Yes, perhaps an "appropriate algorithm" could, but it is my opinion that we have not found that algorithm. LLMs are cool but I think they are very primitive compared to human intelligence and we aren't even close to getting AGI via that route.