Hacker News new | ask | show | jobs
by gregpilling 2995 days ago
from the article: "In a small town, there are likely a few small restaurants run by a few people who do many things—cook, clean, manage the books, etc. “Some of these tasks are easily enough defined to soon be automatable,” Youn says.

By contrast, in a larger city there will likely be some much larger restaurants that require more specialized knowledge and skills—perhaps a marketing team, or a lawyer who specializes in the restaurant industry—that cannot be easily automated."

I have a small factory, we use a fair bit of automation. I disagree with the above quote - I think that the larger restaurant will be more likely to automate first, and arguably the franchise restaurant model is automated marketing and business management. The larger places will go full robot in the kitchen sooner, the smaller places will continue to have one person wearing many hats for a long time. Larger places have more budget, can afford more automation, and the smaller places would have a harder time justifying a burger cooking robot if it would only be in use a small fraction of the time.

My welding robots are super fast. I have two so that I can reduce setup and changeover time. The robots are so fast, the main headache is feeding them enough jobs - selling the work. The setup and changeover is very slow. I spend 5-10 hours getting a new jig built, programmed and tuned up.

This setup time is never discussed in the media. Sure the robot will be amazing and fast and super productive - once you spent 10 hours programming it to do so. If the task is a 'once-only' and it only takes an hour, who in their right mind would automate that? This setup time will be the biggest impediment - humans are pretty quickly adapted, autonomous fleshy robots, that outperform the best automation when it comes to setup time.

I love automation. I have 6 automated machines, and a couple 3D printers, that you can push a button and go get a coffee, and there is the part when its done. Sometimes one button push can make a couple thousand dollars in products - that is a great feeling! But the setup time is massive, and this is the greater expense over the cost of the machine.

Its like an algebra equation. y=ax+b where 'a' is the setup time, and 'b' is the time for one unit.

With a human welding a simple part, 'a' would be 5 minutes, and 'b' would be 5 minutes.

With a robot welding cell like a Panasonic PA-750, the 'a' would be 500 minutes, and the 'b' would be 2 minutes.

If you need 100 of them, no sense in turning on the robot. If you need 10,000 of them, the robot welding cell will be done in 20,500 minutes instead of 50,005 minutes for the human welder. My Panasonic costs $2284 a month for 60 months on a lease, $1 to purchase at the end. The human costs $1000 per week in perpetuity, with increasing costs due to inflation if nothing else. The robot is half the hourly, and its paid for in 5 years. The human is never paid for, it always needs compensation. In this example, the human takes 500 hours more to do the task, and that is $12,500 more in wages over the robot. My 7 year old son can run (not program) the welding robot and make parts, which he does for Pokemon money on the weekend.

So where does that leave the human? Doing small runs! Which is better for the human anyway! Welding the same small part for weeks at a time makes you a little crazy. Its like watching the same GIF every 5 minutes for weeks at a time. Welding the same part every 5 minutes for 2-4 hours is quite doable, can be fun to get in the groove. I think in the future that production runs will become smaller as CNC machines proliferate. Everyone will want something custom, which is what humans are best at. High volume will be robot welded, because its boring and machines do it faster and better.

I manufacture in America. Using automation, I am directly cost competitive with China and India in my core products that we have been making (and optimizing manufacturing) for 14 years. Its 1/4 steel plate mostly, and our cost to build is directly in line with the best quotes from Asia. Robot welders are cheaper than any welder in Asia. We loose on thinner stuff, like 16 gauge and lower, and we lose on items that need a lot of hand work. But we are right there for heavier items. We did a large welding job for Chinese factory last year that was for the US market. Our price was the same, but shipping was less since the product was already here. We got the deal over their own factory. 28,000 units.

TL;DR - things are better for humans than these articles would have you believe.

2 comments

And you are underestimating the crowd that will prefer to buy "human made" and not "cheap automated robot crap"

Not saying they will be right or wrong, but that's the current dichotomy with US Made / Non US made goods.

There was also the hand made rolls Royce cars and the other robot assisted manufactured luxury cars. Which one suffered from poor quality perception (for good reason)? Automation can reduce errors and shoddiness.
There will of course be plenty of hold-outs, but most "robot made" will actually be better than human-made.
Sure, for most things. I am finding this not to be the case with things like furniture, clothing, small finer metalworks that require something like a blacksmith, food, soap, etc.
Human made soap is better?
For the purposes of being a Veblen good, yes.
Ok, soap removed.
I am not underestimating them, I am planning on it. All my capital investments are in automated equipment, that I know how to operate myself, and that are flexible in their use - they can make things for many industries, out of many materials.

The future is high mix short run manufacturing. Just in Time. Anything that has to be shipped and not flown, is never going to be 'just-in-time'. Locally made, custom just for you. This does not mean that automation won't be involved. The automation is the core of it - without the miracle of CNC (praise be!) the local guy wouldn't stand a chance.

My customers don't seem to care if it is robot welded or not, or made in the US or not. I doubt 99% of them could tell me which parts are robot welded and which ones are manually welded. Everyone claims that they care, but the top sellers in my segment are all made in Asia. So the evidence does not match the claims. The customer just cares if its what they want. Thats it. We use the "long tail" strategy in our market segment to differentiate with offshore competitors. We offer 3x the part numbers, and 2x the vehicle coverage, plus we do custom sizes and the others do not.

Automation, and flexible use machines, let me change styles to follow trends, and get low labor costs without resorting to offshore. We can customize parts one a one-by-one basis for clients if needed, and we do so usually once or twice a day.

I view my factory like a breakfast restaurant. I know I am going to sell a bunch of eggs, bacon, potatoes, toast - but I don't know exactly how much, or in what configuration. I have a menu, and I know what is roughly the most popular on any given day, but I never know exactly what the customer will want, and it doesn't matter since we make it fresh for them right when they order. I keep the raw materials in stock, we build about half our units on a fixed schedule based on statistical high movers, and the other half is built to order from the orders that day. There is minimal WIP on the floor, I try to keep it under 2 weeks.

Given the choice, people want to get things personalized just for them, or to have something a little unique. How many t-shirts do you see people wear on any given day, and how many are the exact same? Almost none, yet the basic form is the same for all.

TL;DR - things will be fine. Humans will have jobs. Have robots, hire humans.

That crowd will change what they consider important as they have overtime. At one time people resisted the car.
Great comment. Minor point: you switched the a and the b