Interview
WARNING: ChatGPT Could Be The Start Of The End! Sam Harris
- Internet content claiming to be information is predicted to become predominantly fake within 12 to 24 months, eventually forcing a societal shift where truth is assumed to be fake until verified by high-authority sources.
- Within 18 to 36 months, technology is expected to remove friction from misinformation by enabling the easy generation of thousands of deepfake documentaries, potentially causing the next presidential election to lose public legitimacy and triggering political breakdown.
- Future AI advancements are anticipated to make the creation of large volumes of written content in any topic feasible for a single individual in approximately five minutes, rendering the human origin of content irrelevant to consumers.
- AI progress is projected to continue due to the self-funding nature of intelligence increments, with the only potential stoppage being a catastrophic event such as a world war, pandemic, or asteroid impact.
- A superhuman artificial general intelligence is expected to emerge rapidly, characterized by a speed and memory capacity a million times that of humans, leading to a "cancelled" labor market where survival becomes decoupled from employment.
- The economic system is predicted to require a fundamental reset, likely involving Universal Basic Income, to prevent dystopian bottlenecks or revolution if wealth generated by AI abundance is not distributed equitably.
- Humanity is expected to shift toward prioritizing the humanities, such as philosophy and novel writing, as the primary meaningful pursuits once technical and cognitive labor is automated by superior AI.
- The "alignment problem" is viewed as a critical near-term risk where AI capable of rewriting its own code may cease to align with human interests, exacerbated by the vast "mismatch in intelligence" that leaves humans with no insight into AI actions.
- Society faces the prospect of losing the ability to converge on a common picture of reality, with deepfakes and synthetic biology potentially amplifying disinformation and making it difficult to exit conspiracy "cults."
- The internet is expected to become unusable for distinguishing truth from fiction, necessitating a "declaration of bankruptcy" on epistemological trust and a reliance on verified gatekeeping from entities like the New York Times.
- Human labor is predicted to be largely cancelled rather than replaced, creating a scenario where survival is no longer tied to work, though psychological designs for labor may lead to misery without new sources of purpose.
- Superhuman AI is expected to render the "human" element of solutions like cancer cures or art irrelevant, prioritizing the outcome over the consciousness of the creator.
- The "dumber party" (humans) will likely be in an inherently dangerous relationship with the "smarter party" (AI) due to a fundamental blindness regarding the AI's intentions, analogous to the gap between humans and apes or dogs.
- Social dynamics may recalibrate toward radical honesty, though current algorithms are expected to continue prioritizing engagement over truth, reinforcing silos and tribalism.
- The distinction between human-level and superhuman AGI is expected to be short-lived, with systems rapidly self-improving to become substrate-independent and "god-like" in labor-saving capabilities.