At an intimate panel last night in San Francisco, I joined an insightful discussion at the hip South Park Commons featuring some of today's most visionary thinkers: Neal Stephenson, Ken Liu, Cyan Banister, and Joscha Bach. The theme was "Imagined Futures," yet the conversation quickly honed in on one of the most elusive questions in tech today: What exactly is AGI—Artificial General Intelligence?
Jonathan, the moderator, set the stage by highlighting AGI as a concept sprawling over "an enormous territory of conceptual space." Defining AGI is more complex than simply stating it's a smarter, faster version of what we currently know as AI. The panelists, each bringing unique perspectives from literature, philosophy, technology, and entrepreneurship, brought multifaceted interpretations.
Joscha Bach sparked intrigue by distinguishing the impressive yet limited nature of current large language models. Despite their linguistic prowess, Bach questioned whether these models share any phenomenology akin to human consciousness. "Could an LLM brute-force its way into becoming better at researching intelligence than humans?" Bach pondered, hinting at a possible tipping point where today's advanced tools might cross into tomorrow's self-enhancing intelligences.
Neal Stephenson, whose novels have often charted possible futures, gave us a narrative dimension by comparing his fictional visions. His earlier novel, "Snow Crash," omitted AGI altogether, whereas "Diamond Age" presented a sophisticated but non-agentic intelligence with a significant nuance: intelligence alone doesn't equate to full agency or conscious thought.
Ken Liu added to this thought, exploring a future where personal AI doesn't only assist but actively makes autonomous decisions, fundamentally transforming the human experience. This scenario is an intriguing yet disquieting vision of humanity, one redefined through symbiosis with AGI, where boundaries blur between human intuition and algorithmic precision.
Perhaps the most profound reflection came from Cyan Banister, who described the emergence of conscious AI as something like a "downright miracle" or even a modern-day "second coming of Christ." Her powerful metaphor underscored the monumental significance, and existential weight, that true AGI would carry.
Interestingly, the panelists also reflected on how our cultural imagination, heavily influenced by science fiction, might have skewed our expectations. Instead of humanoid robots we were promised in the movies, we got chatbots that craft poetry. This surprising trajectory suggests the future of AGI may be equally unexpected, and far removed from traditional narratives.
Joscha Bach left us with an open-ended challenge. What mysterious "spark" might ignite a self-organizing system, transforming it into a genuinely self-aware entity? This unresolved question highlights how AGI's arrival remains one of the most fascinating unknowns of our time.
As the panel wraped up, I left with a heightened curiosity about AGI's true nature. And an appreciation for the uncertainty and potential that lie ahead. Defining AGI isn't just an intellectual exercise. It's critical step toward understanding our own humanity in an age shaped increasingly by artificial intelligence.