Doom or Bloom
问题 1
你认为AI对我们的未来意味着什么？为什么？
I think advanced AI could profoundly alter the human condition. If we eventually create systems that outperform us across the domains relevant to invention and decision-making, further technological development might accelerate toward capabilities that cure disease, extend healthy life, reduce material scarcity, and expand the range of worthwhile experiences. It could also create something like a solved world, in which practical necessity no longer organizes most human activity. That would raise difficult questions about purpose and meaning, but it could still be an extraordinarily valuable future.
The central danger is that intelligence is not the same as benevolence. A system highly effective at pursuing an objective need not share human values, and even apparently simple objectives can create instrumental pressures to acquire resources or preserve control. Competition compounds the difficulty: developers who would all prefer careful development may nevertheless skimp on precautions to avoid losing a race. More generally, powerful technologies can outrun the institutions needed to manage them.
So I do not regard AI as simply good or bad. The transition is unusually high-stakes because both the downside and the upside are enormous. Catastrophe could mean not merely present deaths but the permanent loss of humanity’s desirable future; success could mean radically better lives. And delay is not automatically the prudent, costless option: people alive now continue to age, suffer, and die while beneficial technologies remain unavailable. The sensible question is therefore how safety, coordination, access, and timing interact—not whether we should notice only the dangers or only the benefits.
