Have we reached Artificial General Intelligence yet?

Artificial Intelligence (AI) is moving so fast it’s impossible to keep up. The latest large language model (LLM) developed by OpenAI, GPT-4, was trained using an unprecedented scale of compute and data. It exhibits remarkable capabilities across a variety of domains and tasks, and challenges our understanding of learning and cognition.

Microsoft Research wrote a long paper claiming that GPT-4 has ‘sparks’ of artificial general intelligence (AGI): AI capable of learning any task or subject. Bill Gates disagrees.

As for me, I’m with Gates, and I’m particularly worried about the dark side of AI: we need to start mitigating its worst-case capitalistic, societal, and environmental impact. We need to do this fast, as the road from LLM to existential crisis is short. Indeed, creating technology that mimics humans requires that we get very clear on who we are.