Both of those sound kinda dystopian. Because you just know the first one will start getting gamed by every company from the grocery companies trying to SEO the AI, to the big fossil fuel companies trying to get you to drive your car more.
The same technology can be used for widespread, low-cost, highly convincing misinformation and propaganda campaigns
I can’t wait for the technology to get basic enough where I can roll my own self hosted instance of it without it taking months. Because I can see a way it’s doable without a centralized service to get around that. But for mass consumer level, I can see that becoming true. But this can be applied to every bit of software currently. All of it can be ran by you, if you have time. Hell I’ve got my own cloud (hosted at my home ) music streaming service.
A lot of that is doable now - like, how many grocery stores are even nearby to someone, so writing a custom bit of code to check the website of each, one by one, and looking for previously manually-identified items could be automated.
One major downside is prioritization of large chain stores at the expense of smaller mom & pop ones that don’t maintain a constant inventory system accessible via the web. Someone could even volunteer their time to build them a database backend, but still they’d have to see the value in actually scanning the items every time or else it would quickly fall behind.
You don’t need “AI” for that. All you would need is some standardized APIs for the various shops, and you could easily solve this with computer technology from 20 years ago.
The reality is, though, that there are no such APIs. LLMs on the other hand could be a valid tool for the use case.
It’s not that there’s no API. It’s that there’s probably a different API for every single grocery store. And they make random changes and don’t have public documentation. That’s why we need the AI.
The stores don’t want you to have easy comparable access to their prices.
They’d quite like it if you just came in, saw that the item you wanted is out of stock, and then just buy some shit you didn’t need.
Indeed. LLMs read with the same sort of comprehension that humans have, so if a supermarket makes their website compatible with humans then it’s also compatible with LLMs. We have the same “API”, as it were.
You just need someone to do it. Here in Austria someone did it: https://heisse-preise.io
It’s only in German and most of the prices aren’t from a public API but crawled from different sources.
It’s open source. Nothing except greed is stopping them from providing something like this.
LLMs are not a good tool for processing data like this. They would be good for presenting that data though.
Make an LLM convert the data into a standardized format for your traditional algorithm.
there are no such APIs
Yes there are. You can obtain access to the Kroger API, the Meijer API, the Walmart API, and I’m sure others that I didn’t bother to Google. Failing getting access to the actual APIs, there are tons of web scraper projects that just parse those stores’ websites for product information, and web scrapers are still orders of magnitude more efficient than LLMs.
At the cost of huge amounts of wasted energy and the whole litany of concerns that are always co-morbid with AI, but technically yes they could work for this lol. Ideally we’d have standardized APIs and mandated pricing transparency, but unfortunately we live in a capitalist society where that will literally never happen ever.
We need somebody to wear a 360 camera and go walk every aisle every day. Use image recognition to get the SKU and price from the labels + estimate stock level. Upload the data to an API that’s accessible to all for like $5/month.
Kind of like the Streetview cameras but for spying on actual in store prices.
And it’s a service because AI
And the service costs a subscription fee
And the service quality drops once it saturates the market
And the service now contains ads
And the grocery stores can pay to promote their store when it is not the most affordable option
And now it’s not economically feasible to not use their service
I was working on this with a friend over 10 years ago but the only grocery store that made a decent effort at organizing their website to be scrapeable was Loblaws and all the others had APIs that cost $100,000
Which is one area ML models might (with the right investment) actually be useful. A model trained to look at web pages and relay information from the content visually like we do would be very powerful. The newer ChatGPT models have visual capabilities, I wonder if you could give it a website screen capture and ask it for prices.
Why would you want a model trained on outdated prices? This is not really something LLMs are particularly suited for.
Maybe to crunch historical data, but not for daily comparisons.
Why would the model be trained on outdated prices? I’m not talking about LLMs, but separate model designed to parse visual information - specifically websites - and extract particular elements like prices. My comment about ChataGPT was in reference to the newer models which can relay visual information, I’m not suggesting that would be the right approach for training a new model.
The applications would be broader than just prices - this would allow you to scrape any human-readable website without needing to do bespoke development.
The cheapest way to get groceries in the States has always been do all your grocery shopping in the same store, preferably a discount store like an Aldi, instead of cutting coupons and going to multiple different stores due to the simple fact that the gasoline used for driving around is most likely going to cancel out any saving from shopping around, an unfortunate side effect of America’s car centric infrastructure.
You don’t really need an AI to make this list, plus, I think there are apps that already trying to do exactly that.
However, getting a computer to draw yourself in ridiculous situations (usually with an equally ridiculous number of fingers) is great entertainment.
This kind of small scale optimization is not really the best use case for AI anyway. Considering the actual cost of running that kind of code at a large scale… I’m not convinced the savings are worth it even setting aside the petrol issue.
AI doesn’t need to be in the hands of consumers. It should be a step removed, working behind the scenes to make all those basic foods cheaper before you even go shopping. It should be optimizing supply chains, reducing production costs, and otherwise making us more efficient at a societal level.
Which, well, in some cases it already is. Sadly many companies just use it to optimise their marketing 🙄
going to multiple different stores due to the simple fact that the gasoline used for driving around is most likely going to cancel out any saving from shopping around
I wouldn’t jump to that conclusion. Here in suburbia, there are different stores every couple miles. Figure even a 5-mile detour to go to another store, and that “simple fact” of gasoline used turns out to cost less than a dollar. I save that much on a pair of salad kits by going to one store over another, and it’s really more of a one-mile detour anyway. Plus, there are simply things that one store does better than the other and I like to take advantage of that too.
Seriously. Sale items are often several dollars cheaper per item. It is well worth the time and gas driving to several stores unless they are very far apart, then just roll that into another trip. Some big “what could it cost, 10 dollars?” vibes off that comment.
You also need to factor in opportunity cost or concede that your free time doesn’t have value.
If you value your free time at the same rate that you work hourly, then suddenly it’s very hard to save money by spending more time. If you value free time as overtime equivalent, it gets even worse.
Standard IRS reimbursement rate per mile driven is 67¢ per mile this year, which is essentially the per-mile average cost for driving a car. But like, with this sort of thing everyone has their own personal calculus for what they want to optimize for. Do they want to save as much money as possible? Do they want to have fun while shopping? Do they want to shop as quickly as they can? A lot of people will balance these priorities against each other and come up with a solution that isn’t optimal in any one specific area.