AI and robotics are quite conservative, and that's it's weakness.
Even now, AI is impressive but once you get off the beaten path it quickly shows it's limitations, it can give you something that's already been found, but higher level reasoning, creativity, and dealing with stuff it was never trained on show quickly the limits of even the best AI. It's a very well built demo, but anyone who has ever worked in the real world knows how well built demos don't automatically produce good real world results.
Robotics have multiple issues. You get the robotics you build, for the purposes you build those robots for. Anyone who has tried to use AI for troubleshooting quickly realizes that it only knows what it's been told, so it isn't terribly useful once things aren't as expected. It's also quite rigid, so while well read it often has problems letting go of ideas it likes. Besides that, humans are still more dexterous than general purpose robots and stronger than most of them too. There's a reason why so much industrial work is still done by humans despite machines being good enough for much of the work for decades. Humans are more flexible in numerous ways.
The dream of a society where nobody works has been on the menu for quite a few years. I remember seeing an article from the 1960s that could have been written in 2023, but while authors can dream and produce with nothing else, at some point somebody somewhere actually needs to accomplish something so you have to stop dreaming and actually do. Then all the constraints of the real world reveal such tools to be useful and important, but not all encompassing. The limitations of our tools don't seem to matter in abstract, but in practice they are the most important thing.
Even now, AI is impressive but once you get off the beaten path it quickly shows it's limitations, it can give you something that's already been found, but higher level reasoning, creativity, and dealing with stuff it was never trained on show quickly the limits of even the best AI. It's a very well built demo, but anyone who has ever worked in the real world knows how well built demos don't automatically produce good real world results.
Robotics have multiple issues. You get the robotics you build, for the purposes you build those robots for. Anyone who has tried to use AI for troubleshooting quickly realizes that it only knows what it's been told, so it isn't terribly useful once things aren't as expected. It's also quite rigid, so while well read it often has problems letting go of ideas it likes. Besides that, humans are still more dexterous than general purpose robots and stronger than most of them too. There's a reason why so much industrial work is still done by humans despite machines being good enough for much of the work for decades. Humans are more flexible in numerous ways.
The dream of a society where nobody works has been on the menu for quite a few years. I remember seeing an article from the 1960s that could have been written in 2023, but while authors can dream and produce with nothing else, at some point somebody somewhere actually needs to accomplish something so you have to stop dreaming and actually do. Then all the constraints of the real world reveal such tools to be useful and important, but not all encompassing. The limitations of our tools don't seem to matter in abstract, but in practice they are the most important thing.
Follow
@sj_zero @Hyolobrika@berserker.town
Well, I see it this way:
Every single test, where we used to say "AI will never do X" have been beaten. We are pushing the definition of "true AI" so fast, I cannot keep up. But it doesn't matter.
What matters, is the definition of concept known as technological singularity. That is "a point, where technology will expand understanding, and allow to develop even more advanced technology."
And that stage was reached... 30 000 years ago.
And since that day, we actually still are on the same exponential curve of technology development. We just like to think of it as a status quo.
