陶哲轩谈AI如何改变数学
TL;DR · AI 摘要
陶哲轩指出AI正显著降低数学研究的认知摩擦，使数学家能更高效地实验、协作与文献检索；他强调AI工具已进入‘主流应用阶段’，并呼吁共享探索路径以提升集体知识积累效率。
核心要点
- AI工具使数学家可跳过繁琐计算（如黑板推导后交由AI完成），大幅提升实验自由度
- 陶哲轩称其当前工作已转向‘大量AI辅助数学与协作项目’，并认为AI‘已准备好进入主流应用阶段’
- 他建议未来应公开研究过程中的多种探索路径（而不仅是最终成果），以增强知识复用价值
结构提纲
按章节快速跳转。
陶哲轩描述AI使其能尝试更激进的猜想，并将重复性计算任务交由AI处理，从而释放创造力。
过去所有智力任务都依赖人脑承担全部认知负荷，而AI正将这种摩擦降至接近零水平。
陶哲轩主张未来应公开研究过程中的多条探索路径，而不仅发布最终结果，以提升集体学习效率。
OpenAI关注的是赋能百名数学家自主突破，而非争夺诺贝尔奖或菲尔兹奖等个人荣誉。
思维导图
用一张图看清主题之间的关系。
查看大纲文本（无障碍 / 无 JS 友好）
- AI对数学研究的变革影响
- 认知摩擦降低
- 重复计算自动化
- 实验自由度提升
- 工作流重构
- 黑板推导 + AI补全
- 文献检索精准化
- 协作项目规模化
- 知识共享新规范
- 发布探索路径而非仅结果
- 促进集体智能积累
- 机构目标转变
- OpenAI：赋能百人而非争奖
金句 / Highlights
值得收藏与分享的关键句。
你可以在黑板上自由发散思路; 若遇到双方都不愿手动计算的部分，直接交给AI工具收尾即可。
我目前正开展大量AI辅助的数学研究与协作项目；现在我认为它已准备好进入主流应用阶段。
我们直到最近仍生活在一个‘认知摩擦’的世界里……如今AI及其他技术可将此类摩擦降至近乎为零。
我希望当AI使用普及后，人们不仅发布最终成果，也公开通往该成果的多种探索路径——这些信息同样极具价值。
视频笔记
[0:00] [music]
[0:02] > 我是 Terence Tao。我 是 IPAM 的项目 director， pure 和 应用 数学stitute。AI 已经 realatively been improving very rapidly。
[0:07] 它 has allowed我 to实验。我将尝试更 crazy ings.你可以 Ivan on the blackboard and then如果有计算 neither of us wants to do, we can just get our AI tool to finish that.
[0:18] >> [music]
[0:19] > 我可以搜索文献 much more accurately and effectively than I could before. So, I'm doing way more AIAssisted mathematics and and collaborative projects. And now I think it's ready for prime time.
[0:30] >> fundamental, at OpenAI, we care about being at the frontiers in terms of automating science, the economy, and ourselves. We care less about winning a Nobel Prize or a Fields Medal, and more about enabling 100 mathematicians out there to do that for themselves.
[0:45] > We lived in a world of cognitive friction until very recently, where every task required us to use our brain. And so, we didn't real about it. We just thought this was the cost of doing something intellectual.
[1:00] I hope when AI usage becomes more common, people will also post not just the final product, but all the different paths that they used to get there, because that's also very useful information.
[1:09] I think we can find some way to have the best of both worlds.
[1:10] >> [music]
[0:02] >> I'm Terence Tao. I'm director of special projects here at IPAM, the Institute for pure and applied mathematics. AI has real been realizing very rapidly.
[0:07] It has allowed me to experiment. I will try more eccentric things. You can Ivan on the blackboard and then if there's a calculation that neither of us wants to do, we can just get our AI tool to finish that.
[0:18] >> [music]
[0:19] >> I can search literature much more accurately and effectively than I could before. So, I'm doing way more AIAssisted mathematics and and collaborative projects. And now I think it's ready for prime time.
[0:30] >> fundamental, at OpenAI, we care about being at the frontiers in terms of automating science, the economy, and ourselves. We care less about winning a Nobel Prize or a Fields Medal, and more about enabling 100 mathematicians out there to do that for themselves.
[0:45] >> We lived in a world of cognitive friction until very recently, where every task required us to use our brain. And so, we didn't real about it. We just thought this was the cost of doing something intellectual.
[1:00] I hope when AI usage becomes more common, people will also post not just the final product, but all the different paths that they used to get there, because that's also very useful information.
[1:09] I think we can find some way to have the best of both worlds.
[1:10] >> [music]
[0:02] > 我是 Terence Tao.我 是 IPAM 的项目 director, pure and applied mathematics. AI has real been realizing very rapidly.
[0:07] 它 has allowed我 to experiment.我 will try more eccentric things.你可以 Ivan on the blackboard and then if there's a calculation that neither of us wants to do, we can just get our AI tool to finish that.
[0:18] >> [music]
[0:19] > 我可以搜索文献 much more accurately and effectively than I could before. So, I'm doing way more AIAssisted mathematics and and collaborative projects. And now I think it's ready for prime time.
[0:30] >> fundamental, at OpenAI, we care about being at the frontiers in terms of automating science, the economy, and ourselves. We care less about winning a Nobel Prize or a Fields Medal, and more about enabling 100 mathematicians out there to do that for themselves.
[0:45] >> We lived in a world of cognitive friction until very recently, where every task required us to use our brain. And so, we didn't real about it. We just thought this was the cost of doing something intellectual.
[1:00] I希望 when AI usage becomes more common, people will also post not just the final product, but all the different paths that they used to get there, because that's also very useful information.
[1:09] I think we can find some way to have the best of both ways.
[1:10] >> [music]
[0:02] >> I'm Terence Tao. I'm directed of special projects here at IPAM, thestitute for pure and applied mathematics. AI has real been realizing very rapidly.
[0:07] It has allowed us to experiment. I will try more eccentric things. You can Ivan on the blackboard and then if there's a calculation that neither of us wants to do, we can just get our AI tool to finish that.
[0:18] >> [ music]
[0:19] >> I can search literature much more accurately and effectively than I could before. So, I'm doing way more AIAssisted mathematics and and coll
[0:30] >> fundamental, at OpenAI, we care about being at the frontiers in terms of automating science, the economy, and ourselves. We care less about winning a Nobel Prize or a Fields Medal, and more about enabling 100 mathematicians out there to do that for themselves.
[0:45] >> We lives in a world of cognitive friction until very rapidly, where every task requires us to use our brain. And so, we didn't real about it. We just thought this was the cost of doing something intellectual.
[1:00] I希望 when AI usage becomes more common, people will also post not just the final product, but all the different paths that they used to get there, because that's also very useful information.
[1:09] I think we can find some way to have the best of both ways.
[1:10] >> [ music]
[0:02] > I'm Terence Tao. I'm directed of special projects here at IPAM, thestitute for pure and applied mathematics. AI has real been realizing very rapidly.
[0:07] It has allowed us to experiment. I will try more eccentric things. You can Ivan on the blackboard and then if there's a calculation that neither of us wants to do, we can just get our AI tool to finish that.
[0:18] >> [ music]
[0:19] >> I can search literature much more accurately and effectively than I could before. So, I'm doing way more AIAssisted mathematics and and coll
[0:30] >> fundamental, at open AI, we care about being at the frontiers in terms of automating science, the economy, and themselves. We care less about winning a Nobel Prize or a Fields])]) and more about enabling 100 mathematicians out there to do that for themselves.
[0:45] >> We lives in a]) Trent of cognitive friction until very rapidly, where every task requires us to use our brain. And so, we didn't real about it. We just thought this was the cost of doing something])]
[1:00] I])** when AI usage becomes more common, people will also post not just the final product, but all the different paths that they used to get there, because that's also very useful information.
[1:09] I think we can find some way to have the best of both ways.
[1:10] >> [ music]
[0:02] > I'm Terence Tao. I'm directed of special projects here at IPAM, thestitute for pure and applied mathematics. AI has real been realizing very rapidly.
[0:07] It has allowed us to experiment. I will])** more eccentric things. You can Ivan on the blackboard and then if there's a calculation that neither of us wants to do, we can just get our AI tool to finish that.
[0:18] >> [ music]
[0:19] >> I can search]) accurate and effectively than I could before. So, I'm]) way more AIAssisted mathematics and and coll]
[0:30] >> fundamental, at open AI, we care about being at the frontiers in terms of automating science, the]) and itself. We care less about winning a]) or a]) and more about]) 100 mathematicians out there to do that for themselves.
[0:45] >> We lives in a]) Trent of cognitive friction until very rapidly, where every]) requires us to use our]) and so, we]) about it. We just thought this was the cost of doing something])**]
[1:00] I]) when AI usage becomes more common, people will also post not just the final product, but all the different paths that they used to get]) because that's also very useful information.
[1:09] I think we can find some way to have the best of both ways.
[1:10] >> [ music]
[0:02] > I'm Terence Tao. I'm directed of special projects here at IPAM, thestitute for pure and applied mathematics. AI has real been realizing very rapidly.
[0:07] It has allowed us to experiment. I will])** more eccentric things. You can Ivan on the blackboard and then if there's a calculation that neither of us wants to do, we can just get our AI tool to finish that.
[0:18] >> [ music]
[0:19] >> I can search]) accurate and effectively than I]) before. So, I'm]) way more])Assisted mathematics and and coll]
[0:30] >> fundamental, at open AI, we care about being at the frontiers in terms of automating science, the]) and itself. We care less about winning a]) or a]) and more about]) 100 mathematical])** out there to do that for themselves.
[0:45] >> We lives in a]) Trent of cognitive friction until very rapidly, where every]) requires us to use our]) and so, we]) about it. We just]) this was the cost of doing something])]
[1:00] I]) when AI usage becomes more]) people will also post not just the final product, but all the different paths that they used to get])** because that's also very useful information.
[1:09] I think we can find some way to have the best of both ways.
[1:10] >> [ music]
[0:02] > I'm Terence Tao. I'm directed of special projects here at IPAM, thestitute for pure and applied mathematics.])** has real been realizing very rapidly.
[0:07] It has]) us to]). I will]) more eccentric things. You can Ivan on the black board and then if there's a]) that neither of us wants to do, we can]) get our]) tool to finish that.
[0:18] >> [ music]
[0:19] >>]) accurate and effectively than I]) before. So,]) way more])Assisted mathematics and and coll]
[0:30] >> fundamental, at open])isor, we care about being at the frontiers in terms of automating]) and itself. We]) less about winning a]) or a]) and more about]) 1) mathematical) out there to do that for themselves.
[0:45] >> we lives in a]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of]) Trent of