Ed Rensi, former CEO of McDonalds, says a higher minimum wage will bring more robots into fast food making.
it’s cheaper to buy a $35,000 robotic arm than it is to hire an employee who’s inefficient making $15 an hour bagging French fries
Rensi also says ordering Kiosks are already deployed in higher wage areas of Europe and are in limited deployment in the United States. Panera has ordering Kiosks in about 20% of their American restaurants and expect roll-out to all restaurants in a few years. In Japan Pizza Hut is rolling out robot order takers. So we do not even need a high minimum wage for mass replacement of restaurant labor.
The current McDonalds CEO, Steve Easterbrook, claims he does not expect mass labor replacement with robots when minimum wage goes to $15 per hour. But this sounds more aimed at maintaining morale than an objective assessment. The moves by Pizza Hut, Panera, and other restaurants suggest that even without higher minimum wage restaurant automation is happening and will likely accelerate The cost effectiveness of robots is rising every year. A higher minimum wage will incentivize the faster development of better restaurant-automation robots. This becomes easier to do every year due to continuous advances in computer chips and computer software.
Chinese computer equipment manufacturer Foxconn has recently replaced 60k out of 110k workers in a factory with robots. Automating factories is easier than automating restaurants. But as Ed Rensi points out, robots for restaurants are already on the market. Falling robot costs and government-mandated wage increases look set to accelerate their spread. This will ultimately lead to your ability to electronically send a recipe to a restaurant to get exactly what you want from a robotic chef.
Update: Whole Foods has just rolled out a new format of store in Los Angeles which uses less than half as many workers as existing stores. It uses a number of techniques to cut labor including kiosks. Whole Foods is preparing for the rise of California minimum wage to $15 per hour. Minimum wage increases and robotics are going to cause a huge slash in retail store and restaurant employment. I expect this to cut the employment-population ratio of the least cognitively able. Their lives are sad.
A San Francisco start-up named Otto seeks to bring self-driving trucks to market before self-driving cars. Their argument is that the ROI for autonomous trucks is higher because each truck has such a high price ($150k) and high distance driven per year. Makes sense.
What I found interesting: the idea that autonomous vehicles will hit small towns especially hard. Long haul truckers live in smaller towns with lower real estate costs. They do not need to be near major cities since when they go to work they travel long distances. So when the over 3 million long haul truckers in America lose their jobs to autonomous trucks they'll have to move closer to cities to find work. This will cost them much of their equity in their houses (who is going to move in to replace them?) and raise their living costs.
Who will lose their jobs first? Uber drivers? Or long haul trucking drivers? I am thinking it is the long haul truckers.
I assert: Genetic engineering of human offspring will start some time in the 2020s. It will take off sharply in the 2030s. The rapid rate of advance of CRISPR gene editing technology combined with the rapid rate of decline of genome sequencing costs will make this possible. People will jump on offspring genetic engineering because they want kids that will excel at something. The range of desired areas of excellence is quite large (ballet dancing, novel writing, software development, painting, music composition, musical instrument playing, investing, managing, etc). The challenge: what to choose for the genetic endowment of your kids.
The desire for super kids will lead to frustration for many, even if their gene selections do exactly what's advertised.
Why frustration? Consider making babies capable of excelling at sports. Imagine basketball fans who want to have a kid who excels in the NBA (ditto other pro sports such as golf, tennis, football, and baseball). The people who decide to make a super basketball player will be frustrated by the results because others will be doing the same. Sports is made up mainly of losers. That won't change even as the level of play rises. Will someone in the late 2020s be able to select genes that make a basketball player better than Michael Jordan? Sure, get his DNA, identify what makes him great, fix all the harmful mutations, and add some mutations that other great basketball players have that he doesn't have. But lots of people will do that. Making a sports winner really isn't the best strategy for prospective parents. There are just too few winners in sports and way way more losers. Offspring genetic engineering will not change that.
What's worse: careers in sports will get shorter. Suppose you choose the best known genetic variants to make a sports star baby in 2028. Well, every year the babies will get better. Kids born 2032 will grow up to be better athletes than your baby born in 2029. Each year the best known gene choices will get better and people making a baby a few years later will make better players than you did. So 15-20 years latter increasingly better players will show up to compete every year, knocking off the previous best. Winners won't stay winners for long.
Humans compete in other domains such as in business. We will see the same effect of rising performance in other domains as well. But in many other endeavors much larger groups of people collaborate with a single company employing thousands of people to develop a product or service. The result is not quite so much winner take all. Someone making a baby in 2028 or 2035 should think about making a high performing baby for one of these areas involving collaboration in larger groups.
Big internet, pharmaceutical, chemicals, and manufacturing companies employ tens of thousands of engineers and scientists. They want the best. But they can't replace all the 25 and 26 year olds with 23 and 24 year olds. They need too many people to do that. They also hire people who are quite a bit less able than the best because they need too many to restrict their selection to only the best. So producing the best future engineer baby in 2028 seems like a safer bet than producing the best pro sports baby.
You can think about other domains of human endeavor the same way. Ballet dancers have short careers. Their careers will get even shorter when the genetic quality of ballet dancers hitting the market is rising every year. Even if you want to produce a baby with great ballet (or gymnastics) potential better make sure they've got the mental chops to get into medical school or engineering school too.
So what's my advice for the baby maker of 2035? First, you really need to do offspring genetic engineering or your kid is going to be a loser. Accept the need. Know that your wide type DNA isn't going to cut it in the labor market of 50-60-70 years from now.
Second, go to the expense to eliminate every pure harmful genetic variant you (and possibly our mate or donor) have from a cell line. We all have hundreds of purely harmful genetic variants. Then think about niches that are not appealing to others. What could your kid be great at that most people don't want to do? That's a really tough question because who knows what demand the labor market will have in 2060 or 2070? Still, its the right question to ask. Avoid the most obvious choices where the crowds will stampede.
What makes specialization especially difficult: intellectual abilities seem closely linked. All the components measured in general intelligence tests are positively correlated. But it is likely to be the case that some genetic variants will be found to confer specific abilities and that there will be some trade-offs between abilities. Look for trade-offs that few will find appealing that result in aptitudes in areas that most prospective parents won't consider. Off the top of my head I think of higher functionality in weightless conditions or perhaps neurological attributes that would make it easier to directly interface your nervous system to a computer.
What to look for: work is unappealing to the vast majority of humans but pays really well and is enjoyable by a small number of people. I think of forensic scientists who investigate crime scenes and
Starting with an initial cohort of 15 obese and 18 lean Labrador retrievers, Raffan and her colleagues selected three obesity-related genes to examine, all of which were known to affect weight in humans. This first analysis turned up a variation in a gene called POMC. In more of the obese dogs, a section of DNA was scrambled at the end of the gene. The deletion is predicted to hinder a dog's ability to produce the neuropeptides β-MSH and β-Endorphin, which are usually involved in switching off hunger after a meal.
POMC isn't the only gene contributing to canine obesity. Others are still waiting to be discovered.
In a larger sample of 310 Labrador retrievers, Raffan and her colleagues discovered a host of canine behaviors associated with the POMC deletion. Not all Labs with the DNA variation were obese (and some were obese without having the mutation), but in general the deletion was associated with greater weight and, according to an owner survey, affected dogs were more food-motivated--they begged their owners for food more frequently, paid more attention at mealtimes, and scavenged for scraps more often. On average, the POMC deletion was associated with a 2 kg weight increase.
Use of food to train assistance dogs selects for assistance dogs that are perpetually hungry. Seems like a problem.
Notably, the POMC deletion was markedly more common in the 81 assistance Labrador retrievers included in the study, occurring in 76 percent of these dogs. "We had no initial reason to believe that the assistance dogs would be a different cohort," says Raffan. "It was surprising. It's possible that these dogs are more food-motivated and therefore more likely to be selected for assistance-dog breeding programs, which historically train using food rewards."
Once all the obesity-boosting genetic variants become known and genetic editing technology matures some people will do genetic engineering to produce puppies without these genes. I also expect lots of genetic load (purely harmful) mutations to be weeded out. A dog 30 years from now will be vastly genetically superior to a dog today.
I am concerned about genetic improvements to cats more than to dogs. Humans have moved cats all over the world and this has made cats into predators in numerous areas where they did not use to exist. Domestic cats much more heavily breed with wild cats. So genetic improvements to domestic cats will leak out into wild cats, making them far more capable of killing birds and other wildlife globally.
"Our results show that, while hurricanes can cause flooding and destroy city infrastructure, there are two sides to the story," said Barros, the James L. Meriam Professor of Civil and Environmental Engineering at Duke University. "The other side is that hurricanes recharge the aquifers and have an enormous impact on photosynthesis and taking up carbon from the atmosphere."
So if we harden our infrastructure to reduce the damage from hurricanes we can get the benefits without much damage.
Tropical cyclones are a needed antidote to droughts.
"There are a lot of regional effects competing with large worldwide changes that make it very hard to predict what climate change will bring to the southeastern United States," said Barros. "If droughts do become worse and we don't have these regular tropical cyclones, the impact will be very negative. And regardless of climate change, our results are yet one more very good reason to protect these vast forests."
Bad news is often good news if you just know how to look at it and benefit from it.
Lelieveld and his colleagues have investigated how temperatures will develop in the Middle East and North Africa over the course of the 21st century. The result is deeply alarming: Even if Earth's temperature were to increase on average only by two degrees Celsius compared to pre-industrial times, the temperature in summer in these regions will increase more than twofold. By mid-century, during the warmest periods, temperatures will not fall below 30 degrees at night, and during daytime they could rise to 46 degrees Celsius (approximately 114 degrees Fahrenheit). By the end of the century, midday temperatures on hot days could even climb to 50 degrees Celsius (approximately 122 degrees Fahrenheit). Another finding: Heat waves could occur ten times more often than they do now. By mid-century, 80 instead of 16 extremely hot days In addition, the duration of heat waves in North Africa and the Middle East will prolong dramatically. Between 1986 and 2005, it was very hot for an average period of about 16 days, by mid-century it will be unusually hot for 80 days per year. At the end of the century, up to 118 days could be unusually hot, even if greenhouse gas emissions decline again after 2040.
What I wonder: How hard would it be to boost the albedo of a large enough portion of the Arabian peninsula or North Africa to substantially lower temperatures?
Alternatively, picture massive underground cities deep enough to be cool whatever the temperature on the surface. The cities would be powered by solar panels.
Another idea: how about using the power from solar panels along coast lines to spray sea water up into the atmosphere to cause could formation and rain? Could this cause substantial cooling?
Will the affected countries just crumble? Or will they get enough resources together to engage in climate engineering and avoid collapse of their societies?
Your self-driving car could decide you have a bad job, send your resume and work history to an AI at a different company,and negotiate a new job for you. The first you'll find out about it is when, on the way to work, the car will take an unexpected turn, drive you to a different building in a different town, and a robot will roll out to greet you as you get out of the car. The robot will tell you what your new job is as you follow it into your new office building or factory.
Then there's the new autonomous vehicle blind date. You get in the car, it pulls up in front of an apartment building, and someone else gets told they have a car waiting for them to take them out for the evening. The other person gets in, you seem them. They see you. Your aural implants inform both of you that you are great matches for each other, and the car drives off.
Moving will be interesting as well. The AI running your life will, unbeknownst to you, decide you ought to move, choose a new house, negotiate terms. Then when you ride off to work one morning then right after you are out of sight trucks full of robots roll up to your place, go into your home, and start moving your stuff out. The first you will find out is when your ride home takes an unexpected turn and take you to a different house.
Vacations will be a similar surprise. You won't know you are going on vacation until your car picks you up from work and unexpectedly drives you to the airport. You won't have to get your luggage out of the trunk of course since a robot will already have delivered it to the airport while you worked. The robot concierge at the airport will recognize you and direct you to the correct departure terminal. You won't need to be searched for bombs or weapons because computers will have watched you since you god dressed and they'll know you aren't carrying anything. Plus, they'll know the location and purpose of everything in your home.
Also, when an old friend comes to visit of course it'll be a surprise to both of you.
Offspring genetic selection for IQ will come sooner than widespread adoption of electric cars. Razib thinks we'll know enough about the genetic structure of intelligence to start doing embryo selection for higher intelligence by 2020.
We will see people accept genetic research results by their own choices when they come to reproduce. Says Razib: "Revealed preferences do dictate history at the end of the day".
Say in the year 2021 we can select embryos for higher IQ. What will be the cost of testing a few dozen genomes well enough to select the ideal embryo? g
Razib says the rate of advance of CRISPR genetic editing on embryos will be practical in 5 years: "CRISPR is getting better and better and better and better". The ability to fix some bad genetic load mutations in an embryo would make embryo selection much easier. Got great stuff on one of your chromosomes mixed in with a few bad things? Fix the bad things. Give your kid only positives.
Parenthetically, after one recent series of dentist visits I've been taking notice of people's teeth. Some people have perfect gleaming white teeth with no cavities. Others like me are not so lucky. Lots of people like me would choose to give their kids better tooth genes if they had the choice. Other people have back problems or skin problems or food sensitivities that make their lives difficult. Why pass that along to your offspring if you can avoid it?
If (or, rather, when) we get more really smart people we will also get more super smart people like John von Neumann. Razib is a little scared at the idea of a substantial fraction of the population being as smart as John von Neumann (one of the smartest people to ever live). By 2030 people might be making babies of his intellectual caliber. So imagine these babies grow up and in 2055 a lot of active scientist operating at von Neumann's level, making many original contributions to many fields. What happens to the world then?
Razib thinks the Chinese will pressure the West to go faster on CRISPR and other genetic technology. Otherwise our geniuses will be far fewer and less able than their geniuses. Resistance will be overcome. I think the affluent will travel to other legal jurisdictions to start pregnancies if they can't get the biotech for super babies in their own country.
Razib thinks we will do genetic editing to create meat animals that are much dumber, unemotional, and pain free to reduce concerns about causing animals to suffer.
There's another aspect of cognitive differences and genetic engineering that Miller and Khan did not touch upon: cognitive attributes heavily influenced by genes that are not about intelligence level. My concern is that 50+ years from now we could have clashes between populations because different populations went in different directions with genetic choices that influence moral and social behavior. We could have groups of brilliant people with strong innate differences in moral reasoning style battling each other for domination within and between societies with an intensity that makes the last century's conflicts seem mild by comparison.