@liron - From today's @loganbartshow, worth a watch
Liron Shapira✓@liron
Dario Amodei's P(doom) is 10–25%.
CEO and Co-Founder of @AnthropicAI.
video not captured
Liron Shapira✓@liron
“I often try to focus on the 75–90% chance where things will go right.”
image not captured
Read 3 more tweets
Liron Shapira✓@liron
@loganbartshow It’s not just Anthropic.
OpenAI’s leadership is similarly candid about the imminent existential risk of what THEY are creating.
Liron Shapira@liron:OpenAI Alignment Team Lead @JanLeike's P(doom) is 10-90%.
Liron Shapira✓@liron
@loganbartshow These AI leaders have every incentive to say “the risk of doom is less than 1%”.
Yet they’re saying the chance that their companies’ efforts amount to mass suicide is high.
I believe they’re honest people, and that these statements are their real beliefs.
Let that sink in.
Liron Shapira✓@liron
@loganbartshow Reminder that @OpenAI's former safety lead has a P(doom) of 50%.
https://x.com/liron/status/1650980681953198080
Liron Shapira@liron:Meet the guy who invented @OpenAI's RLHF technique (https://arxiv.org/abs/1706.03741). His name is Dr. @PaulFChristiano, PhD Comp Sci from UC Berkeley. Paul says there's a 50% chance of doom once AI systems are human-level intelligent. So do we keep training smarter AI systems or no?🤔
Liron Shapira✓@liron
2023-10-07From today's @loganbartshow, worth a watch: https://www.youtube.com/watch?v=gAaCqj6j5sQ