, for what it's worth, so don't take me as authority. It's just the common thread I've been able to gather from context over time. If you're gonna use it (I rarely do) it'd be worth researching it to make sure you're using it correctly...
[flagged]
>It's not personal.
I didn't say it was personal, I said it was irritating.
It's a common rhetorical technique that never convinces me. I'm not convinced this guy's anecdote is an exception to some rule.
I knew what he meant and still thought it spawned an interesting discussion. Mainly because I've never quite intuitively understood that saying. So, I did not take it as OP being tedious about it at all.
That’s why we’re all here together friend.
It’s not an exception. People seem very short sighted when it comes to AI. Unable to think outside the box for how AI can be helpful or useful.
Yeah, I use AI for this stuff all the time. Found a visa agency, accountant, great cafes to be working, etc just in the past week.
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)
How many of the things you've listed here are $20/month better than a search engine? That's the actual deal here.
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?
I’m already paying for it, this is just one of the many ways I’m using it.
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.
I’m not entirely convinced that they will be able to monetize that effectively with ads. Right now I can buy more chat than I can use in a month for $10 in API credits via commodity open model providers.
Given the growing distrust of ad supported tech, I could see AI remaining as a paid product.
Kagi is a better (arguably) search engine without ads. Unmetered search is available for $10, and also includes an AI search tool.
> I use it to sift through Amazon.
...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/
Don't you think that's backwards from how utility usually works? Most effective solutions come from attempting to solve a known problem, not by searching for problems to apply an available solution. Even thinking outside the box is usually in service of a particular problem - just applying creative or unorthodox solutions to that problem.
You're thinking about it the wrong way. Have you never come across some successful business idea and go, 'Huh, I never realized this problem even existed' or even 'People are paying this much for this? Wow'
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.
I don't think people doubt what AI _could_ do, they just have been through enough enshitification cycles to know this is not any different. Right now AI is better than Google but only because Google regressed so much. Market forces always prevail. The operating costs are just too high to offer AI for free for everyone but people will refuse to pay, so AI (at least for the masses) will become just an other marketing funnel companies can buy out. I also don't see how AI will change the fact that clothing companies target average users and don't serve the long tail.
Yeah it's a helpful and useful tool. It's the people who use it in annoying ways and marketing pushing it too much. It's natural for people to think like that. It's strange being there. Isn't exception a word that describes it well?
Twitter would make you think it’s dooms day with AI yet there are pockets of “wow that’s incredible”
It can be useful.
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.
ChatGPT has helped me find multiple niche products and vendors. It is really good at that. Products I fruitlessly tried to find for years, ChatGPT found right away.
> 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.
> 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 bought a used laptop with the help of ChatGPT last month and was amazed. It helped me narrow the model that suited my needs based on my prompts. I needed to renew my old Thinkpad T480. It also helped me find an ad and negociate with the seller.
I ended up with a T14 Gen 4 and I'm super happy with it.
> ChatGPT has helped me find multiple niche products and vendors. It is really good at that. Products I fruitlessly tried to find for years, ChatGPT found right away.
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?
ChatGPT and similar are, in some sense, a semantic web search engine combined with an operator that's able to jot down its findings, pivot to different lookups, and filter/combine outputs.
I use Kagi to search, and claude to help me find things. These are different tasks.
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.
I wonder if LLMs will actually kill Youtube for those who are like you (and me)? I am curious to see if anything happens to earnings from Youtube over the next few years as people increasingly do not need to sit through whole videos.
It is what they are supposed to do but they are SEO’d so bad that they give you prominent brands rather or least common denominator type stuff.
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
I needed some extra wide toe box shoes. Search results had been SEOed to hell so I could only find a handful of brands (that I'd already tried).
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.
>That is kind of the idea of serving the long tail.
I feel like I see a brand new way of saying “something that people don’t really want” on a near daily basis nowadays
It means Google will show you the top 5 brands for a product category and then give up. If you want something more specific you have to search through reddit threads. Or you can have chatgpt search through reddit threads for you.
No, you don't understand. We're...
- 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.
You and the parent are dismissing an actual customer who really used the “AI” successfully.
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.
Yeah, people are letting their biases get in the way of seeing the reality.
I also did a bunch of shopping with AI to identify clothing recently. I was going to DC for a bunch of meetings, and did not have a good sense of what clothes are appropriate in different DC contexts. I did a bunch of iteration with AI to identify something that communicated what I intended, and then ran the final list by a friend with more context to confirm that it was indeed a readable choice.
Just in the last three weeks I cut buying an used car analysis (1-3 months usually) and a new dryer (usually at least a week) to three days total -- this is "time to shortlist" aka "any of the three remaining options will be a great choice".
Using several AI models to cut through the multidimensional sea of options.
It's not all grim, thia technology can genuinely be helpful.
AI helped me shop for some bits and tools that I needed to do my rear differential and brake fluid, and after some nudging, I also got it to do price comparisons for the tools I needed. saved me a lot of time to walk into each store with an exact list on the bits that I needed. And time with getting exactly the tool I needed without overspending.
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.
That's fine? Especially if the AI does the searches for me, and does them more frequently than I would.
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.
Uhh I don’t think you shop for clothes if you think there’s just a half dozen giant mail order clothing vendors.
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.
> I can imagine lots of cases where people with specific needs
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.
Great ideas always start with a niche. Moreover, they almost always start out looking dumb.
and therefore, ideas that seem dumb and niche must be great
The example was of someone who is tall and doesn’t find clothes that fit them.
For the average consumer, clothes fit them.
There is no generalizability from a niche possible here, since the mass market is already served.
Actually, I think you're really underestimating how many people struggle to find clothes that fit them well. Most people's bodies are not perfectly average. For clothes not to fit well, you don't need to be a significant outlier in one dimension. Just the accumulation of several smaller deviations from average can be enough to create an awkward fit.
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.
I can assure you having observed the process of clothes shopping for the women in my life, that as far as they are concerned, clothes do not just “fit them”.
Yeah there was no way Amazon could expand from its niche of specialist books to anything else, because that market was already served. Oh, wait...
I will do my best to remember your opinion next time I use AI to buy clothes, but given the fact that IDGAF, it's not going to be easy.
Yeah, that’s how AI should be used. If the ad was using AI as a tool to solve a real problem then I’d be down. But that’s not what this is. This is AI as a shopping cart, or a thing to organize the busy life of a casually rich person who flies to Japan to buy vintage clothes. Basically I’m only saying the ad is wildly out of touch with reality.
What made you believe it is not someone else's real problem? You're simply not targeted audience of the ads
It’s not enough people's problem to make it worth building a mass market advertising campaign around it.
even for advertising it's not even a problem to aspire toward having
[flagged]
There have been several startups focused on helping consumers find clothes that fit properly due to lack of consistent sizing between brands (or dress size "inflation" for women). Some of these used optical or laser scanners, or asked consumers to measure themselves. I think they're all dead or on life support now, but it still feels like there's a profitable business opportunity in there somewhere?
There are multiple Thai tailors that fly around to major US cities. They'll take your measurements, and then sit down and design a bunch of custom clothes for you.
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.
You can also get a local tailor to take your measurements and then order direct from tailors in Thailand or Singapore, and it tends to be cheaper than off the rack.
What is the ballpark, 20%-30% of a $20 shirt, or 20%-30% of a $200 shirt?
It has been a long time since I used the services, the shirts I got a decade ago are still good, but it was like $130 for a really nice shirt where I got to customize everything including the stitching and the buttons.
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.
Link?
Here is the guy I used last time
http://www.ravisehgal.com/
For dress clothes at least, Maxwell's is great. I have some of their shirts and a jacket and they all fit and feel great. They do tours and measure you in a hotel conference room:
https://www.maxwellsclothiers.com/
Big issue that also seems to unfortunately be more and more common is variations in sizing within the same brand and article of clothing! Different batches with minor variations of the same exact size, or sizes changing over time.
Love the idea but difficult problem to fix.
Quite possibly! But Google Gemini, who obtains the specs from the same flawed, inconsistent, contradictory, or absent size charts that I have to look at, is not positioned to be the solution to this problem.
I don’t think there is. People who care will go out and try the clothes on in the fitting room or just order online and return. That’s a much nicer experience and more foolproof.
Heh when I came to this country I was overjoyed to see that they had a "Big & Tall" store. Until I realized they actually meant the conjunction there...
I use chatGPT to track my nutrition goals, and adjust exercises. I also let it code review my personal projects to (at worst) gain exposure to new patterns.
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.
Researched men's sneakers last night. Super conflicting TMI for my odd size so going to a store for human sizing and gait evaluation. Info on durability was complete garbage. Suspicious about tuning for favored brands but AI recommended shoes will have the edge in my purchase decision since I've done some research.
Did it work? Did you buy something?
Undecided. One of the sites didn't have an LT but the LLM flagged that chest dimensions on their large were narrower than others, so could be worth trying.
That you didn't just search for "Big and Tall" or some-such tells me the search engines aren't as decent as I might think. Are search engines really that useless? Or did you start with a search for "clothes, tall" and then use AI to scrape the hits to find more details?
I did that first, which didn't work for a couple of reasons. Many "big and tall" stores are the intersection only, and many of the stores didn't quite have the right vibe. I do have some of my wardrobe from places like 2tall.com but I was looking for something very silly for an in-joke for a friends vacation.
Yes, you used it but in a way not even remotely close to how they envision you should use it.
But I did this with Google before the LLMs
I recently used AI to shop for clothes. A T-shirt I liked and wanted more of had doubled in price due to tariffs, while some shorts that needed replacement had been discontinued. AI helped identify alternatives with comparable fit and fabric that were respectively domestic and available, which would have been a much bigger hassle pre-AI.
Did it find Girav [1]?
[1] https://www.girav.com/
What did you use a s what was the procedure. I want to find pants similar to ones i got.
Imagine you could take a few photos of yourself and a system would find your real-life doppelgangers around the world. Then see what they wear and easily copy them. They get a commission.
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"
This sounds like a terrible idea. Why not just have a simulated fashion shoot of yourself rather than requiring a database of the entire populations likeness to find your doppelganger? Very dystopian
I mentioned a "simulated fashion shoot" .. maybe after you posted the initial comment?
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.
My brother in size Small-Tall, tell me where you shop!
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.
Dunno if it's the best brand in terms of bang for your buck, but I've bought a lot of shirts from "Have it Tall" on amazon and I have zero complaints about the fit. 6'4" and a pretty average build.
Hi. What AI and procedure did you use for this? I am also looking for good formal clothes that fit my broad shoulders but narrow waist than typical mass market clothes shoot for
A good friend of mine has the opposite problem. I'm 5'11" (180cm) with a slender frame and long arms. A small Patagonia jacket fits me great.
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.
It used to be so local (regional) brands had sizes adapted to their demographics. There used to be a thing like Italian, French, German, Scandinavian, etc sizing. I guess for global brands like Patagonia it's going to be challenging to fit everybody into the same – let's say – "M" size.
You scrape sites, okay, but what's "ai" got to do with that (I assume ai means chat bots in this context?)
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)
What "ai" got to do with that would be that he didn't write a scraper and a clothing style ("vibe") categorizer to build a database to process entries in to pick a shop. They just prompted the "ai" (I really don't know why you're putting that in quotes), and it in turn did that for them.
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.
Ah so what they meant was like a 'vibe coded' a scraper? I thought they meant something like turning descriptions/reviews/photos of clothes into embeddings, as in like sentiment classification but way beyond that? Because the latter would be somewhat cool if it's actually achievable (I doubt it is tho…)
(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…?)
You could turn clothing descriptions into embeddings and have a fashion vector database, but doing that would mostly just net you the ability to find adjacent clothings, rather than the ability to navigate available clothing or clothing fitting certain requirements.
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.
> You could turn clothing descriptions into embeddings and have a fashion vector database
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'])
good luck scrapping very specific information in a bunch of dynamic websites before AI.
Okay but then 'AI' is just a noise generator when it comes to very specific information… I mean just try asking any chat bot to search for something like specific photography gear for some specific scenario and in my experience it's just as good as simply picking some random stuff, except the chat bot will also gaslight you into thinking you made the right choices, so you don't question them… :/
What were your results? I'm nearly the exact same height with a shorter torso than leg length but super long arms, so I tend to need a medium tall, 36" inseam pants.
You realize that if anything, this way of using AI for shopping is something corporate would rather hate?
By using AI to filter out results, you don't see ads, upsells, other products, recommendations, reviews, dark patterns, etc.
Wow, I’m in the same boat - do you mind sharing more about how you did it? I was thinking about that too (I’m 197cm) and would love to learn!
I asked Gemini for help with something similar recently and it just made up a bunch of stores and items. When I pointed it out it said sorry and that it won't do it again. Then it did it again.
It's called Gemini because there's two of them responding to you. One that tells the lie, and the other to apologize for telling the lie (which, of course, is just a different lie).
I use AI regularly to consider new looks. Just have it render someone like me in different outfits. Super useful.
Sounds interesting.. can that be done easily?
Hilariously I've done something similar for the same reason. Medium shirts/sweaters are generally too short on me but large sizes feel baggy. I only travel occasionally to the US for work, so last trip I had ChatGPU look at several US-based retailers (eg Land's End, LL Bean, American Tall) to see if there was stuff in stock I might want to have shipped to my office/hotel.
Just curious, did you check the stores’ sites afterwards for false positives or negatives? eg, “no this store doesn’t have anything for you” but it did?
Yes it was a mix of LLM and manual searching.
Your valid use case doesn't contradict the point that so far most consumer-focused "AI features" are rarely useful and often just get in the way. I'm pretty sure a specific "AI Shopping Feature" wouldn't actually do what you're already doing, or if it did, it would add more steps/distractions than you have now.
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.
I feel the real problem is poor standardized sizing for clothing in store and worse online. I swear every store has their own unique sizes and when it comes to no names on sites like Amazon it's just pure good luck.
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?
Hacker News demographics isn’t real!
I mean, same on the struggle as a tall person, but doing that kind of research is pretty easy even without AI. Just find a couple brands that fit and some shops that sell them and you're pretty much set. I buy almost all my t-shirts from a specific company for tall people now, that I found on amazon by typing in something like "tall t-shirts"
I’m sorry but I’m not buying this argument. This problem was solved in 2005 with search engines.
Search engines are good for finding shops, but not individual items. There also aren't many shops that are dedicated to only tall, usually big and tall instead.
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 used AI to scrape a bunch of clothing stores
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.
I'm a similar build and I could imagine using ai for this purpose but come on. 194cm is like top 1% of human height. It's not a solid business model.
As 204cm human, I have also built a thing to scrape all the major brands for LT sizes.
It is deeply annoying we have to do this.
6'4" with relatively proportionate body-parts: buy the tall/long and you're likely good. At your height, all bets are off!
At my height, I have to do custom on a lot of things, though LT sizes can work for some pieces (short sleeves anything, some long sleeve items if the cut is intended to have longer sleeves).
It's a frustrating existence.
Can you share it?! Will save me the trouble …
AI is good for shopping today because all other platforms are fully enshitified but AI is still in the pre-enshitification phase. It will be infested with ads soon enough. Enjoy it while you can.
[dead]