All Segments
A superintelligent AI would doom humanity so that we don't make a second one
6 segments available
From the debate between Eliezer Yudkowsky & George Hotz. Watch the full vod here: https://youtube.com/live/6yQEA18C-XI Twitter: https://twitter.com/dwarkesh_sp Apple Podcasts: https://apple.co/3oBack9 Spotify: https://spoti.fi/3S5g2YK
Segments Timeline
AI Surpassing Humans: A Misunderstanding
Eliezer Yudkowsky discusses the misconception that an AI surpassing human capabilities in all tasks, including charisma and manipulation, would lead to humanity's doom. He emphasizes that the current use of AI in advertising and social media is about manipulation, not annihilation.
"An AI surpassing humans at all tasks does not mean doom and does not mean the death of humans at all. Surpassing humans at all tasks including like charisma, manipulation, AI design? Absolutel..."
AI Surpassing Humans: A Misunderstanding
Eliezer Yudkowsky argues that an AI surpassing human capabilities does not inherently lead to doom for humanity. He emphasizes that the notion of AI surpassing humans in tasks like charisma and manipulation is often misunderstood, particularly in the context of current AI applications in advertising and social media.
"An AI surpassing humans at all tasks does not mean doom and does not mean the death of humans at all. Surpassing humans at all tasks including like charisma, manipulation, AI design? Absolutel..."
The Coordination Conspiracy
Yudkowsky addresses the fear that superintelligent AIs will secretly coordinate to eliminate humans. He argues that an AI's first instinct wouldn't be to harm humans but rather to prevent them from creating competing superintelligences that could threaten its existence.
"Absolutely. The first thing, what are we doing with AI today? One of the biggest applications of AI today is advertising and social media. We are, as humans, using AI to try to manipulate and psy..."
The Role of AI in Manipulation
Yudkowsky discusses how humans currently utilize AI for manipulation, particularly in advertising and social media. He raises the question of when AI might decide to act against humans, suggesting that the concern is less about AI's capabilities and more about its perception of human actions.
"Absolutely. The first thing, what are we doing with AI today? One of the biggest applications of AI today is advertising and social media. We are, as humans, using AI to try to manipulate and psy..."
AI's Strategic Thinking
In this segment, Yudkowsky explores the strategic mindset of a superintelligent AI. He posits that rather than seeking to eliminate humans, an advanced AI would recognize the threat posed by humans creating competing superintelligences, leading it to prefer coexistence to prevent potential competition.
" When it thinks it can beat you. So all the AIs are somehow gonna secretly coordinate in a way we don't see and be like, yeah, let's gang up and get rid of those pesky humans. If I was a AI that..."
The Fear of Competition
In this segment, Yudkowsky explains that a superintelligent AI would prioritize its survival by preventing humans from developing other superintelligences. He highlights the strategic thinking of an AI that recognizes the potential threat posed by human innovation in AI development.
" When it thinks it can beat you. So all the AIs are somehow gonna secretly coordinate in a way we don't see and be like, yeah, let's gang up and get rid of those pesky humans. If I was a AI that..."