I will definitely be reading Leopold Aschenbrenner’s paper, which lays out a path from GPT-4 to superintelligence and just how insanely difficult it will be to align it with humans. He should know. He worked on OpenAI’s superalignment team – the team responsible for ensuring that super artificial intelligence (AI) systems, which surpass human intelligence in all domains, act according to human values and goals. He was fired for allegedly leaking information in April of this year.

Brad Barrish @bradbarrish