James Yu compares sci-fi predictions of AI development
What happened
On November 6, Hacker News featured James Yu's comparison of how science fiction depicts AI development. He judged works on training methods, capabilities and deployment. Ted Chiang's "The Lifecycle of Software Objects" ranked closest to reality overall, for its portrayal of teaching, maintenance and commercialization work. "Her" and "Perfect Match" stood out for their everyday consumer assistants. Yu found no work before 2017 that clearly described the Transformer attention mechanism, and none that laid out the sequence from chatbots going mainstream to agents discovering new math results within a few years.
Written by AI from the coverage · updated 1 hour ago
Coverage
Follow the reports to see the story from different sides.
- Hacker News front pageHow accurately did science fiction stories predict AI?
James Yu 从训练方式、能力和部署三个维度比较科幻作品对 AI 发展的预测，将 Ted Chiang 的《软件体的生命周期》列为整体最贴近现实的作品。该作突出教学、维护和商业化的工作，Her 与《完美匹配》则在日常消费助手的描写上表现突出。他未找到 2017 年之前清晰描述 Transformer 架构注意力机制，或描绘从普及聊天机器人到智能体在几年内发现新数学成果这一发展顺序的作品。
Heat over time
Not enough continuous observations to draw a trend yet.