'Eat the future, pay with your face': my dystopian trip to an AI burger joint | Hacker News Reader

'Eat the future, pay with your face': my dystopian trip to an AI burger joint

theguardian.com

29 pointsBrajeshwar45 comments

I worked for years at a KFC, all the way up to assistant manager. It was actually quite a challenging job, far harder than my senior software engineering role I have now. You have to manage at least 10 different products, and about 10 other variables such as labor and food cost and speed of service in real time, all while typically running a position yourself, and doing inventory in your down time.

That said, the actual physical making of the food is a small and trivial part of any operation of a fast food restaurant and this robotic help would be laughed out of any kitchen I was in.

Until the robots can do all of the tasks above and quickly switch from one to another and be highly adaptable in real time based on the needs of the restaurant, I think we’re okay.

This tells me more the founders of this startup should have gone to work in a fast food restaurant before thinking they are smart enough to innovate it. This feels like ivory tower innovation at best.

Per the article, their original goal was to man a grill station but they realised it was far too difficult. Still, manning a deep fryer is possible apparently.

Dropping a fry bag is such a tiny part of what you do in a restaurant it’s almost a joke.

> Until the robots can do all of the tasks above and quickly switch from one to another and be highly adaptable in real time based on the needs of the restaurant

Ideally if your robots cost less than $20/hour you can buy enough of them to do one job each.

Physical space limitations may prevent you from cramming too many machines inside the restaurant.

I worked the kitchen at McDonald's for about six months once, and I think a better application of AI in this case would be to keep track of the number of customers queued up versus the amount and kinds of food already staged, then tell people what to cook and when. Simple automation like a light above the deep fryer that means "drop a batch of nuggets now" would be a game changer.

Another way to say that is as a young man I had loads of manual dexterity but zero experience running an assembly line with multiple products and unpredictable surges in demand. Use the AI to patch what's actually lacking.

I know that Chick Fil-A does exactly this, based on an engineering post of theirs.

Yeah, reading the posts above it's like the people above have never worked in logistics where all those things have been a solved problem for 40 years with a database and heuristics.

The absurdity of the fry robot is that they’re automating around cooking stations designed for humans, rather than modifying the cooking stations for automation. These are consistent processes that can be automated with repetition - conveyors and timers and such. We manage to fry millions of KGs of chips (crisps) around the world, and none use robotic arms to gently shuffle a fry basket in a small vat of oil.

Seems like a “performance” is baked into the price. You’re right, surely there are better machine optimized methods of making fries.

Automate the process, then scale up... doing both at the same time would probably cost more.

Then why don't you eat a tv dinner instead of getting McDonald's? That's what you're talking about.

Parts of fast casual food really isn’t that different to a TV dinner already. Starbucks will literally just warm your breakfast sandwich in front of you using a combo-microwave. McDonald’s uses microwaves and warming ovens for certain items, that are literally just single prepackaged portions straight from the freezer.

Even sit down chain restaurants like Applebees and Olive Garden will use microwaves and poaching to bring preprepared items to serving temperature. Not exactly TV dinners, but closer to that than from-scratch cooking.

Because you don't have to clean up after yourself.

I'm reminded of the "donut robot" that a local place breaks out to fry apple cider donuts in the fall.

It's a ring-shaped nozzle with a hopper full of batter above a long but shallow vat of hot oil with a conveyor belt running through it. Halfway through, the belt hits a bump thst flips the donuts. At the end they land in a pile of cinnamon and sugar.

I'm sure the wiz-bang inefficient fry robot is just to attract attention. We could have built a machine to make fries a hundred years ago. But now a robot is doing it with AI. Cue the ooohs and ahhhs.

> But manning a grill – keeping track of burgers and cheese and buns and onions, and being able to flip the different objects at the proper time – turned out to be a tremendously sophisticated robotics problem, one too tricky for the startup to tackle

I suppose if my software job gets automated out of existence, I can go to work flipping burgers.

10+ years ago when I worked in fancy kitchens we would every once in a while get a white collar guy (always a guy) interested in changing careers to something they could be "passionate" about. I saw it work a couple times but never with tech people. We stopped even letting them try lol. They'd come in expecting to be in a position of authority, expecting us to be grateful for all the changes they'd immediately want to make to how we did things.

To some extent this was also true of the other professionals. But they'd usually shake it once they got humbled by seeing up close how fast a good cook can actually shuck perfect oysters or just calmly handle the intensity of the line during dinner. Never saw a programmer adapt to it.

One of my former coworkers is now a high-end chef after he abruptly quit the company I was working at. He's married now and he seems happy.

It's cause he made all his money in tech first. It's a hard life otherwise.

I never worked in a kitchen I'd remotely describe as fancy, but I've certainly worked a fast food kitchen or two. I know way better than to go to one looking for meaning; it'd strictly be more of a last-ditch paycheck thing for me at this point, and not an option I'd like to be forced into; that's for sure.

Mostly I'm just disappointed that apparently we're determined to make humans toil in order to free up the robots to create art; I rather preferred the opposite vision.

> Mostly I'm just disappointed that apparently we're determined to make humans toil in order to free up the robots to create art; I rather preferred the opposite vision.

Are we "determined" to do that, or did art just turn out to be vastly easier to automate than anything involving the manipulation of actual physical objects?

Didn't McDonald's solve this problem decades ago?

Yeah, apparently Flippy is now only for working a frier. The article says it can use intelligence to properly cook different things like wings or onions. The fast food way of doing this would just be to put a timer over the frier with a preset and a picture of a wing/fry/whatever on it that you press when you drop the basket in.

I'm sure it could save some labor, but you know the thing probably costs a hundred thousand dollars and also comes with a SaaS contract that's twice what a worker would cost already. And you still need a worker to get stuff out of the freezer and load up whatever hopper the machine is working out of.

Fun fact: McDonalds cooks don’t even flip burgers. The process involves a grill press which cooks both sides at once. It’s less work in half the time!

Almost all cooking steps in a McDonalds kitchen involves a timer and some form of automation. The humans run the overall process and final assembly, but automation - microwaves, warming ovens, toaster belts, auto-rise fry baskets, grill timers - runs all the repetitive and standardized cooking sub-processes.

What is most interesting to me about automation in the restaurant industry, is its almost complete absence in the United States.

If—as I understand it—labor costs dominate the cost structure of all restaurants, then why isn’t the average restaurant already mostly robots?

Is it really a technology issue? We haven’t been able to build a burger making robot until today? Seems unlikely.

Is it financial? The capital deployable to build a single restaurant can’t cover the initial costs of the robot? Seems like a McDonalds could spread the costs over thousands of restaurants.

So what gives?

Wages for fast food workers compared to those of the people ordering the food.

In most countries the wage spread between minimum wage and average wage is a lot lower than the one in the US. It makes no sense to automate something that your customers are happy to pay for.

I once witnessed a noodle making bot absolutely YEET a plate of spaghetti with sauce at the wall of its box. It went everywhere. Quite an amazing sight to behold. Also watching the humans having to clean up afterwards.

So what you're saying is: Develop a noodle cleaning bot before you work on your noodle cooking bot? ;)

I really like the word "yeet". 10 years old this month I think [0]. I really need to look up the origin sometime

0 - https://trends.google.com/trends/explore?date=all&q=yeet&hl=...

What's old[1] is new again. I guess with an AI to take your order, the last piece has been solved. Only took 60 years.

[1] https://www.youtube.com/watch?v=FmXLqImT1wE

Why do jounos treat mundane incremental automation progress as dystopian, while casting the dystopian tech paradigm shifts as mundane?

I'm not seeing anything here that would actually require anything like AI, this looks like old school industrial robotics, a little machine vision, and some basic scheduling - can anybody tell me what I'm missing here, or is it really all things that could been done 30+ years ago without issue?

> After I left CaliExpress, I found myself driving to the closest McDonald’s, where I ordered another cheeseburger prepared by other humans. It was smaller, and the ingredients were clearly cheaper, but I found myself savoring the well-honed flavors of this drive-through classic.

I wonder if the author eats at McDonald's regularly or just reserves her visits after eating at automated fast food vendors she's reviewing.

In all seriousness, these kinds of articles fetishize human workers and read more like activist pieces arguing against tech. The tech journalist that hates tech is a phenomenon I noticed a while ago. I wrote about this before but here are 5 of the latest articles from a nyt tech journalist

There’s Nothing to Like in Facebook’s Plans to Hook Our Kids

The Maps That Steer Us Wrong

Apple’s Illusion of Privacy Is Getting Harder to Sell

Why Tesla’s ‘Beta Testing’ Puts the Public at Risk

The Assault on Our Privacy Is Being Conducted in Private

The articles can be summarized as Facebook bad, digital maps bad, Apple bad, Tesla bad and big-tech bad

What other field has journalists covering it that absolutely hate everything about it? Why bother at this point? That's why YouTube personalities like mkbhd and Doug demuro are so popular because they honestly love the things they're reporting on. Who wants to read the umpteenth article about how meta is selling our data?

https://mleverything.substack.com/p/journalism-shouldnt-be-s...

>What other field has journalists covering it that absolutely hate everything about it?

Politics

Aren't most journalists carrying water for certain politicians? Just look at the article Uri Berliner wrote on how politics affected their reporting. The journalists felt a need to cover up a story because it would negatively impact a favored politician of nearly everyone there

https://www.thefp.com/p/npr-editor-how-npr-lost-americas-tru...

> The articles can be summarized as Facebook bad, digital maps bad, Apple bad, Tesla bad and big-tech bad

Even from the headlines, it’s clear the criticism is more targeted than that. To generalise so aggressively is either bad faith (ironic, no?) or poor reading comprehension.

Are youtube personalities journalists? Is it the job of journalists to be popular? Should it be?

Sure but where's the article about how uber reduced drunk driving, made cabs stop ripping people off and profiling. Or how google maps liberated so many people to feel comfortable exploring new places? Or the million other benefits that these new technologies have. Why is everything negative? Imagine if science reporting didn't report on drug discoveries and just how new drugs are leading the inequity in healthcare outcomes between those that can afford them and those that cant

It's a lot easier to criticize than it is to create.

Food for thought is that it's impossible to tell if you are commenting on the comment, or reinforcing its message.

Perhaps there are societal issues that make it so that advances in technology tend to be hostile to regular people. If that were the case, then it wouldn't be surprising if objective journalism on technological advances skewed negative.

Personally, I think that the issue may be that technological advances are being driven by corporate profits so they don't necessarily prioritize the concerns of average people. This could be exasperated by issues such as corporate consolidation, regulatory capture, network effects, etc. that insulate corporations from the negative effects of upsetting average people.

> fetishize human workers and read more like activist pieces arguing against tech.

Agreed. The writing style, including adjective/adverb use, is clearly intended to convey an emotional tone of disgust and displeasure.

I'd love to give these people a chance to review a car from the 1900s or 1910s. They'd undoubtedly walk away convinced that the newfangled 'automobile' was a heap of rubbish destined for the dump.

(Of course the successes we have today are built on top of endless heaps of rubbish and a few good ideas.)

"Pay for it with your face?" At a burger joint?!

As an occasional connoisseur of Nickelodeon's You Can't Do That on Television back in the day, I can only state that I hear the obvious punchline / catchphrase resounding loudly in my head: https://youtu.be/cATh91BYjjA?si=D6X97xDwZwff3izz

To file under “oh wait that actually wasn’t AI”, as the journalist herself subtly points out.

The bit about computer vision and being able to adapt to a new kitchen without reprogramming or to a human employee moving things around is interesting.

Wouldn't you want your robot precisely designed to the spec of your kitchen as it throws hot oil around, rather than figuring it out on the fly with computer vision? As another commenter pointed out, robots have been frying potatoes at immense scale for decades, without a single LLM or AI in sight. Just "normal" computer programs can do it.

Reply on news.ycombinator.com