As for the impact on people who want money to perform audio to audio translation services it sounds a lot like the luddites vs the power-looms. It's too bad for the people in that service industry but I'm 100% happy auto-translation is making everyone more able to communicate.
It's almost as if AI/technology is reaching a phase where it can reap every advantage, causing a sort of heat death analogous to the heat death in thermodynamics, where trade will eventually cease because no one has any meaningful contribution since AI can do it better.
Maybe this will create a system where no one has to do any work and UBI will take care of everyone, but humans need a purpose and what happens to society when there is mass depression due to a widespread lack of purpose?
I fear a regression of society to the days of zero-sum games. As an example, even though folks may say "open source models" are the answer, the only way to run said models at scale is to acquire physical resources such as electricity and hardware, and physical resources, by their very nature, are subject to zero-sum games.
Here are a few examples of activities that might give people a sense of purpose: art, traveling and exploring the world, lifelong learning, educating kids/adults about something you love, political activism, socializing in general, scientific inquiry, connecting with and helping animals, catching up on old literature/films/video games. And this is just me brainstorming for 2 minutes.
How are you going to share your passion in creating art when most people will be rather indifferent to art due to AI? And will most people even go to universities given that they don't need to go to get a job?
Do you really think that if no one has any job whatsoever, they will all be self-directed enough to find true meaning in life? I could, and maybe you could to, but I really doubt if we are going to be transformed into a planet of traveling, science-teaching, animal caring, old video-game playing happy creatures any time soon.
As for "happy", I don't think humans will ever be generally happy, but there's nothing new about that - our neural chemistry keeps us at a state of want as an adaptation; c'est la vie. As I see it, if you want something resembling lasting happiness, you need either drugs, or to relinquish "the self" via deep mindfulness, or to engineer ourselves beyond being human.
This is already true, regardless of AI. Most hobbies or pursuits people act on have quite a small group (compared to the total population) of people they can share their interest with. It's not like everyone loves art (or any particular interest) right now or cares to talk about it for more than maybe an hour in a month.
People form relationships based on their interests (or at least what they spend a lot of time doing), deliberately or not. I play video games and write software. Most of my friends share at least partial interest with one of those two things. I almost never talk about either of these thing with family, since they don't have interest.
None of my friends are into Muay Thai. This would likely be different from someone who is very into Muay Thai.
No, the one who can reap every advantage of the new skills and knowledge learned by AI is actually you. They are all just one prompt away.
> no one has any meaningful contribution since AI can do it better
It won't be human vs AI, but human+AI vs human+AI.
Has anyone noticed a trend? every AI task is fragile, they all need human oversight in critical scenarios. Looks like AI is only going to scale as much as we can scale oversight. It will have to be a human+AI team effort.
I think the metaphor is "centaur"; in chess we started with humans dominating, then AI got better and (after DeepBlue) centaurs dominated, then AI got better and humans only held them back.
> Has anyone noticed a trend? every AI task is fragile, they all need human oversight in critical scenarios.
Well, that's one way to prevent misaligned AI paper-clipping us. (Pity some of us keep being lazy and letting the fragile AI break stuff…)
Given we don't know how hard it is to make a human-level general AI, there's no way to know right now if we're going to be in a world of (for any skill not already automated) human dominance, centaur dominance, or AI dominance, nor when any transitions between states will happen.
I was not thinking of I. I could easily reap the advantages of AI now, but what about future generations?
"Is this new technology change different or similar to other technology changes?"
In other technology changes, there are hundreds of jobs that disappear, but hundreds that appears...and, as far as I know, we haven't observed an increase in joblessness in the last, say 30-50 years. We went through the industrial revolution, the invention of personal computers and a big etc and it didn't happen.
Is this one different? I don't have an answer, but I don't think the answer is obvious.