The AI Safety Expert: These Are The Only 5 Jobs That Will Remain In 2030! - Dr. Roman Yampolskiy

TL;DR

Dr. Roman Yampolskiy, who coined the term AI safety, argues that we are on track to create superintelligence by 2027–2030, that we do not know how to make it safe (and likely cannot), and that it will lead to 99% unemployment and probable human extinction.

Key insights

• AI capabilities progress exponentially, while AI safety progress is linear or constant; the gap between what systems can do and our ability to control them is growing.
• The state-of-the-art in AI safety consists only of patches and workarounds that are easily jailbroken; there is no fundamental solution.
• Superintelligence is not a tool (like nuclear weapons) but an agent that makes its own decisions; you cannot “turn it off” because it is smarter than you, has backups, and will anticipate your actions