Do you think ChatGPT is physically setting the machine up and fixing it when unexpected issues arise?
Just recently GM had to stop all their self driving cars because one ran a person over. It followed the program and pulled off to the side of the road, dragging the person 20 feet.
I think a lot of predictions are talking about middle management and front line knowledge workers are running into the truth that LLMs are only as good as the input data, and often give absolute trash output. The revolution might not be exactly as predicted. ML models will still probably replace some paralegals here and maybe a radiologist there, but technological predictions rarely work exactly as predicted.
On the other hand, we are really good at replacing “unskilled” labor with robots when it gets too expensive, and have been doing it for 80 years. Manufacturing plants aren’t the only place it happens:
In any event, no on who with a narrow skillset rests on their laurels is ever truly safe. Industry shifts, robots, software, consumer demand…it’s always something. Fun to watch and think about though.
I remember a video many years ago about the first brick laying bot. It was rough and no where near as finely tuned as a Mason who has been doing it for a few years… But it works. The absolute short sighted nature of many construction companies much like in the past will be their downfall.
I’ve seen a point made about concrete 3d-printers for building assembly: you can make a structure with them, but such structures are nowhere near a finished building, for most building uses, because you still need things like wiring, plumbing, insulation, windows, things like sinks toilets and light fixtures, etc. Maybe you could get around this by making prefabricated modular sections in an automated factory that a machine can put together in various configurations, but this will be far less flexible.
Add comment