You can't automate to a level to replace them. You can automate order taking, but the human order taker who does it also cleans the store, fills and packages orders, leaves the store to clean the outside and dumps trash, answers customer questions about food, troubleshoots register issues, balances their drawer, and more.

And something as simple as checking straws and napkins or making sure the floor by the soda machine is clean can't be automated cheaply. Automation in general is too fragile to handle the demands of that kind of work.

These are all environments designed for humans. Vending machines already get warm food, and fresh coffee to me. The system once automated will not look just like the current human paradigm.

There's a lot to operating a physical establishment that machines can't do without reducing the experience. Something as simple as cleaning the tables will not be automated. You need people to provide service in a restaurant, which is why automats generally never take on.

And we already have a model of what it will look like, laundromats. That's what people are aiming for, the patron doing a lot of the work with maybe one person to oversee and provide special services like dry cleaning. I don't think that would be an improvement in food service.

Agreed. Watch "The Founder". The amount of workplace efficiency brought by the McDonald brothers would have been, and almost certainly was considered inconceivable by other restaurateurs of the era.

"The system" was so good that it revolutionized the fast food market to the point that almost every attempt to buck the trend is now considered some sort of synonym to 'craft' or 'artisanal'.

(Note: I'm taking the movie as canon, which is basically stupid, but I have a grandfather who would gladly tell you the story of his first McDonald's visit, and how revolutionary it was relative to everywhere else he'd been, but for his passage a few years back.)

This will change just like one of Philips "lights-out" factories have.

Automation is already being explored essentially everywhere, so I'm not sure motivation is a big part of it. The question is more whether it's feasible, both technically and economically, in a given case. In principle, yes, a minimum wage helps the economics of automation by making the "competition" more expensive. But with the kinds of modest wage changes typically proposed, it's a pretty marginal impact: +10% to wages might move up automation that was going to happen anyway by 6 or 12 months, but it's not the kind of change that meaningfully impacts the progress of automation on a larger scale.

Do you believe there's a somewhat fixed wage threshold (for now or 20 years later) that when it's passed McDonalds begins to consider replacing workers with self-service machines?

Or that the self-service machines are going to arrive regardless of the wage, with the only difference being McDonalds keeping workers on $10/h wage 3-5 years longer than they would keep the $15/h workers?

What happens when the entire self-serving machinery costs as much as it does to pay one or two employees for a year? Does it make sense for McDonalds to keep human employees at any wage?

It does if it produces more sales. If people stop coming because they dislike the machines 15/hour can be a small cost.

On the other hand if people like the machines more they might raise prices.

>To zeroth order, that seems reasonable. But the added expenses might shift the cost-benefit analysis enough to add motivation to explore automation

Automat restaurants have existed since the 1890s and were first touted as a cost cutting measure.

Ironically enough, they were most popular around the 1930s and disappeared roughly when the minimum wage was highest (1970s) and are now making something of a comeback, when the minimum wage is again, historically very low.