In the mid 2000 I was confidently predicting that home telerobots would soon be arriving. Mobile phones were getting capable enough, and had all the sensors which would be needed in a convenient and battery efficient package - video, audio and gyro. The price of all of this was also getting within the reach of being a consumer product. My futurology seemed bulletproof.
I also made predictions about the evolution of telerobots. Initially it would just be teleconferencing type things and then there would be manipulators added and remote manual work becomes possible, and so on. Soon enough all kinds of manual tasks can be done from your desktop without any physical hazards, etc. No big advances in AI would be required either. There would be a competitive economy in labor hours of telerobot work too, and the whole thing could be an internet platform.
In 2019 I can see it was all bullshit. I actually realized that by 2013, and then changed what I was doing accordingly. There's still a chance that telerobots might be ubiquitous somewhere in an undefined future, but it doesn't look close or realistic. Today if I type "telerobot" into ebay the number of results I get is a grand total of...zero. "home robot" only returns the usual Roomba-style cleaners, which already existed in the mid 2000s.
It's a sobering reminder of just how wrong it's possible to be about the future. It's possible to fool yourself into believing all manner of follies. There is no inevitable path. Just because something is technologically possible doesn't mean it will happen.