4 ms·

For the other 63%, I think an interesting possibility is to use a robot Turk machine type of arrangement. Take for example, the "flag person" at a road constru

For the other 63%, I think an interesting possibility is to use a robot Turk machine type of arrangement. Take for example, the "flag person" at a road construction site, it is an outdoor job and hence would be categorized as not possible to do from home, but, one could easily design a very simple machine with a camera and the stop/slow sign that could be operated remotely. That could actually be safer and lower cost. There are lots of other possible examples, window washer, etc. There is probably some companies working on it already, if not, it might be an interesting idea I think.

PraetorianGourd 5y ago
Have you ever worked as a flagger? I have. It isn't a binary "move the sign and then flip it back". Here are a list of things that you would need to consider: 1. Flaggers often go from site to site throughout the day. You would need someone to set up and take down the machine 2. The flagger is in constant radio contact with the work crew. They need to know which equipment is blocking which lane etc. And this is always changing. 3. Flaggers need to respond to emergency vehicles, busses, their own work trucks etc. You have to have your head on a swivel at all times 4. Pedestrians/bikers/other road users need to be considered as well, and may be outside of camera shot 5. What if the connection is choppy/equipment breaks. Does a tech go site to site? Does the crew stop working? Labor is the single biggest cost of construction by a large margin. Everyone minute the crew isn't working is CrewNumber * PrevailingWage lost. 6. Flaggers work at night/all weather/complex intersections. All this would need to be accounted for. 7. Like self driving cars, there are about 12 billion edge cases that take a combination of experience/reasoning. I don't mean to be dismissive of your idea or anything. But this attitude is pervasive in our industry, that "I saw someone doing a job that looks easy, I bet I know everything about it and can design a better way to do it". That attitude is both great (inspires bold ideas) and terrible. The right question isn't "I bet this could be automated" but rather, "I should job shadow someone doing this job, see if there are any ways to improve it".

PraetorianGourd 5y ago

Have you ever worked as a flagger? I have. It isn't a binary "move the sign and then flip it back". Here are a list of things that you would need to consider: 1. Flaggers often go from site to site throughout the day. You would need someone to set up and take down the machine 2. The flagger is in constant radio contact with the work crew. They need to know which equipment is blocking which lane etc. And this is always changing. 3. Flaggers need to respond to emergency vehicles, busses, their own work trucks etc. You have to have your head on a swivel at all times 4. Pedestrians/bikers/other road users need to be considered as well, and may be outside of camera shot 5. What if the connection is choppy/equipment breaks. Does a tech go site to site? Does the crew stop working? Labor is the single biggest cost of construction by a large margin. Everyone minute the crew isn't working is CrewNumber * PrevailingWage lost. 6. Flaggers work at night/all weather/complex intersections. All this would need to be accounted for. 7. Like self driving cars, there are about 12 billion edge cases that take a combination of experience/reasoning. I don't mean to be dismissive of your idea or anything. But this attitude is pervasive in our industry, that "I saw someone doing a job that looks easy, I bet I know everything about it and can design a better way to do it". That attitude is both great (inspires bold ideas) and terrible. The right question isn't "I bet this could be automated" but rather, "I should job shadow someone doing this job, see if there are any ways to improve it".