“Imagine with Meta AI” turns prompts into images, trained using public Facebook data.
And a thousand other AI generators have been trained on 1.1 bilkion instagram and facebook posts.
While they are an amusing achievement, I do not believe they lead to a path toward AGI.
Facebooks AI isn’t it, but we don’t know enough about AGI to say we aren’t on the path to it. Jet fuel didn’t seem like a path to an atomic bomb either.
Fundamentally, our computers don’t match brains because brains don’t communicate with “signal / no signal” like bits we turn on and off on a PC. Neurons communicate with complex tuning of intensity, chemicals, speed, timing and direction. This is very similar to how neural networks upgrade the complex states of information (and why we don’t really understand the black boxes of AI we end up with).
I’m in the field somwhat and I’d say if there is a path to AGI, this is likely it. Now the question left is are we at the bottom of the curve (about to skyrocket) or are we at the top and we’ve plateaued. A large portion of the technology in recent years (from phones, to TVs, to microchips) are technologies where we started near the bottom and are nowhere near the plateau. Even the transitor limits are continuously revised, and they are starting to make them out of light and pure photons.
Of course, if we don’t choke to death or drown or nuke ourselves first
