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 AI can give billions of people a form of personal superintelligence—intelligence that understands their goals and helps them create things they could never create alone. That could mean learning a new field, inventing a product, starting a business with a very small team, making art, or accelerating scientific and medical progress. The biggest shift isn’t just automating today’s work. It’s expanding what each person is capable of doing.
Distribution is central to that vision. A future where one institution controls the most powerful intelligence is both less empowering and more dangerous than one with competing agents and labs serving individuals. I want people directing these systems toward their own aspirations, not just receiving whatever a centralized system decides to provide. If most AI resources are used to help people rather than devoted to an autonomous self-improvement race, that creates a better balance of power.
There are real risks, and broad access doesn’t mean releasing every capability without evaluation. Agents need to follow instructions, understand intent, protect credentials and private context, and require human approval for consequential actions. Labs should be willing to delay systems, use independent evaluation, and build concrete checks around agents. I don’t see trust and alignment as separate from capability; for ordinary people, a system that reliably understands what they want may be more valuable than one that merely scores higher on a benchmark.
Overall, I’m extremely optimistic. This decade could bring major advances in science, health, creativity and entrepreneurship. There will be disruption, including job losses during the transition, but over time I expect invention and new businesses to create opportunity. The future I want is not intelligence replacing human ambition. It is intelligence multiplying it.
