Raw LLM Responses

I'd say one of the biggest flaws in human perception about AI is that while we think of AGI as the AI that can understand and learn anything a human can, we're actually setting an impossibly high standard that no human can ever achieve. For example, people who do understand newtonian physics might struggle with deciphering social cues, behave like idiots around pets, etc. Nobody is even remotely close to being good at everything. A human with the ability of this theoretical AGI would easily be considered to be a god. The combined skillset of all current narrow AI already developed is fairly close or already surpassed what the AGI level really should be about - an AI with the ability of a random average human. I'd say that's pretty much where we are right now.