this post was submitted on 25 Dec 2023
1882 points (97.8% liked)

People Twitter

5162 readers
2440 users here now

People tweeting stuff. We allow tweets from anyone.

RULES:

  1. Mark NSFW content.
  2. No doxxing people.
  3. Must be a tweet or similar
  4. No bullying or international politcs
  5. Be excellent to each other.

founded 1 year ago
MODERATORS
 
top 50 comments
sorted by: hot top controversial new old
[–] [email protected] 164 points 10 months ago (1 children)

I feel like since that's a very useful product it will not be made available to me.

[–] [email protected] 30 points 10 months ago

It'll be kept within product marketing and, I dunno how, but it would absolutely be used to see what they can raise prices on

[–] [email protected] 155 points 10 months ago (11 children)

I want said AI to be open source and run locally on my computer

[–] [email protected] 38 points 10 months ago (5 children)

It's getting there. In the next few years as hardware gets better and models get more efficient we'll be able to run these systems entirely locally.

I'm already doing it, but I have some higher end hardware.

load more comments (5 replies)
[–] [email protected] 20 points 10 months ago* (last edited 10 months ago) (8 children)

I can run a pretty alright text generation model and the stable diffusion models on my 2016 laptop with two GTX1080m cards. You can try with these tools: Oobabooga textgenUi

Automatic1111 image generation

They might not be the most performant applications but they are very easy to use.

load more comments (8 replies)
[–] [email protected] 8 points 10 months ago (6 children)

This technology will be running on your phone within the next few years.

load more comments (6 replies)
[–] [email protected] 8 points 10 months ago

I want mine in an emotive-looking airborne bot like Flubber

load more comments (7 replies)
[–] [email protected] 89 points 10 months ago (6 children)

The bad news is the AI they'll pay for will instead estimate your net worth and the highest price you're likely to pay. They'll then dynamicly change the price of things like groceries to make sure the price they're charging will maximize their profits on any given day. That's the AI you're going to get.

[–] [email protected] 12 points 10 months ago (5 children)

If capitalism brought me into this world, will it also take me out of it?

[–] [email protected] 12 points 10 months ago

Are you subscribed to Existence Premium or just Basic?

load more comments (4 replies)
load more comments (5 replies)
[–] [email protected] 49 points 10 months ago (1 children)

AI could do this. Conventional programming could do it faster and better, even if it was written by AI.

It's an important concept to grasp

[–] [email protected] 15 points 10 months ago (10 children)

Cameras in your fridge and pantry to keep tabs on what you have, computer vision to take inventory, clustering to figure out which goods can be interchanged with which, language modeling applied to a web crawler to identify the best deals, and then some conventional code to aggregate the results into a shopping list

Unless you're assuming that you're gonna be supplied APIs to all the grocery stores which have an incentive to prevent this sort of thing from happening, and also assuming that the end user is willing, able, and reliable enough to scan every barcode of everything they buy

This app basically depends on all the best ai we already have except for image generation

load more comments (10 replies)
[–] [email protected] 42 points 10 months ago (3 children)

I'm sure there are companies who'd love to develop something like this. And collect that information about exactly what groceries you currently have and statistics of how you consume them, so they can sell it to advertisers. Not advertisers that sell these groceries, of course - for these the AI company could just make the AI buy them from suppliers that pay them.

load more comments (3 replies)
[–] [email protected] 40 points 10 months ago (12 children)

I want AI to control traffic lights so that I don't sit stopped through an entire cycle as the only car in a 1 mile radius. Also, if there is just one more car in line, let the light stay green just a couple seconds longer. Imagine the gas and time that could be saved... and frustration.

[–] [email protected] 30 points 10 months ago (2 children)

That's already a thing, though it isn't AI driven. Many intersections have sensors that detect traffic and can change the lights quickly or let them stay green longer if you're approaching it. It's only getting more advanced as time goes on.

https://youtu.be/pTR3Cn5DnHY?si=nXtqHqVMsw47FmQZ

https://youtu.be/P_EmAKRrlBc?si=McCO_tE43FaPkRL9

[–] [email protected] 10 points 10 months ago (3 children)

As always, such systems need infrastructure investment to make them widespread.

load more comments (3 replies)
load more comments (1 replies)
[–] [email protected] 18 points 10 months ago (1 children)

Doesn't need AI, and there are countries that already have a system in place with the same result. Unsurprisingly the places with more focus on pedestrian, cyclist, and public transit infrastructure have the most enjoyable driving experience. All the people that don't want to drive will stop as soon as it is safe and convenient, and all those cars off the road also help with this because the lights will be queued up with fewer cars.

load more comments (1 replies)
[–] [email protected] 6 points 10 months ago (2 children)

That’s been a thing for decades already in a lot of places.

load more comments (2 replies)
load more comments (9 replies)
[–] [email protected] 38 points 10 months ago (1 children)
[–] [email protected] 24 points 10 months ago (2 children)

I mean, when that xkcd was made, that was a hard task. Now identifying a bird in a picture can be done in realtime on a raspberry pi in a weekend project.

The problem in the op isn't really a limitation of AI, it's coming up with an inventory management system that can detect low inventory without being obtrusive in a users life. The rest is just scraping local stores prices and compiling a list with some annealing algo that gets the best price to stops ratio.

[–] [email protected] 17 points 10 months ago (3 children)

It's too bad the actual users will be the grocers price fixing for maximum profit.

load more comments (3 replies)
[–] [email protected] 9 points 10 months ago

I think you focused too much on the details...

AI image manipulation is entirely based in a computer where an image is processed by an algorithm. Grocerybot involves many different systems and crosses the boundary between digital and physical. The intertwined nature of the complexity is what makes it (relatively) difficult to explain.

[–] [email protected] 38 points 10 months ago (1 children)

google used to do this type of stuff then you get SEO shit and in the same way people would try to game the system and ruin it

[–] [email protected] 7 points 10 months ago (1 children)

Aye, this be the problem. As long as there is a profit motive the AI is going to steer you to whatever makes them money, be it whoever works the SEO game or pays for API access.

load more comments (1 replies)
[–] [email protected] 36 points 10 months ago (10 children)

This is surprisingly difficult problem because different people are okay with different brand substitutions. Some people may want the cheapest butter regardless of brand, while others may only buy brand name.

For example my wife is okay with generic chex from some grocery stores but not others, but only likes brand names Cheerios. Walmart, Aldi, and Meijer generic cheese is interchangable, but brand name and Kroger brand cheese isn't acceptable.

Making a software system that can deal with all this is really hard. AI is probably the best bet, but it needs to be able to handle all this complexity to be useable, which is a lot of up front work

[–] [email protected] 17 points 10 months ago (1 children)

As long as the AI has access to their ongoing purchase histories it's actually quite easy to have this for day to day situations.

Where it would have difficulty is unexpected spikes in grocery usage, such as hosting a non-annual party.

In theory, as long as it was fine tuned on aggregate histories it should be decent at identifying spikes (i.e. this person purchased 10x the normal amount of perishables this week, that typically is an outlier and they'll be back to 1x next week), but anticipating the spikes ahead of time is pretty much impossible.

load more comments (1 replies)
load more comments (9 replies)
[–] [email protected] 36 points 10 months ago (2 children)

We were already robbed of the brief value stage of AI, it came out of the gate with a corporate handler and a ™

The internet had a stretch where it was just useful, available and exciting. This does not.

[–] [email protected] 14 points 10 months ago (2 children)

Local models are a thing, and GPT is extremely useful in some cases, even with the corporate handholding. I find the whole space super exciting, personally.

load more comments (2 replies)
load more comments (1 replies)
[–] [email protected] 34 points 10 months ago (1 children)

Wasn't the first part what "smart fridges" were supposed to solve like 10 years ago?

load more comments (1 replies)
[–] [email protected] 22 points 10 months ago (1 children)

So many people in this thread saying "you can already do this if you just do all these extra steps" like avoiding extra work isn't the whole point.

[–] [email protected] 9 points 10 months ago* (last edited 10 months ago) (4 children)

Exactly. But also I'm blown away that most grocery stores don't list inventory and prices on the website. I can only think this is because they don't want to show prices in an attempt to get you to go to the store.

[–] [email protected] 8 points 10 months ago

They absolutely don't want to make automatic comparison shopping that easy. The goal of every grocery store is to get you there with one or two specific good deals they advertise and then have you do the rest of your shopping there because nobody wants to go to a second store and MAYBE get a slightly better deal but also maybe get a worse deal.

load more comments (3 replies)
[–] [email protected] 17 points 10 months ago (2 children)

And also the hottest thing a woman can say:

"It runs on lokal hardware"

[–] [email protected] 11 points 10 months ago

Men want one thing and it’s fucking disgusting.

load more comments (1 replies)
[–] [email protected] 17 points 10 months ago (1 children)

I'm sure Sara is not ready to be served the optimal outcome from a competitive multi-agent simulation. Because when everyone gets that AI, oh boy the local deals on groceries will be fun.

load more comments (1 replies)
[–] [email protected] 12 points 10 months ago

Cant I get both? Here are your weekly projections, sir. You will need to get this list of items at these locations and here is what you would look like as a latin american dictator. Enjoy

[–] [email protected] 12 points 10 months ago

I want both.

[–] [email protected] 11 points 10 months ago (1 children)

There are websites that do things like this. For some time I worked in one. The part missing here is analyzing the user’s pantry.

load more comments (1 replies)
[–] [email protected] 10 points 10 months ago (5 children)

It's not exactly an ai-task, I guess? Like pretty much the only ai-related thing there is to classify stuff in ocr-ed receipts (technically, one can opencv whatever is in the fridge, but I suspect it won't be reliable enough).

[–] [email protected] 28 points 10 months ago (11 children)

Bruh. If AI is being taught to drive cars on the open road then I feel like cameras to detect what's in your fridge is pathetically easy in comparison and very much an AI task

[–] [email protected] 7 points 10 months ago (1 children)

That's how you get weird things like the AI determining that your favorite items are jam, baking soda and whatever you left at the back of your fridge to rot for six months.

It is easy to detect what's in your fridge. We have that today on some smart fridges.

The problem to be solved though is

  • what's in your fridge
  • what's not in your fridge
  • what do you consume vs throw away
  • what do you buy
  • where do you shop
  • what prices are available
  • what's the best way to minimize cost and store trips
  • what's your metric for how to balance that

Of those things, AI is really only helpful for determining the metric for how much money you need to save to add another grocery stop, and knowing that the orange blob is probably baking soda.

Most of the rest of that is manual inputs or relatively basic but tedious programming, and those are the parts that would be the most annoying.
I say this as a person who has repeatedly utterly failed to use https://grocy.info/ because actually recording what you eat vs throw away is painful.

This isn't a great AI problem not because AI can't help, but because the tedious part isn't the part it can help with right now.

load more comments (1 replies)
[–] [email protected] 6 points 10 months ago* (last edited 10 months ago)

Probably would make sense to start with the receipts for what you purchase and aggregate lists from there (pantry, freezer, fridge, etc.).

load more comments (9 replies)
[–] [email protected] 11 points 10 months ago (8 children)

What is an "ai task"? Their task is anything we assign them to do.

load more comments (8 replies)
[–] [email protected] 9 points 10 months ago

AI could potentially do, “write me a python script that scrapes a website for grocery prices and compares them with another” or something.

[–] [email protected] 9 points 10 months ago (3 children)

Why are so many of you trivializing the fact that providing perfectly formatted input data that having set logic figure something out is a VERY different thing than providing a firehose of data and then asking the software to make sense of it? Like have you been paying attention here at all?

load more comments (3 replies)
load more comments (1 replies)
load more comments
view more: next ›