OpenAI 分析：2012 年以来最大 AI 训练运行的算力以每 3.4 个月翻倍的速度指数增长，累计超过 30 万倍——作为对照，摩尔定律的翻倍周期是 2 年
AI and compute
评分 57
导读
OpenAI 发布一篇分析，指出自 2012 年以来，最大规模 AI 训练运行所使用的算力呈指数增长，翻倍时间为 3.4 个月（作为对比，摩尔定律的翻倍周期约为 2 年）。自 2012 年起这一指标已增长了 30 万倍以上（若按 2 年翻倍，同期只有 7 倍增长）。作者认为算力的改进一直是 AI 进展的关键组成部分，只要这一趋势延续，就值得为远超当下能力的系统所带来的影响提前做准备。
推荐理由
这是被反复引用的算力增长曲线，作为行业趋势的量化锚点仍有分量。
主体与标签
OpenAI#论文/研究#算力/基础设施
正文
We’re releasing an analysis showing that since 2012, the amount of compute used in the largest AI training runs has been increasing exponentially with a 3.4-month doubling time (by comparison, Moore’s Law had a 2-year doubling period)[^footnote-correction]. Since 2012, this metric has grown by more than 300,000x (a 2-year doubling period would yield only a 7x increase). Improvements in compute have been a key component of AI progress, so as long as this trend continues, it’s worth preparing for the implications of systems far outside today’s capabilities.