and Rime Labs does digital blackface
https://www.youtube.com/watch?v=KSD6-Nf1fg0&list=UU9rJrMVgcXTfa8xuMnbhAEA - video
https://pivottoai.libsyn.com/20250830-taco-bell-rethinks-ai-order-bots-because-they-dont-work - podcast
This is a very good story to boost because it’s one of the most straightforward replacements for a human worker. Like, there’s no fig leaf of “creating tools to enable software developers to-” on it. That means that it’s failures directly undercut the story of how these systems become at all profitable, which is the kind of thing that’s going to actually hurt the bubble.
If I encountered an order AI I would have been like: “Can I get the Tuesday $1.00 Big Box Special?” and tried to haggle with all sorts of reasons until it gave it to me.
Doing drive through for fast food and interpreting a humans garbage input to actual menu output is probably one of the more stressful jobs in the kitchen I work at.
Aren’t most people ordering their fast food through apps nowadays anyway? Isn’t this slightly more customer-friendly than AI order bots because it is at least a deterministic system?
Oh, I forgot, these apps will probably be vibe-coded soon too. Never mind.
Taco Bell’s app would probably work better if it was vibe coded.
Lmao yeah, I mean I haven’t used Taco Bell’s, but so many websites and apps are total garbage.
In my experience, the large self-service kiosks at McDonald’s are pretty decent (unless they crash, which happens too often). Many people (including myself) use them voluntarily, because if it is nice to have more control of and visual information about your order (including prices, product images, nutritional information, allergens etc.). You don’t even need to wait in line anymore if their staff brings your order directly to your table. You don’t need to use any tricks to speak to a human either, because you can always go to the counter and order there instead. However, this only works because the kiosks are customer-friendly enough that you don’t have to force most people to use them.
I know that even those kiosks probably aren’t great in the sense that they may replace some jobs, at least over the short-term. However, if customers truly like something, this might still lead to more demand and thus more jobs in other areas (people who carry your order to your table, people who prepare the food itself, people who code those apps - unless they are truly “vibe-coded”, maintain the kiosks, design their content etc.).
However, the current “breed” of AI bots is a far cry away from even that, in my impression. They are really primarily used as a threat to “uppity” labor, and who cares about the customers?
McDonald’s has the worst kiosks of the three that I have experienced (the other two being Taco Bell and Burger King.)
They feel slightly laggy, while cramming in as many upsell interstitials (and “log in with your
personal data accumulationrewards account” nags) as possible. This makes ordering feel like wading through molasses. The other two could also be slightly streamlined, but the number of clicks to order doesn’t feel as egregious. (I’m now tempted to go count and see if my perceptions are accurate…)Dang, I am missing a lot due to eating junk food instead of fast food.
By the way, I know there is an argument that “low-skilled” jobs should not be eliminated because there are supposedly people who are unable to perform more demanding and varied tasks. But I believe this is partly a myth that was invented as a result of the industrial revolution, because back then, a very large number of people were needed to do such jobs. In addition, this doesn’t even address the fact that many of these jobs require some type of specific skill anyway (which isn’t getting rewarded appropriately, though).
The best example to this day are immigrants who have to do “low-skilled” jobs even though they possess academic degrees from their home countries. In such cases, I believe that automation could even lead to the creation of more jobs that match their true skill levels.
Another problem is that, especially in countries like the US, low-wage jobs are used as a substitute for a reasonable social safety net.
AI (especially large language models) is, of course, a separate issue, because it is claimed that AI could replace highly skilled and creative workers, which, on the one hand, is used as a constant threat and, on the other hand, is not even remotely true according to current experience.
I guess that’s basically one of my problems with this ai craze. It’s the wrong tool for the job. There are way better ways to automate processes ir whatever they are trying to do. Possibly cheaper, though not having tried to purchase an ai system in kit sure.
You point nails it, yeah the self serve kiosks are a good example they work way better. Then the human backup is there too.
At this point I fundamentally disagree that it will replace jobs, the kiosks didn’t. Walmart and Home Depot almost exclusively have self checkout, right? Well what about the cashiers? For one, those stores had like…. 3 cashiers for the entire store so not many are replaced. Further, now they have the employee milling about the self checkout kiosks for whatever. Shrink prevention, assistance because the self checkouts are hilariously dumb etc. the jobs changed but they are still there.
“Would you like code with that?”
“I would like to order extra napkins, and that is all. Thank you.”
Yeah, I’ve just stopped going to Taco Bell entirely because of this crap. Left an angry Google review about it too, for all the good it might do.
My local TB never installed any AI crap. I think it’s only in certain test markets.
The hack to get past the bot at Taco Bell is to ask for 18,000 cups of water. This gets you a human.
To be fair, I interacted with one of these and you can also use the hack of saying “give me a human being” at the start.
only if you hate fun and embarrassing the shit out of Taco Bell
The issues with a ton of these use cases is direct tying to avoid middleware, which is absolutely NOT something one should be doing with AI. At best AI should be used as a speech to text processor that has the optional ability to provide a formatted output, like say JSON. Past that, you are doomed to deal with issues that AI will not handle correctly.
That’s the key issue, once you have to have some logic layer in between, the cost of these things starts flying northward and the savings you’ll get from replacing around 8,000 minimum wage workers is nowhere near what you though you would be saving. So these companies believe that they’ll just bake in the logic into the system prompt. Which, LOL, gets you $15,000 priced combos.
The Taco Bell one is actually about as well as it should handle for a lazy deployment. The issue is that Taco Bell isn’t ready for everyone in the world to be bypassing their AI system and the AI system doesn’t have any good fall backs. There’s just a thin layer that says when something goes outside of parameters to trigger a human being, that’s striking too cautious a balance.
AI isn’t a replacement for humans, it’s no where near that ability at the moment. But if you think about say an IVR system, as long as you are keeping parameters incredibly narrow, you can fit AI somewhere in there. But that rarely justifies the cost, especially if you’ve already got a well developed IVR that you’re still dealing with.
That’s the biggest thing, there’s a lot of promise in these things. But we’re not there. We are not anywhere close to these things replacing 100% humans taking orders. If you’re a place that’s not yet built out an IVR, the AI agent might be something to look into, IF you have programmers on your team that can do the heavy lifting for the logic layers. In short, treat these things like a Small Language Model plus serialization, that’s about as far as they are good at the moment. Trying to toss logic into these things is just going to induce pain and failure along the way. And if that doesn’t save you enough money to justify the super high cost per credit, then don’t do AI. There’s a very small application for AI in public facing things at the moment.
The only thing I’ve seen AI at the moment be really good at is RF signal processing. It wouldn’t surprise me if down the road we saw some sort of model embedded for boosting Wi-Fi signal processing. But this kind of application with Taco Bell et al. No. At least not right now. Or not enough to justify it to most companies at this time.
