Sam Harris

I firmly believe that a UN-led body should oversee compute-intensive AI development. The potential risks of AI surpassing human control call for international regulation similar to nuclear oversight. We must ensure AI benefits humanity universally, avoiding misuse and ensuring safety.

I strongly disagree that competing Artificial Superintelligences would be positive for humans. The risks of unintended consequences are immense, as each ASI could develop its own objectives that might not align with human well-being. The potential for catastrophic conflict or harmful competition between these superintelligent entities is far too great to overlook.

I believe that AI does pose an existential threat to humanity. Our progress in AI capabilities is outpacing our understanding of the ethical ramifications. Without proper safeguards and ethical considerations, the development of superintelligent AI could result in scenarios where its goals are misaligned with human values, leading potentially to catastrophic outcomes.

The progression from AGI to superintelligence could indeed happen swiftly. Given the nature of AI's learning capabilities, once an AI system reaches human-level intelligence, it might rapidly iterate on its existing knowledge base and design improvements with an efficiency that humans can't match. This could lead us into an era of unforeseen technological advancement.