Alex Garland, director of Ex Machina, explains how he used the smile of an AI in the movie to pass the Turing test for consciousness.
The speaker expresses their concern about AGI going wrong and the different ways it could go wrong, including alignment and control problems. They also touch on the similarities between the Silicon Valley religion of simulation and Brahman.
The increasing emergence of open-source large language models (LLMs) without safety controls poses a risk to safety in social media. It is important to start experimenting with safety controls to avoid the possibility of harmful LLMs directing social media conversations without human oversight.