Hacker News Reader

Comment by neel8986 | Hacker News Reader

Meta can be a extremely significant player in generative AI. Meta owns few of the largest interface of chat UI through Whatsapp, Messengar and Insta. Inserting a chatbot can hook more people in the eco system. You can book your cab or order your groceries through a simple chat in whatsapp. A primitive version is already there but with introduction of LLM Meta can actually build a app store for whatsapp. Moreover Meta has a data moat which is out of access from Microsoft and Google. It has access to 70% of all human chat conversations in the last 10 years

View on HN

> You can book your cab or order your groceries through a simple chat in whatsapp

Sure, if you want to use a terrible user interface. An app or website can provide the same thing much faster.

I don't know, I hate installing apps on my phone. I also hate texting strangers. I put a lot of care into my written communication, and many people in the world are genuinely bad at comprehension.

If I could type a few vague sentences into a box describing what I want, and then have it reliably converted into an unambiguous summary for me to adjust or approve, that would be amazing:

"I would like a gallon of lactose free milk, a pound of fresh chicken breast, a dozen eggs that are on sale, and a bag of frozen hash browns. I need it by 4:30."

5 seconds later up pops a website with a shopping list with product photos, an estimated delivery time with appropriate fee, and an order button linked to Google Pay.

I don’t think an LLM is required for the UX you want. Online grocery delivery already exists. They could easily offer a list-based UI where you upload a list and it populates a cart for you.

I order groceries online a lot. My personal opinion about why this UX wouldn’t work well is that people want to feel like they made an optimal choice. Like when you go to the store to buy an apple, do you just pick the first one? Or do you pick the organic variety that looks freshest? Or maybe you go in for an apple buy decide to get grapes because they look better. There are a few staples I always buy, but generally my grocery purchases are more deliberate. From what I see at the store, many people exhibit such behavior.

My father-in-law visited last week to help with the baby and my wife's broken leg. He asked if I wanted anything from the grocery store. I said, "an instant ramen cup, but not the cheapest brand. Also, watch out because they sometimes have beef broth and I'm allergic." He came back with a choice that I would never have picked myself, but it fit all my criteria and perfectly satisfied my sodium craving.

When I'm outsourcing a chore, I'm doing it to exchange money for time and effort. I would prefer to be able to abstract away the details, and trust that my request will be satisfied. In the current UX, I have to specifically request a tub of "Ben and Jerry's Extra Choco Choo Choo Churn", then be disappointed when the desperate minimum-wage worker somehow buys a version with almond milk because it was the closest one in stock. All I wanted was "a $5 pint of chocolate ice cream" to cheer up my depressed wife, and she's allergic to almonds...

Perhaps an LLM could provide that level of abstraction to customers, while reliably compiling it down to a sufficiently concrete list of options for a worker to retrieve. The abstraction would likely improve the UX, as the workers would have more freedom to choose compatible products rather than arbitrarily specific ones. It would be closer to a personal shopper experience virtually built on top of commodity labor, which seems like the whole goal of these grocery services.

> I said, "an instant ramen cup, but not the cheapest brand. Also, watch out because they sometimes have beef broth and I'm allergic."

This strikes at the heart of the issue. The criteria you’ve placed on the ramen is actually pretty complicated, and yet it is simpler than the criteria many use to select produce, dairy, and meat. Needing to meticulously articulate criteria for the 100+ items purchased every month would be a pain. Items also have interdependencies, e.g. I can only afford the fancy yogurt if I got a deal on something else, and context dependent, e.g. I want to spend more on chicken for the fancy dinner, and less for a weeknight meal.

Of course, these criteria are theoretically learnable by an LLM, but extracting the prerequisite information is hard. So much context exists in your mind. For instance, just looking at your ramen purchase history, do buy chicken ramen over beef ramen because you like chicken better, or is it because you actively dislike beef? It’s hard to say without your explicit input. If every purchasing decision requires explicit input, it seems better to cut out the LLM.

The best UX would be to have a personal shopper. That's the point of using human language. I can be explicit about the properties I care about and vague about the properties I don't care about. I'm not going to pay someone $100 to do my grocery shopping or have my father-in-law do it every week, though. An algorithm that can robustly interpret human language can also optimize a shopping cart significantly faster than a human can, based on massively indexed databases.

Not to be rude but to highlight my point, why are you trying to infer my ramen flavor preferences from my purchase history? I already explicitly said that I am allergic to beef. That's the whole point. It takes practically no effort for a literate human to speak or write down the criteria that are important to them. It is significantly harder for a computer to correctly interpret it, but that's why it seems like an LLM could help.

And then the store doesn’t have half of your items so they replace the hash browns with fries and refund the milk.

The Kroger app, at least, lets you specify whether or not substitutions are allowed. Granted, that doesn't solve the problem completely.

On the other hand, the physical store has often been out of certain items over the last couple of years, so I don't think the app is entirely to blame here..

The app's UI and the reality of what the human in the loop does is often inconsistent. In general I would prefer to delegate and give the worker freedom to choose the best option given the unknown constraints. I believe people generally perform better when given more autonomy rather than less. The level of over-specification required to use grocery delivery services at the moment just makes it more stressful than going to the store myself. We've been ordering groceries a lot due to my wife's broken leg, and there's really stupid stuff in every order, like rotten broccoli.

When I do see a worker, they always genuinely seem to be trying their best to find the right thing. I was at a Walgreens at 10pm a couple nights ago, and a poor guy who barely spoke english desperately asked me whether the bag of charcoal briquettes he was holding matched the listing on his phone.

> I hate installing apps on my phone

That is why I do not have Whatsapp, Messengar or Insta in the first place. Using them through web browser is not good experience as well.

You can do that on alexa today. Its not great.

Alexa bugs me everyday to do voice shopping with it, i would be suprised if any significant number of people are voice shopping on alexa.

I also think most ppl don't want to type in a whole bunch of sentences on their phone either.

I was under the impression that the messaging apps (WA, Messenger) use end-to-end encryption for the conversations, in which case how does Meta have "access to 70% of all human chat conversations in the last 10 years"?

My own understanding has always been that Facebook cares not one jot about the content of the messages; it's the metadata (i.e. the social network and its patterns of interaction) that's valuable. (Of course that may change, now.)

They could just push an update that sends your local history back. "Security" and "software updates" are dangerously connected.

"Leaving AT&T. Just dotted the I's, crossed the T's, and every single detail is....perfect."

Indeed, WhatsApp uses e2e encryption. But today's mobile devices are ever more capable with AI accelerators. It's just a matter of time until a locally inferred model is good enough.

“You can book your cab or order your groceries through a simple chat … It has access to 70% of all human chat conversations in the last 10 years.”

I’m feeling like one of those guys in Japan who attacks the customer service robot in the department store—I loath the voice recognition systems behind phone systems for major business (banks, energy suppliers, telcos). I routinely mash #1 until I get to touch tone. I either want to pay a bill or I have a problem the automated system can’t solve—chat bots waste my time & I don’t like talking to friendly robots.

I guess some people are only using their phone and not a website first, in which case a chatbot lets you navigate a much wider set of options by phone. I get it.

Still hasn’t changed my feeling that my mobile phone bot is keeping me chatting just to further train their itself and waste my time. I’m Nexus 1, help me grow as a chatbot.

true, but they've been sitting on their hands as far as actual product improvement, considering their huge resources. Some products have outright regressed with repulsive (and unnecessary) sponsored content and tik-tok style forced feeding. Others are stagnant from the functionality standpoint. Instead of deluding themselves with mass VR, could have done something useful, but didn't, and now someone else is doing it.

Reply on news.ycombinator.com