← On The Frontiers of Intelligence

The difference between AI and AGI is that AI is bound to rules while AGI is unbound, i.e., AGI can think and decide for itself. This is a huge difference. An AI cannot be general enough to be called AGI without allowing it to do unsupervised things. So we will see in the next 5 years (perhaps even sooner) the proliferation of many more AIs but not an actual AGI as the regulations and soft controls around it are still being cooked. AGI may be achieved in some workshops, but it is unlikely that the world will get to experience it.

That being said, what is even more interesting is the question of what intelligence is and whether it is the inevitable product of life. I think the answer is yes. Any living thing has the potential to be generally intelligent given the right circumstances and long enough time to develop. This brings an interesting question of whether an AGI is alive.

And that is why some believe AGI is dangerous. What is alive and cannot be hurt (by traditional means) is terrifying.