Skip Navigation
27 comments

  • Pretty sure McD's already tried this, with disastrous results.

  • If the menu is limited, I could see this working, the problem is how will it handle special requests?

    I don't do cheese, is Taco Bell AI going to be able to process that request? Because every damn thing there has cheese. The humans do a decent enough job, not sure about AI.

  • I mean, good.

    This is a trivial task, so AI could do it reasonably well and frees up people to do other things. Nobody who is working at McDonald's is dreaming of having to repeat the same script twice per minute for the whole shift.

    And you already know they'll have things like promotional voices so you can get your order taken by the Taco Bell dog or The Hamburgler.

    If you're explicitly anti-AI and also immune to irony then you can use the app on your smartphone to order without AI.

    • Yes, it seems like a reasonable approach. The comments are acting as if they're just going to just throw your order into a generic LLM with a fancy prompt and believe whatever it says, but if they approach this properly, it could be viable.

      Most orders are just from a fixed list of available items. They need a speech model that can identify which of the available menu items are mentioned e.g. "two tacos and one french fries" to make sure that the AI doesn't hallucinate anything the customer didn't mention, a LLM-based part which parses the grammatical structure and relates words to their context (the counting word "two" refers to the amount of tacos; the context suggest that the tacos are part of what the customer wants added to their order), and finally a check that errors if the customer's order does not fall in one of the regularly seen order patterns and thus might be too complicated for the LLM to understand, in which case a human has to check the order.

      Most orders are simple and will fall in one of the same few patterns and could be handled by an AI. If 95% of the orders can be understood by an AI and a human needs to be involved for the last 5%, then I think that counts as societal progress.

      [Edit: oops, I didn't notice what community this is.]

      • I stopped at speech recognition. That's the only important part of this that needs to involve any complex AI. The rest is basic programming and doesn't need a neural net at all. Think of a modern phone tree. There's some things that will get recognized as menu items, and anything beyond gets dumped to the human still monitoring the activity.

        Good luck with the first part though. Having worked drive thru in my days (long ago, but nothing much has changed) the noise level on the input is all over the place. The human ear is very good at picking out things, so usually you can piece together what the order was, but even today's phone trees that I mentioned or even a smartphone that's tied to Google/Siri in real time can screw basic words up, and that's from an algorithm that learns the user's voice, not a different person with a different accent/ volume, car sound, etc. each time.

        Also let me add - even though the human ear is far superior at picking out nuances in a high noise environment, many people working the drive thru still suck at understanding even clear speech. To get AI/LLM/whatever past even the basic incompetent human order taker will be a monumental accomplishment that could filter into so many other things. Read that as sarcastic - it hasn't happened yet, it won't happen via a drive thru cost replacement to increase bottom line profits.

      • [Edit: oops, I didn't notice what community this is.]

        They ain't the boss of me.

        Machine learning has uses, including LLMs and diffusion models.

        Capitalist abuse AI and fuck them when they do. But that's completely different than trying to fight against an entire branch of human knowledge and technology.

27 comments