Would training LLMs on either smaller batches of text first instead of the (current) large batches of text or on easier subjects/topics first then later train them on subjects/topics with increasing order of difficulty...enable the LLMs to learn better. I ask this because the hum
In Marvel's VisionQuest, a superpowered robot tries to better understand humanity by getting lost in elaborate, digital daydreams in which he and other artificial intelligence-powered beings live together in a home. The series is deeply fantastical, yet it plays with some of the