Sam Harris: Trump, Pandemic, Twitter, Elon, Bret, IDW, Kanye, AI & UFOs | Lex Fridman Podcast #365

AI Risks

Sam Harris shares his unchanged concerns about artificial general intelligence (AGI), emphasizing the existential risks it poses. He highlights the challenge of aligning superintelligent AI with human values, a problem he believes remains unsolved. Harris draws an analogy to humans' disregard for insects, suggesting that superintelligent AI might similarly overlook human interests without malicious intent 1.

Unless we find some way of permanently tethering a super intelligent, self improving AI to our value system, I don't believe anyone has figured out how to do that or whether that's even possible in principle.

--- Sam Harris

The discussion underscores the urgency of addressing these concerns before AGI becomes a reality 2.

   

AI Progress

The conversation shifts to the broader implications of AI advancements, with Harris expressing skepticism about the current trajectory towards AGI. He questions whether deep learning techniques will suffice to achieve true general intelligence, noting their limitations in generalization and abstraction 3.

I think we have reason to be skeptical that these techniques of deep learning are actually going to be sufficient to push us into AGI.

--- Sam Harris

Despite these doubts, Harris acknowledges the impressive progress of AI models like Chat GPT, which leverage the collective intelligence of humanity. The potential for AI to surpass human capabilities remains a significant concern, prompting discussions on how society should prepare for such advancements 4.