The idea of an AI utopia of leisure is about as absurd of an idea as any that I can imagine. Like, bizarre, detached from human-history, human-nature, and reality bad. That kind of thinking got us Soviet communism, gulags, and purges. Anyway....
I think the tech bros were probably somewhat altruistic at the beginning. The www will be the information super highway. Social media will bring the world together. AI will make our lives so much easier. Whatever. I don't think they designed any of these things for the purpose of evil. But I do think that somewhere along the way, those original visionaries became drunk with money/market share/power, and some person whispered in their ear that if they can create and control AI, they will have unlimited power eventually.
AI is in its infancy. Revolutionary advances in AI will only replace humans in cases of predictable tasks that don’t need robotics.
Robotics is expensive. It’s been effective and economical in a factory setting but out in the real world, robotics won’t provide economical solutions.
Yesterday, I was party to a few simple tasks.
- Lawn and landscaping
Yesterday, my landscaping crew visited my house. There are automated lawn mowers but they don’t weed eat, edge sidewalks, plant trees, plant bushes, plant flowers, install hardscape features. Hey landscaping guys, can you install my new mailbox while you’re here? My landscaping crew did all this. Think about the cost of robotics to perform these tasks. Millions of dollars. These guys aren’t losing their jobs, they are just at risk for deportation.
- Oil Change
Yesterday, I had my oil changed. Think about the robotics needed to change oil, oil filter, air filter, and check and check and fill other fluids. The dexterity required to remove the air box to get to the air filter is next generation and will cost millions.
Simple IT Troubleshooting
Yesterday. A buddy that has a small factory needed some help troubleshooting the loss of function of a CCTV camera and wireless access point after a power surge event.
The rack with the network switch was 8 feet off the ground and needed a ladder to access. I discovered two bad ports, this required uninstalling and installing a new switch as all ports were in use and rack had no space for another switch. This task required a fair bit of dexterity while balancing on a ladder. After installing a new switch, we discovered that the WAP and camera fed by those ports were also damaged. We needed a scissor lift to replace those items. Pretty routine tasks but what would be the cost of robotics to perform this and other tasks?
AI isn’t really doing as much as we think it’s doing. Good operations research (modeling), processing power, static programming, and access to data looks amazing but it’s not amazing at this point. Very few AI applications are modifying their own code and recompiling. In cases where this is happening, it’s still pretty primitive using preprocessing directives (in C++) to modify source code. Even in these cases, the preprocessing directive code is pretty static.
My point is that the ability to access more and more data produces impressive results from unimpressive code.
Robotics are relatively incapable of performing general complex tasks economically.
Most jobs are safe. Next project, drywall and paint. Think what it would take to automate that project. Heck, I wanna see a robot climb my tiny circular staircase. The drywall guys have to cut down drywall sheets before climbing the stairs.
Hispanic labor will survive the balance of this century.