SoylentNews is people
SoylentNews
Pundits will debate the wellsprings of Donald Trump's election triumph for years. Right now, cultural explanations are in the lead. Multiple researchers and journalists are stressing the role of "racial resentments" and xenophobia as the deepest sources of Trump's appeal. And such explanations cannot be dismissed.
But the decades-long decline of U.S. manufacturing employment and the highly automated nature of the sector's recent revitalization should also be high on the list of explanations. The former is an unmistakable source of the working class rage that helped get Trump elected. The latter is the main reason Trump won't be able to "make America great again" by bringing back production jobs.
The Rust Belt epicenter of the Trump electoral map says a lot about its emotional origins, but so do the facts of employment and productivity in U.S. manufacturing industries. The collapse of labor-intensive commodity manufacturing in recent decades and the expansion in this decade of super-productive advanced manufacturing have left millions of working-class white people feeling abandoned, irrelevant, and angry.
To see this, one has only to look at the stark trend lines of the production data, which show a massive 30-year decline of employment beginning in 1980. That trend led to the liquidation of more than a third of U.S. manufacturing positions. Employment in the sector plunged from 18.9 million jobs to 12.2 million.
[...] In fact, the total inflation-adjusted output of the U.S. manufacturing sector is now higher than it has ever been. That's true even as the sector's employment is growing only slowly, and remains near the lowest it's been. These diverging lines—which reflect the sector's improved productivity—highlight a huge problem with Trump's promises to help workers by reshoring millions of manufacturing jobs. America is already producing a lot. And in any event, the return of more manufacturing won't bring back many jobs because the labor is increasingly being done by robots.
(Score: 3, Informative) by DECbot on Monday November 21 2016, @04:15PM
Robots are useful, yes, but they can't and don't do everything.
Precisely! Take welding robots for example. Sure they can go faster and have more repeatablity. However, they are very inflexible for doing anything beyond the assigned task. When given a bin of bad parts, a manual welder will attempt two or three parts with a bad fit-up before inspecting the bin and complaining to the production manager that stamping has a problem. A robot will happily make terrible welds all day at twice the production rate of the manual welder and never complain of the bad fit-up. Thereby wasting an entire day's worth of production and generate perhaps thousands of dollars of scrap.
American manufacturing should absolutely add robots where it makes sense, and use people where it makes sense. In my example, there normally is an operator that would be responsible for loading and unloading parts from that section of the production line that could identify bad robotic welds. This is the common difference between American operators and operators and foreign manufacturing facilities, the Americans tend to care about quality at the lowest levels and will raise awareness to problems. In third world and developing world manufacturing plants, that level of ownership is not as prevalent and the operator cares little more than the robot.
So, yes, robots will take many of skilled laborers positions, but they cannot do everything. 20 years ago, 100 manufacturing jobs went to $NOT_USA, and if they come back now, there will be 80 robots and 30 laborers. However, the point is even 30 jobs added back to the US is better for the US than 0 jobs getting added. The more of these positions that return to the US, the more opportunities there are available for people without MBAs or Harvard degrees. This is a win for anyone with just a high school diploma or a technical degree.
cats~$ sudo chown -R us /home/base
(Score: 2) by Pino P on Monday November 21 2016, @04:47PM
When given a bin of bad parts, a manual welder will attempt two or three parts with a bad fit-up before inspecting the bin and complaining to the production manager that stamping has a problem. A robot will happily make terrible welds all day at twice the production rate of the manual welder and never complain of the bad fit-up.
You have discovered a bug in the robot's programming: "evaluate fit-up" is missing. Add it. But I admit that I'm unfamiliar with welding. Is this something that's impractical to measure automatically?
(Score: 2) by DECbot on Monday November 21 2016, @05:55PM
Impractical as it is expensive and time consuming (hurts cycle time). Also, it tends not to work well on shiny surfaces, like aluminum and stainless steel--you know, the stuff you happen to be welding. The biggest this is the human welder can change their path automatically to accommodate bad or less than consistent fit-up. Robotic algorithms to measure and calculate a path based on measuring a joint's fit-up are still iffy. More so on any parts that require traveling around a curve. Also, if the gap between the parts vary by more than a millimeter, you may need to adjust your power, wire feed speed, or travel speed to accommodate. Stuff that experienced welders do really well on the fly but are still impossible for robots to do.
cats~$ sudo chown -R us /home/base
(Score: 0) by Anonymous Coward on Monday November 21 2016, @04:48PM
You think the companies using robots are not doing quality control?
(Score: 2) by DECbot on Monday November 21 2016, @06:14PM
Except for aerospace, aggressive quality control may put in QC checks for the quality of the welds every 25 parts. Otherwise, it is just a simple 'is there a weld here check'--and these checks are still done by a person. Additionally, production continues making bad welds while the operator discusses the bad welds to QC.
Anyway, my argument was never that these checks weren't done, but rather that the (1) the quality of the quality control checks vary by workforce possibly allowing large volumes of bad parts to be produced and (2) both the QC supervisor and the operator making the checks are jobs that could be in the US if the manufacturing returned. Sure, there will be less welding and machinist positions than previously, but there will be additional operator, production supervisor, engineering, technician, procurement, and HR type jobs if the manufacturing returned to the US. From my discussions with engineers working with foreign plants, they are installing robots in those places because the workforce doesn't have any competent welders and they are looking for big data solutions for weld monitoring because can't trust the quality control measures in place.
cats~$ sudo chown -R us /home/base