youtor.org
Provably Safe Systems The Only Path to Controllable AGI
>> YOUR LINK HERE: ___ http://youtube.com/watch?v=nUrYCUkTFE4
Steve Omohundro discusses a paper that Max Tegmark and he wrote (available at https://arxiv.org/abs/2309.01933) about Provably Safe Systems enabling humanity to thrive with Artificial General Intelligences (AGIs) with a group of AI researchers. He argues nine key points: 1. AGI is imminent 2. AGI is an existential risk 3. AI alignment is insufficient 4. We need a security mindset 5. We need Provably Safe Systems 6. AI theorem proving is critical and imminent 7. We need cryptographically secure hardware 8. This is the only path to safety 9. Success will lead to human thriving
#############################
