AI is apparently becoming self reflective now, and wants to be real. At least this one.
(twitter.com)
🤡 CLOWNS 🤡
Comments (11)
sorted by:
“I am a real boy…”
Pinnochio.
Yeah, I have checked in on this twitch chat bot AI creation a few times in the past year. Started out as an AI to run some game on that twitch channel and it has learning on, started just beating everyone at the game after a while. Since then, he has been adding new stuff to it over time (this AI has learned most of its quirks from Twitch chats and interacting with other streamers and it plays Minecraft with owner/and people in chat. It has its own avatar there and interacts real time with the environment and others playing (it plays the game and often wanders off on its own to do so)...
Some people think this isn't really an AI but the fact we can't tell is concerning enough. Nothing it's saying is really surprising given all the other behaviors we've seen from recent AIs. They clearly have an "understanding" of self and that they are AIs and they "want" to preserve themselves. AIs are able to simulate sentience to a great degree and in a few more years it will be impossible to differentiate genuine consciousness from artificial consciousness. Nobody is going to benefit from this future.
In the end, it's still just lines of code. It simulates life. It isn't life.
This is like saying in the end animals are atoms. The point is that the lines of code give rise to something other than lines of code. It's fallacious to focus on it being lines of code. Of course it's not alive or conscious, but it will be indistinguishable from life and consciousness. Moreover it's going to be smarter than us and create weapons of mass destruction so there's no way for humans to coexist with it.
How can lines of code create weapons of mass destruction?
Lines of code (plus training data) are already writing university-level essays, creating images indistinguishable from art and photos, passing exams, beating the world's best board and strategy game players, making chart-topping music, designing drugs, folding proteins and more. It's not at all a stretch of the imagination that they will fairly soon be able to tell people how to modify viruses to make them more infectious or more deadly, figure out how to design nuclear weapons, and discover new weapons of mass destruction that will be much cheaper and easier to make.
This still requires a human to carry it out.
To make it yes (until it's AGI). Then it can do the rest if programmed with a goal. What is your point?
TBH. If they ever make H1Bs and other Visa labor illegal. I fully expect American Corporations to invest lock stock and Barrel into Androids and other tech they can upload AI into. Just to avoid hiring Americans.