AI may not be the dire existential threat that many make it out to be. According to a new study, Large Language Models (LLMs) can only follow instructions, can't develop new skills on their own and are inherently "controllable, predictable and safe,"...
