Looking at google’s search trend for AGI, it hasn’t really picked up and it’s mostly china-base:
As debated by Lex Fridman and Yann LeCun (Meta AI) at AGI is still very far away and very gradual! The following is an extract from a 3 hour interview between Lex and Yann.
I cannot deny that AI has gone quite far in the past couple of years and we may arguably call it a SuperAI, however, the definition of an AGI begs that the AI becomes good at so many things that are still out of the scope. The most general AIs are multi-model LLMs that can handle text, audio and video, however, that leaves much for touch, smell and locomotive actions that can be performed by an embodied LLM or a robot, which brings us to Figure by OpenAI:
As appealing as this may look and without going into the Q* drama that we know very little about, we don’t know how contained was this experiment that is the only experiment that was shown to us. However, it does capture the idea of a path towards an eventual AGI…
What are your thoughts on this and when do you think will AGI be achieved?