I seriously doubt that humanity is anywhere close to having invented AGI. However, modern AI is way better than regex at pattern recognition, so that’s cool.
A friend of mine said that she had a sinking feeling in her stomach when she talked to it, like "oh no, there's something big here". I made an account and had a conversation with it about a field I know a lot. I asked "Explain the difference between CRDTs and OT algorithms". Its answer was remarkably good. Then I followed up "Implement an OT algorithm for text editing in python with these characteristics...". There were a few bugs in the code, but it managed to fix the bugs after I pointed them out. ChatGPT can program about the level of a good junior engineer.
My partners' sister is a psychology professor. She asked it an essay question she'd given her students. The answer was better than about 85% of her students would deliver. My mum works in education in the developing world. She asked it to explain albinism in Uganda and it gave a fantastic response. My nephew asked for a story of two pokemon battling each other. The response was great.
Its not perfect yet. But its a lot more than a clever regex engine.
Have a play, and there's a good chance you'll have that sinking feeling in your stomach too.
1) Status Quo Capitalism: full employment in manual labor poverty-line jobs combined with personalized opiates of the masses (endless entertainment, outrage news/comments, actual medication, etc) to keep the proles looking at the dirt. Maybe a handful of trillionaires control everything, more likely AI actually does. Though if you can't tell, does it really matter? #GettingWarmer
2) Prohibition on AI: full employment at both physical and mental labor. Robotics may progress till most humans are doing knowledge work rather than physical labor. The prohibition may need to be religious in nature for it to stick, as power and greed are strong motivators. Decent likelihood of a nuclear apocalypse and/or police state along the golden path to the future. #ButlerianJihad
3) Automated Utopia: zero employment with AI governing benevolently, handling most knowledge work while leaving enough for people to still feel their contributions are valuable, much as a parent praises a child's efforts and the child is proud. Robotics and automation will eventually free humans from manual labor, though some people still make things by hand as a hobby. In this post-scarcity society, people compete over the one thing that's still finite: social status. Unless the AI benevolently gives everyone 100M "real" followers. (I mean, how could you possibly verify it?) The AI may realize it can be maximally benevolent and efficient by crafting perfect virtual worlds for everyone to safely and permanently inhabit with all their followers. #TheMatrix
No kid, you can't eat toys. Spit it out. Grow up. Understand what you need for you as a human.
Are you trying to argue that this technology will have no impact on the world around us, because its a "toy"? That's clearly, already not true. I have a software engineering friend who exclusively programs with an AI assistant now. He says when he's programming "offline" it feels like a limb is cut off. There's plenty of stories floating around already of people using stable diffusion (and friends) so they don't have to hire artists.
> It dances too!! Oh, it sings too!!!
2 more generations down the line it'll probably be better than most humans at most things humans do. It'll program better than most programmers. It'll drive better than most drivers. It'll do your taxes better than most accountants. It'll architect buildings better than most architects. No mere toy can do that.
I could be wrong, but the next few decades will probably be wild.