21 comments

[ 3.8 ms ] story [ 65.1 ms ] thread
Slightly unclear what's even being proposed here. One person just takes orders, and the other person makes drinks? Isn't that what happens normally?
Yes, exactly. It's hard for me to hold back from complaining about all the other terrible parts of the paper. But yes, that's the key.

TL;DR:

- if there's only one worker.

- and a worker immediately makes the last order they took.

- people wait less to get their order taken.

- when the worker consumes the entire "needs to order" queue before starting to make anyone's drink.

The formatting and layout of MS Word hurts my eyes.
Not exactly related, but I thought a great use of facial recognition would be for cashiers: as soon as you step in front of the counter, the register shows the cashier your name (or maybe top likely names), and what you ordered last time. With this information, the cashier could press just a single button for your most likely order (making them faster), and refer to you by name for a personalized experience.
I would be very pissed if they did this to me. Luckily it's illegal where I live.
(comment deleted)
I would be very pissed if they did this to me. Luckily it's illegal where I live.
Gives me Minority Report vibes. Which…probably isn’t good.
you literally scan your palm to enter and exit amazon fresh stores while ML powered cameras watch then shop and people do it without blinking an eye
Not all people, though I admit that it’s a lot more than makes sense to me.
Yeah that’s crazy to me as well
Less creepy would be to just have a RFID chip in a vendor card, like a loyalty card, that is scanned upon admission by RFID readers
A moderately experienced cashier can input your order faster than you can speak it. That part is absolutely not a bottleneck.

Also in general people actively do not like being called by their first name when they did not give their first name. This is a broad but not universal social norm and I don't see any benefit in pushing to change it. I certainly don't think retail food service is the correct front for it.

This is the line where we need to figure out as a society what we want technology to replace or not.

A good barista is already doing this in their head, and also building and maintaining a human relationship with those customers.

Part of the appeal of a good coffee shop to is that it feels like a friendly place to spend time with people you know, and it makes it feel special when you've got that one place where everyone knows who you are and how you like your coffee. Coffee being an important daily ritual for many builds this shared understanding and empathy around the craft - I'm here because you know this specific subset of me intimately and I can get something valuable that isn't just "the screen says your name is X and your drink is Y is that right?"

You and I have very different relationships with our local coffee spots.

I'm fine with "screen says X and drink is Y". That's sort've the most I'm after when I go to get a coffee.

This sounds like the "warm sound of vinyl records", or "pleasantly musty smell of paper books", or "comforting flicker of candle light".

I.e. not wrong -- but probably just another future anachronism.

I am also confused about what this paper is describing. My read is that having two staff (a dedicated barista and a dedicated cashier) outperforms having one staff (working as both barista and cashier). This seems like a trivial result, am I missing something?
Yeah, that seems to be the gist of the entire paper. It's a two year old submission as well, so a bit odd overall.
IIRC from what I've read about queue theory, it's not that having two staff halves the time, but that it can be multiple factors faster. Though I haven't read this particular paper yet.
Open your Starbucks app when you decide to go to the store and initiate your (re) order. It considers your travel time and the Starbucks queue and schedules your order to be ready just in time for your arrival (at the front of the line), while indicating how long (if at all) you will need to wait.