Post

Just watched a video of someone scared of LLMs becoming sentient/AGI. From my understanding, as long as an AI model is unable to learn in realtime, this will never happen. I see no reason to be scared before we change the "algorithm" to be able to learn on the fly. Thoughts?