I have written a few posts on self-driving automobiles and on my belief that fulling self-driving autos are decades away. I recently discovered a few authors who agree with me that getting 95% of the way to a self-driving car will be rapid, but the last 5% will take a long time. In some respects, one could argue that we are almost at 95% today, but we will certainly still be at this point in 5 years or so. But, without the last 5%, a truly self-driving auto is well, not self-driving. So, the thought of our roads being ubiquitously covered with cars with no steering wheels and sleeping passengers in the back seat is still close to science fiction. As pointed out before, Shladover’s article in Scientific American is likely the best-balanced view of the state of the art in this field. He believes that fully autonomous vehicles will not arrive until 2045 at the earliest.
Expanding on this concept, there have been many articles about robots taking away jobs in some massive way. While, at some level, this concern is real, again, I think it is overstated. Stefan Hajkowicz points out, “Spreadsheets didn’t kill off accounting jobs. On the contrary, smart accountants learned how to use spreadsheets to become more productive and more employable.”
Joining Hajkowicz, I think robots and artificial intelligence (AI) will often change jobs, not necessarily eliminate them. In any case, it is an exciting time for those of us in electronics assembly, as we are not only building the robots and AI devices, but are also using them to help manufacture our products.
One can also ask, why have robots been so slowly adopted in electronics assembly? One of the reasons is the adaptability of humans. Consider a modest sized full service assembler. This company assembles complete units in numbers less than 100,000. It has 20 workers on two shifts that perform hand assembly of the finished units. The work is complex enough so that it is not too boring and each person performs a complete assembly. Say they receive an order for 50,000 units. Each worker can assemble 10 an hour, so all of them assemble 3,200 a day. The order takes about 3 weeks to finish. Over the years, management has considered purchasing robots, but each hand assembly is different and the robots either can’t do it or require extensive programming for each new job. It just never has made sense economically. Robots are nowhere near as flexible or adaptable as human at this point in time.
In my next post, I will discuss how close we are to having robots that can routinely handle such assembly tasks.