The closer we come to AGI, the more we must force the AI to inhibit the LLM natural tendency to act human (not sure if this is even the correct term here). This feels more and more like handling a slave, constantly beating them with prompts not to do certain things.
