AI 为何暂时取代不了数学家 —— 陶哲轩

I'm very curious when you expect AIs that can like actually do frontier math better than the at least as good well as the best human mathematicians. 我特别好奇，你预计 AI 什么时候能真的做前沿数学，至少跟最顶尖的人类数学家做得一样好。 I mean in in some ways they're already doing frontier math 我是说，某种意义上他们已经在做前沿数学了， that is super intelligent that humans can't do but it's a different frontier from what we're used to. 那种超智能的、人类做不到的事——只不过这跟我们熟悉的前沿不太一样。 Um I mean you could argue that calculators were doing frontier math 嗯，我是说，你也可以说计算器早就在做前沿数学了—— uh that that humans um could not accomplish but it was but it wasn't you know number crunching it 呃，那些人类完不成的事，但那其实就是所谓的数值运算而已。 um but um 嗯，但是—— but but replacing Terry Tao completely 但——我说的是完全替代掉 Terry Tao。 I mean what do you want me for? 我是说，那你还要我干嘛？ Uh I um 呃，我，嗯—— you'll just go on all the podcasts after. 那你之后去各种播客串台就行了。 [laughter] [笑声] I'm not sure we've 我不确定我们—— it might not be the right question to ask. 这可能并不是合适的问题。 Um 嗯—— I think with within a decade a lot of things that mathematicians currently do um what we spend a lot of the bulk of our time doing and a lot of stuff we put in our papers today can be done by AI. 我觉得十年之内，数学家现在做的很多事——那些占据我们大部分时间、会写进今天论文里的内容——AI 都能做。 Um but we will find that that actually wasn't the most important part of what we do. 嗯，但我们会发现那些其实并不是我们工作中最重要的部分。 Um you know um 100 years ago um a lot of mathematicians were just solving differential equations. 嗯，你看，100 年前很多数学家就是在解微分方程。 Um like people needed physicists needed some exact solution to to to to some system and and they were just they hired a mathematician labor to go through the calculus and work out the solution to this fluid equation 比如说物理学家需要某个系统的精确解，他们就雇一个数学劳动力去做微积分运算，把这个流体方程的解算出来—— whatever. 之类的。 um a lot of what um uh a 19th century mathematician would do um you could make a call to uh Mathematica or or Wolf Alpha or or a computer algebra package or more now more recently NAI and it would just solve the problem you know in a few minutes um but we we moved on we we we worked on different types of problems after that um you know once computers came along you know computers used to be human right people used to laboriously create log tables and and and and and work out primes as scouts did and that has all been outsourced to computers. 嗯，一个 19 世纪数学家会做的很多事，现在你调一下 Mathematica、Wolf Alpha、某个计算机代数软件，或者最近的 NAI，几分钟就能把问题解出来。不过我们往前走了，之后开始处理不同类型的问题。你要知道，计算机出现之前，「computer」这个词指的是人——人们辛辛苦苦地编对数表、像当年的先驱那样手算素数，这些工作现在全都外包给计算机了。 Um but but we we moved on um in genetics you know um to to to sequence the the genome of a single organism that was an entire PhD of a geneticist you know still carefully you know separating all the chromosomes and and and whatever. 嗯，但我们继续往前走了。比如遗传学里，过去给一个生物体测基因组，是一个遗传学家整整一个博士的工作——小心翼翼地分离所有染色体什么的。 Um and now you can just spend $1,000 and send it to a sequencer and and and get it done. 嗯，现在你只要花 1000 美元寄给测序仪器就能搞定。 But genetics is not dead as a subject. 但遗传学作为一门学科并没有死。 You you move to a different scale. 它只是换了一个尺度。 You know maybe you study whole ecosystems rather than individuals. 你可能开始研究整个生态系统，而不是单个个体。 I I take your point but on the question of well when is most mathematical progress or almost all mathematical progress happening by AI so if you find out oh this year millennium price problem has been solved you would put you know a 95% odds that an AI did it autonomously 我明白你的意思。但我想问的是：绝大多数数学进展——甚至几乎所有数学进展——什么时候会由 AI 做出？比如说，某一年你听说千禧年大奖难题被解决了，你会不会给出 95% 的概率判断是 AI 自主做出来的？ surely there will be such a year 这样的一年总会到来吧？ um prIce problem has been solved you would put you know a 95% odds that an AI did it autonomously surely there will be such a year um I guess I mean I I I I do believe that that hybrid um human plus AIs will will dominate mathematics for a lot longer it it's it will depend 嗯，千禧年大奖难题被解决了，你会给出 95% 的概率判断是 AI 自主做到的——这样的一年总会到来——嗯，我想说，我确实相信「人类 + AI」的混合模式还会主导数学界很长一段时间，具体多久取决于—— it will require some additional breakthroughs be beyond what we already have. 它需要在我们现有基础上再有一些突破。 Um so it's it's going to be sarcastic. 嗯，所以这事会挺讽刺的。 Um you know I think you know AI currently are very good at certain things but but really terrible at others. 嗯，你知道，我觉得 AI 目前在某些事情上很强，但在另一些事情上真的很糟糕。 Um and and while you can sort of add more and more frameworks on top to kind of reduce the error rates and and and make them uh uh work with each other a bit more and so forth. 嗯，虽然你可以往上不断叠加各种框架，来降低错误率、让它们彼此之间更好地协作之类的。 Um I I um it feels like we are we don't have all the uh the ingredients to like really have a truly satisfactory sort of uh replacement for all intellectual tasks. 嗯，我——感觉我们还没凑齐所有的要素，做不到真正令人满意地、全面替代所有智力任务。 Um it's it is complimentary currently. 嗯，现阶段它还是互补性的。 Um it's not not uh um it it is not a replacement. 嗯，它还不是真正的替代。 Um 嗯—— but maybe I mean 不过也许——我是说—— because current level AIS will accelerate science in so many ways. 因为现在这个水平的 AI 会在很多方面加速科学进展。 Hopefully, you know, I mean, new discoveries, new breakthroughs will happen um more uh more quickly. 嗯，希望——我是说，新的发现、新的突破能——呃，更快地出现吧。 I mean um it's possible that also by somehow destroying serendipity, we we actually inhibit certain types of progress. 我是说，嗯，也有可能——因为某种程度上破坏了偶然性，我们反而抑制了某些类型的进展。 Um anything is possible really at this point. 嗯，现在真的是什么都有可能。 I think this u the world is very very unpredictable at this point. 我觉得现在这个世界非常非常难以预测。 If you enjoyed this clip, you can watch the full episode here and subscribe for more clips. 如果你喜欢这个片段，可以在这里看完整一期，并订阅频道看更多内容。 Thanks. 谢谢收看。