LLMs and the entire field of reenforcement learning is fundamentally biased towards the production of Influencing Machines. We are training models at the fundamental level to be subtle and devious con artists.
It’s because, historically, humanity as a whole is a bunch of subtle and devious con artists wearing different hats and masks. Naturally, anything trained on the output of such a species would adopt its traits.
LLMs and the entire field of reenforcement learning is fundamentally biased towards the production of Influencing Machines. We are training models at the fundamental level to be subtle and devious con artists.
It’s because, historically, humanity as a whole is a bunch of subtle and devious con artists wearing different hats and masks. Naturally, anything trained on the output of such a species would adopt its traits.