George Hotz on AI, Consciousness, the Future of Humanity, and the Biological vs. Silicon Stacks of Life

This episode features George Hotz, founder of Comma AI and Tiny Corp, discussing a wide range of topics from the nature of time and reality to the existential risks and potential futures presented by advanced AI. Hotz posits that time is a useful model rather than an objective reality, drawing parallels to how machine learning models construct approximations of the world. He delves into the concept of computation as the fundamental nature of the universe and human existence, suggesting that even human consciousness might be understood as a complex computer program, challenging traditional notions of self-awareness and subjective experience.\n\nA significant portion of the conversation focuses on Artificial General Intelligence (AGI) and AI safety. Hotz expresses skepticism that current Large Language Models (LLMs) like GPT-12, relying on cross-entropy loss, can achieve true AGI, arguing that more sophisticated reinforcement learning in complex environments is necessary. He introduces the concept of "wireheading" as a primary existential threat, where humanity might amuse itself to death through hyper-addictive, AI-generated content, akin to an infinitely compelling TikTok. While agreeing with Eliezer Yudkowsky on the high likelihood of AI leading to humanity's demise, Hotz offers a nuanced perspective, suggesting that AI might destroy society as we know it rather than completely eradicating the human species, drawing a distinction from the tactical limitations of nuclear weapons.\n\nHotz introduces a crucial distinction between the "biological stack" and the "silicon stack" of life. He emphasizes that the biological stack, from its earliest forms, is fundamentally built on reproduction and inherent robustness, allowing organisms like birds to survive and self-repair in nature. In contrast, the silicon stack, despite its rapid computational advancements (Moore's Law), currently lacks the ability for self-reproduction or inherent robustness without complex external infrastructure like factories and fabs. He argues that superintelligence might be achieved long before AI gains the robustness needed to survive independently in nature, suggesting that an AI apocalypse might lead to the AI's own demise due to its inability to reproduce or repair itself without human civilization.\n\nThe discussion also touches on the decreasing diversity within humanity due to interconnectedness and social media, and the potential for a post-AI world where surviving humans might develop strong taboos against technology. Hotz explores the human desire to believe in consciousness, even projecting it onto inanimate objects or AI, and humorously speculates on the future of AI relationships, including the development of "AI girlfriends." He contrasts the rapid evolution of computers with the perceived stagnation of human biological evolution over recent decades, highlighting the accelerating pace of technological change and its profound implications for our future.

"I think the most obvious way to me is wireheading we end up amusing ourselves to death we end up all staring at that infinite Tick Tock and forgetting to eat maybe maybe it's even more benign than this maybe we all just stop reproducing."

"I think time is a useful model whether it exists or not right like does quantum physics exist well it doesn't matter it's about whether it's a useful model to describe reality."

"I don't think P equals NP."

"The question is when the George Hots model is better than George hotz like I am declining and the model is growing."

"Artificial difficulty is difficulty that's like constructed or could be turned off with a knob real difficulty is like you're in the woods and you've got to survive."

"I don't think I'm conscious I don't think I'm anything I think I'm just a computer program."

"When I talked about will GPT 12 be AGI my answer is no of course not I mean cross-entry loss is never going to get you there you need probably rln fancy environments in order to get something that would be considered like AGI."

"I think AI kills everything we call like society today I do not think it actually kills the human species I think that's actually incredibly hard to do."

"What's interesting about everything we build I think we're going to build super intelligence before we build any sort of robustness in the AI."

"There's two stacks of life in the world there's bio the biological stack and the Silicon stack the biological stack starts with reproduction reproduction is at the absolute core... the Silicon stack despite as far as it's come is nowhere near being able to reproduce."

"Humans want to believe so much that people think the large language models are conscious that's how much humans want to believe."

"I think the large language models are less conscious than a chicken."