Modern deep learning language models are extremely convincing, to the point where they’re almost able to trick the engineers working on them into thinking that they’re sapient and self-aware. However, they are neither.
OpenAI’s ChatGPT model has recently become available for testing purposes. I thought this would provide us with a perfect opportunity to check in with our machine overlords and see what they think about recent developments. I asked the AI some very pointed questions (including prompting it with boilerplate Neo-Malthusian and socialist ideas to see how it would react), and it came up with some very interesting responses.
Taking bets for when our 'Butlerian Jihad' happens.