Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Taco Bell AI Drive-Thru

aidarwinawards.org

Post reply on HN
Print view View original

141–150 of 237 posts

Re: Taco Bell AI Drive-Thru

#141

Post by jgalt212 »

BobbyTables2 wrote: post #33
PapaPalpatine wrote: post #7

Earlier quoted context omitted.

I have never heard someone describe drive-thrus as a “famously bad experience.”

There’s a StarBucks near me that takes about 3-4 minutes per car at the drive thru. Frequently there would be 3-6 cars in line. Yes, people literally wait 15-20 minutes in line before they can even order, much less get their order. Sure, maybe they’re just inefficient and shouldn’t be rewarded. However the people there are indeed working feverishly (and paid poorly). Going inside and ordering isn’t any faster. I’d pu…

People are "famously bad" at correctly valuing their time. I can't tell you how many times I've seen someone spend $x of their time making a business case to purchase something that costs $y, where x > y.

Re: Taco Bell AI Drive-Thru

#142

Post by saagarjha »

sneak wrote: post #23
jameslk wrote: post #15

When the AI gladly accepts orders that cost hundreds of thousands of dollars or 18,000 cups of water, it’s probably not production ready https://youtube.com/shorts/FDZj6DCWlfc https://www.tiktok.com/@90daygrinder/video/75355084374472983... (another example from a different chain)

This is solved easily by one additional sanity check API call to a different AI. I’m not sure why people think these bugs are like, complete showstopper insurmountable things. It’s a quick fix.

AIs are not resilient against deliberate attacks, even if you use multiple different models.

Re: Taco Bell AI Drive-Thru

#143

Post by ares623 »

HarHarVeryFunny wrote: post #136
dangus wrote: post #97

Earlier quoted context omitted.

500 stores isn't really "scale," that's only about 6% of their locations. To be honest, if LLMs are good at anything, this is the exact kind of thing they are good at. It really isn't dumb that Taco Bell tried this. I could also imagine how great it could potentially be for people to be able to view the menu and/or order in any language. I think long ago I actually read an article posted to HN that essentially argued…

If they want to do automated voice ordering, then using multiple choice A/B/C/cancel (with feedback on screen) would seem less error prone than LLM open ended natural language with some kind of intent interpretation. Of course most customers would prefer to interact with a person, but I don't think "vibe ordering" tacos is going to be the same!

Deep down the PMs know that’s just a touch screen but with voice. So no one’s getting a promotion for that.

Re: Taco Bell AI Drive-Thru

#144

Post by euLh7SM5HDFY »

jameslk wrote: post #68
lordnacho wrote: post #64

Why order when you arrive at the place? Just have people talk to their phones, and make sure the order has some sanity checks and orders something similar to what they can order online? There's less noise talking to your phone, and you can do it without being in line.

Some assumptions it seems you’re making: * The customer has a phone * The customer knows how to use the phone to do what you’re expecting * The customer has cell service in their location * The customer has the patience or ability to order before arriving at the restaurant

Sadly many restaurants already make same assumptions with QR code menu replacements. And the worst thing is they keep them despite everybody hating them.

Re: Taco Bell AI Drive-Thru

#145

Post by ares623 »

stephenlindauer wrote: post #29
nickthegreek wrote: post #18

Earlier quoted context omitted.

I’m shocked that anybody with a smartphone is installing tracking fast food mobile apps, creating accounts, giving emails, and even enabling notifications. Another capitalistic hellscape flimsly masquerading as a great deal for you.

Different folks have different levels of acceptable risk. I generally agree, and won’t install an app for every single one-time purchase, but I eat a lot of Taco Bell. The app makes sense for me. Many other companies also have web interfaces so you don’t NEED to install an app. (Many restaurants/bars that have a Qr Code on the table act this way.)

The sad side effect of these apps is that everyone else who don’t have it (by choice or by circumstance) pay even more because they don’t get personalised discounts.

I know because I didn’t know there was a McDonalds app until my brother got me something one time and he paid decently less than what was on the “offline” menu.

It’s a win-win for the companies. They get to extract more money from a majority of their occasional customers, while getting very accurate tracking and behaviour metrics of more dedicated customers.

Re: Taco Bell AI Drive-Thru

#146

Post by dogman1050 »

akerl_ wrote: post #36
sneak wrote: post #22

Earlier quoted context omitted.

You can’t pay in cash via an app. An alternate reality where nobody can transact without the state seeing it in realtime and having a veto over it (without any burden of proof) is not glorious; that’s called a dystopia. Just because the capability has never been leveled at you personally doesn’t mean that’s a world in which you wish to live.

I’m pretty ok with cash ceasing to exist.

Not me. Around here is spreading a 3-4% surcharge when paying with a card. Everything from a sandwich to car repair to rent. I still use cash and checks when necessary. Originally, cards were used to entice customers to a store with the promise of convenience and ease of short term credit. Now we pay for the "convenience."

Re: Taco Bell AI Drive-Thru

#147

Post by grues-dinner »

toofy wrote: post #60
akerl_ wrote: post #16

I’m shocked that anybody with a smartphone is still ordering by talking into a voice box, regardless of whether it’s a human or an AI on the other end. There’s a Starbucks near me that is pickup-only. You mobile order, and inside there’s just a rack where the employees set out drinks as they’re made. Walking inside felt like I’d stepped into a glorious alternate reality.

the number of places that beg me to install an app is ridiculous. if i installed an app from every single place that begs me to my phone would probably shut down in protest from the thousands of apps i’d have cluttered everywhere. from the different grocery stores, restaurant, to every f’ing gas station, every coffee chain, car wash, home depot, and on and on with everywhere we go in between somewhere is begging us t…

Yes, it's exhausting. I mean I get it, having higher prices for people without the app is the same as member cards with swingeing prices for non-members. It's a surcharge on the unaware (including non-native speakers, the elderly, tourists, etc), the unprepared or time-pressured (who don't have time to fiddle with their phones to install the app at a given shop) and, they hope, the people rich enough that they don't want to fuck about with this today.

"Special offers" also really fuck me off. If I want to buy, say, peanut butter today, there's a statistical upcharge because I'm not carefully synchronising my purchase to a secret schedule of when it will cost 20% less. Whenever I see a special offer now, my immediate thought I'd not what it used to be ("wow, that's great"), it is "ugh, these scammers and their constant games".

Charitably, you could call it an elaborate game to make things cheaper for people with less money who can spend the time, privacy and energy (because poor people always have time and energy...?!) to get the best prices. Realistically it seems like mind games to get you used to overpaying most of the time and overbuying some of the time and handing over the data always.

Re: Taco Bell AI Drive-Thru

#148

Post by ssharp »

josephwegner wrote: post #5

I will say, there is a Wendy’s near me that is piloting an AI drive-thru experience, and I prefer it 10-to-1 to the human version. It had a clear voice, it didn’t disappear randomly, it understood what I meant the first time (even though I was speaking naturally - I didn’t know at first it was AI), and it asked me for feedback (“what sort of sauce?”) in a very understandable way. Drive-thrus are famously a bad experi…

I've used the Tacbo Bel AI drive-thru and came away with the same thought. I kind of groaned at first but it was very accurate, even when making adjustments.

Re: Taco Bell AI Drive-Thru

#149

Post by bboygravity »

potato3732842 wrote: post #130
eco wrote: post #14

Earlier quoted context omitted.

I've had two interactions with Wendy's AI drive-through, and the first time I was pleasantly surprised, but the second time it would not stop suggesting add-ons after every single thing I said. It was comically pushy. A human would have pretty quickly picked up on my increasingly exasperated "no, thanks" and stopped doing it, but the AI was completely blind to my growing frustration, following the upsell directive wi…

It's the speed limit problem. Everyone who's detached from reality whether an MBA in HQ or some two bit in the internet comment section who fancies themselves a central planner thinks that the problem is the people on the ground not following "the rules" when in reality "the rules", in just about any situation where there are rules are crap if followed and often themselves are knowingly crap written in response to ot…

So the solution is more AI: to replace c-suit.

Re: Taco Bell AI Drive-Thru

#150

Post by zamadatix »

sneak wrote: post #23
jameslk wrote: post #15

When the AI gladly accepts orders that cost hundreds of thousands of dollars or 18,000 cups of water, it’s probably not production ready https://youtube.com/shorts/FDZj6DCWlfc https://www.tiktok.com/@90daygrinder/video/75355084374472983... (another example from a different chain)

This is solved easily by one additional sanity check API call to a different AI. I’m not sure why people think these bugs are like, complete showstopper insurmountable things. It’s a quick fix.

This would be better off if the LLM was used for the human interface but traditional logic was used for the ordering API and its sanity checks. I.e. let it be fine the LLM can bug out on occasion, but keep rigorous boundaries around the amount of risk that's associated with.
Post reply on HN