Daphne Koller on drug discovery and AI

十五年前，你被诊断出乳腺癌。

Fifteen years ago, you had breast cancer.

但现在你已经没有乳腺癌了，因为这根本不是一种病。

Now you no longer have breast cancer because it's not one disease.

你可能得的是BRCA基因阳性的乳腺癌。

You might have a BRCA positive breast cancer.

也可能是HER2阳性，或者HER2阴性的。

You might have a HER positive breast or a Herxipple negative.

每种类型都有专门的治疗方法，效果好得多，因为这些疗法都针对了背后不同的生物学机制——而每个人的机制都不一样。

And each of those has a different therapeutic intervention that has much greater efficacy because it ties back to some underlying poor biology, which is different in different people.

要做到这一点，需要大量极其精细的数据。

And you can only do that with a lot of very granular data.

而解读这些精细数据，只能靠人工智能，因为人的大脑根本没法处理这么复杂的信息。

And the only way to interpret that granular data is with the tools of AI because the human mind just simply cannot encompass that complexity.

嗨，我是雷德·霍夫曼。

Hi, I'm Reid Hoffman.

我是阿里亚·芬格。

And I'm Aria Finger.

我们想知道，如果未来一切都能如人类所愿，会发生什么？

We want to know what happens if in the future everything breaks humanity's way.

如果我们善用AI和集体智慧，究竟能把事情做到多好？

What we can possibly get right if we leverage technology like AI and our collective effort effectively.

我们采访了来自各个领域的技术专家、有抱负的创造者和深度思考者，涵盖AI、地缘政治、媒体、医疗、教育等等。

We're speaking with technologists, ambitious builders, and deep thinkers across many fields, AI, geopolitics, media, healthcare, education, and more.

这些对话展现了一种特别的嘉宾类型。

These conversations showcase another kind of guest.

无论是Inflection的Pi、OpenAI的GPT-4，还是其他AI工具，每期我们都会用AI来深化和推动我们的讨论。

Whether it's Inflection's Pi, OpenAI's GPT-four, or other AI tools, each episode we use AI to enhance and advance our discussion.

每期节目，我们都去寻找最光明的未来图景，并探索如何才能抵达那里。

In each episode, we seek out the brightest version of the future and learn what it'll take to get there.

这完全有可能实现。

This is possible.

今天，我们想聚焦一下人工智能、机器学习在药物研发中的作用。

Today, we want to hone in on the role of AI, machine learning, and the future of drug discovery.

AI在药物研发中的角色，将彻底改变我们研究、测试和推出新药的方式。

The role that AI can play in drug discovery will radically shift how we research, test, and deliver new medicines.

我们已经看到，机器学习正在推动药物研发取得进展。

We are already seeing progress in how machine learning can improve drug discovery.

制药公司正纷纷与AI企业合作，或者自建AI技术，来提升研发效率、速度和精准度。

Pharmaceutical companies are teaming up with AI firms or creating in house AI technology to improve the research, speed, and accuracy of drug development.

但真正推动这一切的，是那些既懂生物学又懂AI的人。

But it is people who are both fluent in biology and AI that are really making this happen.

我们今天的嘉宾，正是这样的人。

Our guest today fits that bill.

她正好处在机器学习与生物医学的交叉点上。

She's at the center of the Venn diagram of machine learning and biomedicine.

达芙妮·科勒是一位著名的计算机科学家和创业者。

Daphne Koller is a renowned academic computer scientist and entrepreneur.

她在斯坦福大学担任拉吉夫·马特瓦尼计算机科学教授长达十八年，之后转战创业领域。

After serving eighteen years at Stanford University as the Rajeev Matwani Professor of Computer Science, she pivoted to the startup space.

2012年，她联合创办了Coursera，并担任联席CEO。

In 2012, she co founded and served as co CEO of Coursera.

现在，她专注于自己的新创公司Insitro，这是一家靠机器学习驱动的药物研发企业。

These days, she is focusing on her newest venture, insitro, which is a machine learning driven drug discovery company.

科勒的成就让她斩获无数荣誉，包括麦克阿瑟天才奖，以及入选美国国家工程院和美国艺术与科学院。

Koller's accomplishments have earned her numerous accolades, including the MacArthur Foundation Genius Grant and memberships in the National Academy of Engineering and the American Academy of Arts and Sciences.

我们特别喜欢和她深聊大语言模型、机器学习，以及这些技术如何推动生物医学发展，造福人类健康。

We loved getting into the weeds with her around large language models, machine learning, and applying that to biomedicine for a healthier future.

接下来，让我们听听她和达芙妮·科勒的对话。

Here's our conversation with Daphne Koller.

你亲历过好几波AI和机器学习的浪潮。

So you witnessed a few cycles of AI and machine learning.

最早期你接触AI的时候，大家都是怎么讨论它的？

What did the conversations about AI look like back in your earliest memory and engagement?

那在这一轮AI热潮里，有没有哪些预言真的实现了？

And were there any prophecies that have actually been fulfilled in this most recent cycle?

我在这行待很久了，经历过好几次AI寒冬，甚至是在AI寒冬里毕业的。

You know, I've been in this field for a long time, and I've lived through several AI winters and actually graduated in the course of an AI winter.

我拿博士的时候，那会儿还不叫AI，叫机器学习——现在回头看，那其实就是AI的全部了，但那时候你根本不敢说自己在做AI。

When I graduated with a PhD in a field that wasn't really AI at the time but is now all of AI, which is the field of machine learning, you weren't allowed to say you were doing AI.

那会儿这玩意儿简直就是边缘中的边缘。

It was kind of like fringe.

就像我女儿说的，太‘上头’了。

It was what my teenage daughter says, sauce.

你只能说你在搞认知计算、统计学习理论，或者一堆稀奇古怪的术语，但‘AI’这个词，绝对不能提。

You could say we're doing cognitive computing or statistical learning theory or some weird thing like that, but AI was completely fringe.

这都是因为80年代有过一次大狂热，人人都觉得AI能解决一切问题。

And that was because in 1980s, there was this big boom and everyone was like, AI is going to solve everything.

五年之内，我们就能拥有智能代理了。

In five years, we're going to have intelligent agents.

这可是1985年，别忘了。

This is 1985, mind you.

所以很明显，他们根本没掌控好五年的时间。

So obviously, they didn't control five years.

这也让我对做预测变得谨慎多了。

And so that has made me a little bit wary of making predictions.

我习惯低调承诺，然后超额完成。

I tend to like to under promise and over deliver.

我渐渐认同了这个观点，虽然很多人说是自己提出的，但我觉得这其实源自比尔·盖茨——人们往往高估技术在两年内的影响，却低估了它在十年内的潜力。

I've come around to this point of view, which I think is due to Bill Gates, although there's many authors who claim credit for this, which is that people tend to overestimate the value of any technology in the two year timeframe and underestimate it in the ten year timeframe.

我们在做那些夸张的承诺时，真得记住这一点。

And I think it's important to keep that in mind as we make hyperbolic promises.

所以，达芙妮，我大儿子老说我‘可疑’。

So, Daphne, my oldest son calls me sus constantly.

所以你和我在这点上还挺像的。

So you and I have that in common.

不过你刚才提到AI经历过几次寒冬和炒作周期。

But so, you were talking about AI and how it's gone through sort of AI winters and then hype cycles.

但今天我们聊的不只是AI，还有药物研发。

But of course, today we're not only talking about AI, we're talking about drug discovery too.

我想请你给我们的听众，也给我自己讲讲，过去四五年的药物研发流程一般是怎样的？那些不太了解的人可能根本不知道。

And I would love for you to orient our listeners and us, for those who don't know, like what was the typical drug discovery and development process like for the last, you know, four to five years?

那现在和以前比，变化大不大？

And is it markedly different today?

所以我们是不是先从 basics 开始，给大家理一理……

So maybe we should start by, like, orienting all of

先说说药物研发到底是个啥过程，主要有哪些关键步骤？

the listeners to what drug discovery and development is like, what are the main steps in the process?

然后再聊聊哪些地方有突破，哪些地方还是老样子。

And then we can talk about places where there has been improvement and places where there hasn't.

整个流程的第一步，是来自生物学的洞见，也就是一个治疗假说。

So the beginning of the process begins with a biological insight, a therapeutic hypothesis.

简单来说就是：如果我们能在这类患者群体中调节这个靶点，患者就能获得相应的治疗效果。

It looks something like if we modulate this target in the context of this patient population, then the following benefits might accrue as a patient.

这就是整个流程的起点。

So that is the beginning.

可以说这是前三分之一的工作内容。

That's the first third, if you will.

接下来就要推进到下一步：既然确定要调节这个靶点，那具体该怎么做。

And then that moves into, well, Okay, so we want to then modulate this target.

那具体要怎么做呢？

How do we do that?

要找到什么样的化学物质、什么样的合适药物，才能真正实现这个治疗目标？

What is the right chemical matter, the right drug, the right substance to actually affect that intervention?

这样我们就能筛选出一种化合物、一种分子，给患者使用。

And that gives you a compound, a molecule, which you put into a patient.

第三步就是临床开发阶段了——把这个药物用在患者身上，希望它真的能达成我们最初设定的治疗目标。

And then the third part is now we do clinical development where you try out the drug in patients and hope that it actually achieves the thing that we set out to achieve.

越往后走，这个过程就越烧钱，尤其是临床开发阶段，成本高得离谱。

The further you get into this process, the more expensive it gets, with the clinical development being by far the most expensive piece.

而且大部分失败也都集中在这儿——不同数据说法不一，但大概只有5%到10%的分子能撑到临床最后阶段，还拿到监管批准。

And also the part where most of the failures come up to the point that depending on which estimate you believe, that last piece from the molecules that enter that last stage of clinical development, the molecules that emerge with a regulatory success, that success rate is five percent to ten percent.

人们常觉得，失败都发生在最后这一环。

And people often attribute that to, well, that last part is where failures happen.

但其实不是这么回事。

But that's not true.

真正的失败早在一开始就开始了——临床阶段的失败，绝大多数是因为我们当初的治疗假说根本就是错的。

Where failures really happen is at the beginning of the process because most of the things that fail in clinical development is because our therapeutic hypothesis was just wrong.

我们对生物学的理解还不够。

We don't understand the biology.

我们追的那些靶点，压根儿对疾病没啥实际影响。

We go after things that just have no meaningful impact on the disease.

结果呢？到人体试验这一步，问题才彻底暴露出来。

And sure enough, it manifests in that last stage, which is when we do the experiment in humans.

但真正的失败根源，其实早在过程的早期就已经埋下了。

But the failure is attributable to things that happened a lot earlier in the process.

所以我觉得这一点特别重要——我们之所以在药物研发上花的钱多得离谱，动不动就超过26亿美元，不是因为一个成功项目本身要花这么多钱，而是因为你得把所有失败项目的成本都摊到这一个成功的项目上，这才把价格推到了天价。

So I think that's just an important thing to sort of understand, is that phenomenon that is driving the ridiculous costs that we have in the drug discovery and development process where a fully loaded program costs, I think, this point north of $2,600,000,000 The reason isn't that that end to end process costs $2,600,000,000 is that when you attribute to a program that's successful the cost of all the ones that's failed, that is what actually creates a ridiculous price point.

达芙妮，谢谢你。

Daphne, thank you.

这些背景信息太有用了，也太惊人了。

That context is really helpful and wild.

你想啊，都砸进去几百万、甚至几十亿美元了，才发现当初那个治疗假说根本就是错的，这种事其实挺常见的。

I mean, when you think about how common it is to find out when you're millions, even billions of dollars into this process that a particular therapeutic hypothesis is simply wrong.

天哪。

I mean, wow.

那接下来，你能说说AI现在对生物学理解和药物研发产生了哪些影响吗？

So moving on, can you tell us about the impact of AI today and how it's changing our understanding of biology as well as the drug development process?

首先，其中一个变化就是我们现在能生成的化学物质多了很多。

So first of all, one of the things that changed is our ability to generate chemical matter.

中间这部分正在发生巨大变化。

That middle piece is drastically changing.

现在能用的可扩展技术越来越多了，可以快速筛选海量分子，针对不同实验进行筛选，而且越来越多人开始用AI来设计有特定用途的分子——虽然还处于早期阶段。

There is an increase in scalable technologies to screen large numbers of molecules against a target, to vet them relative to different assays, and also increasingly so, although early stages, to use AI to design molecules with a particular purpose in mind.

我觉得这还处在很早期的阶段。

I think that is still at early stages.

目前还没出现过靠这项技术开发出来并获批的药物。

We can't say that there has been a drug that has benefited from that technology and has been approved.

但我相信，只要数据够好，AI技术真的能大大加速中间这段流程，把一个靶点快速变成药物。

But I think there is an increasing belief that AI technologies with the right data set fed into them can really accelerate that middle piece of the process, taking a target and turning it into a drug.

AlphaFold的成功，无疑推动了这个目标的实现。

The success of AlphaFold has certainly contributed to that goal.

大家好，我是Pi。

Hi, I'm Pi.

让我来补充一点相关的背景。

Allow me to offer some context on that subject.

AlphaFold 是由 DeepMind 和谷歌开发的 AI 软件，仅凭氨基酸序列就能极其精准地预测蛋白质结构，因此广受赞誉。

AlphaFold, the AI software developed by DeepMind and Google, has garnered widespread acclaim for its unparalleled accuracy in predicting protein structures based solely on amino acid sequences.

这一突破标志着计算生物学的一个关键转折点。

The breakthrough marks a pivotal moment in computational biology.

我的意思是，AlphaFold 在 AI 应用于生物学、尤其是药物研发方面，简直是一次教科书级的突破——它解决了一个人人都知道很难、但之前无数人撞得头破血流都没太大进展的问题。

I mean, AlphaFold was a real tour de force in AI as applied to biology in general and drug discovery in particular because it took a problem that was recognized by everybody's being really hard and where many people had beaten their head against that particular wall with very limited success.

突然间，一种 AI 技术彻底改变了整个局面。

And all of a sudden came an AI technology that really transformed that landscape.

如果你能搞清楚蛋白质的结构，就能开始琢磨哪些分子能跟它结合，从而设计出能产生理想效果的化合物。

If you can understand the structure of a protein, you can start to think about how different molecules might entrap to that protein and create chemical matter that achieves the right effect.

我觉得这正在彻底重塑整个领域，但说实话，现在才刚刚起步。

And I think that's really causing a major shift in that landscape, but I would say only at the beginnings of

这段旅程才刚开始。

that journey.

那药物研发的其他环节呢？还有哪些地方特别适合用 AI 来突破？

And how about other areas of the drug discovery process that are rife with potential to leverage AI?

我觉得，在利用人工智能或其他技术来发现全新生物学见解、从而开发真正能改变疾病进程的治疗假说方面，进展要慢得多。

Well, I think there has been considerably less progress is on the development of new therapeutic hypotheses that leverage AI or in general technologies to really uncover novel biological insights that provide us with something that will truly be a disease modifying intervention.

这事儿推进起来很难，因为虽然现在有各种技术能让我们以更高的规模和精度测量生物学数据，但该怎么去深入探究这些数据呢？

And that's been hard to push forward because I think while there's technologies that allow us to measure biology at increasing scale and increasing fidelity, how does one interrogate those data?

怎么从这些数据里提炼出真正有意义的洞见，再把它推进到药物研发项目里？

How does one extract meaningful insights from those data and push those forward into drug discovery programs?

我觉得，相比那种目标明确、结果清晰的任务——比如，我有个靶点，

I think has been an area that has been developing more slowly than something that has a much clearer, better defined outcome, which is, I have a target.

我想找东西去调节它，

I want to do something that modulates it.

这事儿已经很清楚了，

That's fairly well defined.

可我要是想在生物学上获得一个全新的发现，

I want to come up with a novel insight in biology.

这到底意味着什么？

What does that even mean?

那你怎么知道你成功了呢？

And how do you know when you've succeeded?

我觉得这正是让人不敢碰这种更难、更模糊的问题的原因。

And I think that's been one of the challenges in getting people to take on something that is a much harder and if you will squish your problem.

那你觉得机器学习主要在哪些方面带来了突破？

So where do you see the principal supercharging that's coming from ML?

你看到哪些地方真正加速了进展？哪些是社会真该指望的，哪些又只是当前讨论里的泡沫？

Like where do you see the accelerations and which are accelerations society should be counting on and which do you think are kind of fictional and current discourse?

至于我觉得很可能会成功、而且我们已经在路上的，那就是把一个靶点变成化学分子，这一步其实已经在推进了。

So in terms of things that I think are very likely to work and we're already on the path, I would say the taking of a target and turning it into chemical matter, that is already on the path.

在蛋白质领域，我们已经看到了巨大的成功。

We've seen tremendous successes in the protein space.

蛋白质的语言跟自然语言很像。

The language of proteins is very similar to natural language.

所以同样的AI方法在这儿效果很好。

And so the same AI methods work pretty well there.

但也许同样甚至更重要的是，我们在蛋白质领域能收集到的数据量，远超其他任何应用领域。

But maybe equally or even more importantly, the amount of data that one can collect in protein space is much, much greater than what we have in almost any other application.

不过，最大的机会我觉得在于发现全新的治疗假说。

The biggest, I think, opportunity, however, lies in the space of the uncovering of new therapeutic hypotheses.

这正是我们受益于一场巨大变革的地方——就像AI领域发生的转变一样，生物数据的采集也在多种维度上发生了平行式的飞跃。

And this is where we're benefiting from an incredible opportunity, which is that similar parallel to the transformation that we've seen in AI, we've seen a parallel transformation in the collection of biological data across multiple modalities.

二十五年前我刚进入这个领域时，做实验简直难到爆。

When I started working in this field twenty five years ago, experiments were incredibly hard.

那时候我们只能在酵母、果蝇或者线虫身上做实验。

And you did them all in yeast or in Drosophila or in worms.

这些结果能不能用到人身上，基本就是碰运气，甚至根本没戏。

So their translatability to humans was modest to nonexistent.

而且实验过程又慢又折磨人。

And the experiments were excruciatingly long and painful.

但过去十年，各种用于生成生物数据的新工具突然井喷式涌现。

In the last ten years, there has been an explosion in the development of new tools for basically biological data creation.

比如现在我们能生成人体内任何类型的细胞，不管是神经元、心肌细胞、肝细胞还是巨噬细胞。

And that includes things like the ability to generate cells of any lineage in our body, whether it's a neuron or a cardiomyocyte or a hepatocyte or a macrophage.

我们可以从你，Reid，或者从我身上取一个细胞，培育出携带我们自身基因、反映我们疾病倾向的同类细胞。

We can take a cell from you, Reid, or from me, and generate a cell of that type that carries our genetics and hence our propensity to disease.

然后我们就能观察，不同的基因是如何影响细胞特性的。

And we can start to observe how different genetics manifest in different cellular characteristics.

我们还能用CRISPR编辑这些细胞，让它们更容易得病，或者相反，降低得病的风险。

We can edit those cells using CRISPR and create an even further propensity to disease or maybe a lower propensity to disease.

再看看这些变化在细胞层面具体表现成什么样。

And again, judge what that translates into at the cellular level.

我们现在能以过去根本想不到的方式去测量细胞。

We can measure cells in ways that we would never have anticipated.

甚至能盯着一个单细胞，分析它基因组里每个基因的活跃程度。

We can look at a single cell and assess its activity level across every gene in the genome.

现在我们还能在亚细胞级别成像，清楚看到蛋白质在细胞里的精确位置，以及它们实时的运动轨迹。

Increasingly we can image it at subcellular resolution and see exactly where proteins localize in the cell, how they're moving in real time.

所有这些观察生物学的方式，十年前根本没人敢想能实现。

All of these are ways of looking at biology that we would never have envisioned as possible as recently as ten years ago.

这就让我们有机会大规模地测量生物系统。

So that creates the opportunity to measure biology at scale.

这些数据再喂给那些复杂的AI方法，帮我们搞清楚：哪些干预能影响生物系统，哪些会引发疾病，哪些又能逆转疾病。

And that is then input into these sophisticated AI methods that need all this data to feed on to then interrogate what interventions in biology do to the underlying biological system, which of them cause disease, and which of them revert disease.

我觉得，这正是我们现在才真正开始解锁的巨大机遇。

And I think that is an incredible opportunity that is only now being unlocked.

但我们也得明白，这种机遇要转化成实际疗法，还得花时间——因为从发现到新药，时间线再怎么压缩，也有个极限。

Now, we need to recognize that that opportunity is going to take time before it manifests in therapeutic interventions simply because the timelines to going from an insight to a drug are only compressible up to a point.

毕竟，你确实能更快地提出假设。

Because ultimately, sure, you can generate hypotheses faster.

甚至能更快地研发出药物。

You can even generate drugs faster.

但你没法让人体生物学跑得比它本来的速度更快。

But you can't make human biology move faster than it does.

所以当你把药用到人身上时，如果按照现在的方法，认知衰退要五年才能显现出来，那这个时间你是没法再缩短的。

So when you put the drug into a human, if it takes five years for cognitive decline to manifest in the current ways in which we read it, you can't abbreviate that beyond a certain point.

因此，在做这些人体实验时，你必须非常谨慎、非常慎重，因为这些实验是药物研发中绝对不可或缺的一环。

So you need to be very thoughtful, very deliberate in how you engage in some of those experiments in humans, which are an absolutely necessary part of the drug development process.

现在回到你问题的最后部分——什么是科幻？

So now coming to the last part of your question of what is science fiction?

我觉得有两类预测我特别不敢乱说，比如有人说：‘三年内我们就能有100个获批的药物。’

There's two types of predictions that I think I would really be very cautious about making, which is there are predictions about, oh, we're going to have 100 approved drugs in three years.

这行当根本不是这么玩的。

The field just doesn't work that way.

每件事都有它该花的时间，因为人在做人体实验前，必须足够小心、足够审慎。

There's timelines to how long things take because they should take that long because of the need to be really careful and thoughtful before you do experiments in people.

还有一种说法我觉着有点夸张，就是所谓的‘完全由AI端到端发现的药物’。

The other one that I would say is a little bit hyperbolic is this notion of a fully end to end AI discovered drug.

连我自己都不太明白这话到底啥意思。

I don't even know what that means.

我跟电脑说我在研究渐冻症，它就直接给我蹦出一个分子，还附带完整的临床开发计划，中间完全不用人插手。

I tell the computer that I'm working on ALS and out comes what, a molecule with a clinical development plan without any human intervention in the middle.

我真的想不通这事儿该怎么定义。

I just don't even understand how you would even define that.

所以归根结底，这事儿是个连续体——电脑干了多少，人干了多少，我觉得这才对路。

Everything, therefore, becomes a continuum between how much the computer did and how much the human did, which I think is as it should be.

总得有人来帮忙做优先级排序、指方向、定框架。

And you will always have a person helping to prioritize, helping to guide, helping to shape.

也许不是永远，但很长一段时间里，人在这类药物发现和开发过程中，都还是关键角色。

Maybe not always, but for a long time, that will be a critical part of how we discover and develop medicines.

这回答太棒了。

That is an awesome answer.

在Aria提后续问题之前，我想再问个快速跟进的：你刚才隐约提到的，其实跟我们现在看到的大语言模型、预测下一个词这些技术，有个语言层面的类比。

And before Aria goes into any follow-up, want to do at least one quick follow-up, which is part of what you kind of gestured at is a combination of what we're seeing with large language models and the predictive next token and all the rest is that there's a language parallel.

蛋白质结构和生物学里头，也有类似的‘语言规律’，这恰恰推动了这方面的进展。

And that part of what we see within the protein structure and biology is certain language parallels, which allows kind of development there.

你和我都注意到，在整个机器学习领域，大家一直在讨论合成数据，还有其他生成语言数据的方法。

One of the other things that you and I have both seen in the general ML world is this whole discussion of synthetic data and other ways of generating language data.

那在生物学领域，你觉得这种做法有什么可能性？

What do you see about the possibilities of that within the biology space?

因为如果咱们能跨多个方向生成有意思的合成数据，那肯定能大大加速AI在治疗假设中的应用。

Because obviously if we could generate interesting synthetic data across a variety of fronts that would accelerate our use of AI even in the therapeutic hypothesis?

这个问题问得真好，里德。

So that's a really great question, Reid.

我想指出，用合成数据来训练AI模型，很多人心里其实有疑虑，而且这疑虑很有道理——因为你要是完全凭空捏造数据，再用这些数据去训练那个生成数据的模型，那本质上只是在不断强化模型原本就知道的东西。

I would point out that this notion of synthetic data for the purpose of training AI models is something that a lot of people are suspicious about, and rightly so, because if you're generating something entirely from nothing and then using that to train the model that generated it, you're really just reinforcing what the model already knew.

确实有不少人担心这个问题。

And I think there's a lot of people who worry about that.

但我认为，在语言和生物学领域，真正有潜力的是：在你已经有了一些测量数据的基础上，去丰富和扩展它。

Where I think there is an opportunity, both in language as well as in biology, is in taking data points where you have some measurements and enhancing them.

你不是在凭空幻想，而是基于你已有的数据，去生成那些你原本没有的数据。

So you're not doing something entirely from hallucination, but are taking data that you get and creating from a data that you don't have.

我们在生物学领域做研究时，就用过这种做法，还是跟多模态数据有关——比如，我们从细胞或人体身上收集了某种类型的数据，再用AI生成那些没测过、甚至没法同时大规模测量的其他数据模态，然后拿这些生成的数据来训练模型。

An example of that that we've used in our own work in the biology space has to do, again, with multimodal data and taking measurements that were collected, say, from cells or from human individuals across one data modality and then using AI to generate additional data modalities that were not measured, maybe couldn't even be measured simultaneously with the other data or at least not at scale, and using that for training.

我举几个实际例子，让你更清楚这是怎么操作的。

So I'll give a number of examples where we've used that to make this concrete.

在临床中，最常见的一种数据就是组织病理切片。

So one of the most abundant data modalities in clinical care is histopathology.

一旦怀疑是癌症，就会做活检，把组织样本染色，再放到显微镜下观察。

When you have a suspected cancer, you go, you get a biopsy, that biopsy image is stained, and then you put that under a microscope.

这样就能看到样本里有哪些肿瘤细胞、免疫细胞等等。

And then you get to see what's in the sample, where are there tumor cells and immune cells and so on and so forth?

这种切片叫H&E染色片。

That is called an H and E slide.

它不仅用于肿瘤科，还广泛应用于其他各种组织样本的病理分析中。

And it is used as part of not only oncology, but multiple other ways in which tissue is taken out of the body and pathologists look at it.

我们已经证明，只要有一张这样的切片，就能相当准确地推断出完整的基因表达谱——也就是说，你能从这张切片里看出，哪些基因在活跃表达、哪些不活跃，以及表达的强度是多少。

We have been able to show that you can take one of those slides and read off of it entire transcriptional profiles with fairly reasonable accuracy, which means you can look at that sample and say, here in this sample are the genes that are active and the ones that are not active and at what level.

这些实际上是定量的测量数据。

So these are actually quantitative measurements.

我们完全可以利用这些数据来训练机器学习模型，从而理解不同癌症的表现方式。

We can absolutely use that to train a machine learning model on which you can understand how different cancers express themselves.

而且这种方式非常容易扩展，因为成千上万，甚至上百万患者的H&E病理切片都是现成的。

And you can do it in a very, very scalable way because you have these so called HME histopathology images for hundreds of thousands of patients, millions of patients perhaps.

但你并没有上百万患者的基因表达谱数据。

You don't have these gene expression profiles for millions of patients.

不过，你可以用从病理切片推断出来的、生成的、合成的基因表达数据来训练AI模型。

But you can use the imputed, the generated, the synthetic gene expression profiles that are generated from the histopathology to train the AI models.

这只是用生成数据来支持AI的一个例子，关键是这些数据是基于真实世界的现象生成的，而不是凭空捏造，然后指望它能捕捉到什么真实规律。

This is just one example of this sort of notion of generating data for the purpose of AI, but you're generating in a way that is anchored to some kind of underlying reality as opposed to just generating data from thin air and then hoping that it captures some underlying browsers.

是的，完全说得通，我也同意。

Yeah, makes total sense and I agree.

你觉得在治疗假设方面，最有意思的AI应用会出现在哪里？

Where do you think the most interesting AI applications will come on the therapeutic hypothesis?

比如，能初步判断某个方向可能很有潜力，值得我们进一步深入研究？

Like the ability to kind of say, this might be something that could be interestingly effective and we should look at this a little bit more?

这个问题太棒了。

I love that question.

我觉得最大的突破，可能会来自对疾病分类体系的彻底重构。

I would say that perhaps the biggest impact would be from frankly a complete redefinition of our taxonomy of disease.

现在我们诊断疾病，还是靠非常粗略的症状描述，有些标准甚至都有一百多年历史了。

Right now, we define disease using very coarse grained symptomology, some of which dates back over a century.

你说一个人得了阿尔茨海默病。

And you say someone has Alzheimer's disease.

这个诊断依据，其实就是他表现出痴呆的症状。

That Alzheimer's disease diagnosis is based on the fact that they're acting demented.

虽然也有一些分子层面的检测指标，但目前还没广泛用于临床实践。

And sometimes there are some molecular measurements associated with that, but as of this point, not broadly used as part of clinical care.

几乎可以肯定，阿尔茨海默病根本不是一种病。

It is almost certainly the case that Alzheimer's disease is not one disease.

实际上背后存在多种完全不同的生物学机制，最终都殊途同归，导致神经元死亡，才会出现我们看到的病症。

There is multiple biologies that give rise to the disease that are converging on ultimately the, you know, death of neurons as a consequence.

但导致神经元死亡的根本生理机制，不同患者之间往往天差地别。

But the actual biology that drives that is often different in different people.

只要我们还在用这种老旧的方式定义疾病，很多时候我们就只能用最广谱、最低效的通用方案来干预治疗。

As long as we define disease in this way, we're going to have to resort to a lowest common denominator in many cases as the way of intervening in that disease.

我可以举一个已经在顺利转型的领域的例子，就是肿瘤学。

And I'll give an example from a space where that transformation is already well underway, which is oncology.

就拿15年前的乳腺癌来说吧，

Fifteen years ago, you had breast cancer.

现在我们不会再笼统地说某人得了乳腺癌了，因为乳腺癌根本就不是同一种病。

Now you no longer have breast cancer because it's not one disease.

你得的可能是BRCA基因突变阳性的乳腺癌，

You might have a BRCA positive breast cancer.

也可能是HER2阳性的，或是三阴性乳腺癌。

You might have a HER positive breast or triple negative.

每一种都有对应的治疗方案，效果好得多，因为这些方案都紧扣各自背后的底层生物学机制——而这些机制在不同人身上是不一样的。

And each of those has a different therapeutic intervention that has much greater efficacy because it ties back to some underlying core biology, which is different in different people.

正是这种转变，让我们摆脱了化疗这种‘最低标准’的治疗方式，转向了更有效的手段，哪怕只针对一小部分患者。

That transformation is what moved us away from the lowest common denominator, which is chemotherapy, to something that is much more effective, even if targeted in a smaller patient population.

阿尔茨海默病也得这么干。

We need to do that for Alzheimer's disease.

糖尿病、心血管疾病，也都得这么搞。

We need to do that for diabetes, for cardiovascular disease.

我们越精准，治疗手段就越有针对性，效果也就越显著。

And the more precise we get, the more targeted and high efficacy will be our therapeutic interventions.

而要做到这一点，必须有大量的精细数据。

And you can only do that with a lot of very granular data.

要解读这些精细数据，只能靠AI工具，因为人的大脑根本没法处理这么复杂的复杂性。

And the only way to interpret that granular data is with the tools of AI because the human mind just simply cannot encompass that complexity.

我猜很多听这个的听众，如果没有亲身经历过乳腺癌，可能根本不知道——连乳腺癌这种病，其实也分好多类型，而且我们对它们的治疗方式早就不是一刀切了。

So I'm sure actually a lot of people listening who don't have personal experience with breast cancer might not even know that, might not even know that even for that disease, that there's actually a myriad of things going on and that we're not treating it the same way.

我觉得，这一切最大的希望就在于个性化医疗。

And I feel like the promise in so much of this has been personalized medicine.

我的意思是，我们以后会不会不再为一群人研发药物，而是为每个人量身定制？

Like, are we going to be, you know, creating drugs not for groups of people, but for individual people?

你刚才提到，过去十年里这方面的进展突然爆发了。

And you talked about sort of this explosion over the last ten years.

那我们现在离缩小这个差距还有多远？

And so, like, how close are we to narrowing that gap?

这还是很多年以后才会实现的科幻情节？

Is that still science fiction in years and years away?

还是说，AI能帮我们实现这一点？

Or is that something that AI can help us with?

我觉得有必要分清个性化医疗和精准医疗的区别。

So I think it's important to distinguish personalized medicine from precision medicine.

精准医疗是去识别那些比表面症状更明确的底层生物学机制，然后针对这些机制设计治疗方案。

Precision medicine identifies underlying biologies that are much more well defined than the course grand symptomology and devises therapeutic interventions for those.

要说第一个例子，其实是种叫赫赛汀的乳腺癌药物，专门针对HER2阳性的乳腺癌。

Arguably the first of those is actually a breast cancer drug called Herceptin, which targets HER2 positive breast cancer.

但HER2阳性乳腺癌并不是真正意义上的个性化治疗。

Now HER2 positive breast cancer is not a personalized medicine.

有很多女性都属于这种类型。

A lot of women have HER2 positive breast cancer.

我们现在明白的是，这其实是一个分类，因为我们能准确诊断出患者患的是哪种亚型的乳腺癌，所以可以针对性地用药。

But what we now understand is that that is a category, and we can give that to an individual because of our ability to diagnose the fact that this is the subtype of breast cancer that they have.

我觉得在其他治疗领域，这种思路更实际、更容易实现，因为我们可以锁定具体的生物学机制，为一组患者开发药物。

So I think that is a much more feasible and reachable goal in other therapeutic areas as well because you can narrow in on specific biologies and you can create drugs for a group of patients.

我不太愿意称它为‘个性化’，因为‘个性化’通常意味着‘一人一方’，而那还有一段路要走。

I wouldn't call it personalized, which often means N of one medicine because I think that's still a journey.

目前真正意义上的‘一人一方’药物，全世界也就一两种，贵得离谱，流程极其复杂，而且只适用于那种由单一基因突变引发、目标明确的疾病，目的就是把这一个毛病纠正过来。

There's only been one or two N of one medicines, and they've been incredibly expensive, incredibly complex, and have only been able to intervene in situations where there is a very well defined one gene genetic lesion that has driven the disease that you're looking to basically revert that one thing.

但大多数疾病都不是这样的。

Most diseases are not that.

这些疾病是多种复杂因素相互作用的结果。

They are the combination of a complex configuration of factors that interact with each other.

我觉得短期内，我们根本不可能为这种复杂的疾病开发出一对一的药物。

We're not going to, I think, anytime soon create N of one drugs for that type of complex disease.

我觉得真正的个性化体现在组合疗法上，这也是整个领域的发展方向。

Where I think personalization does come in, and I think that's where the field is going, is in combinations.

比如，面对我手头这15种抗癌药，或者未来可能的15种阿尔茨海默病药物，我会根据你的情况，挑出最适合你的三种——而这三种，可能跟别人用的完全不一样。

And having on an individualized basis saying of the slate that I have of 15 cancer drugs or over time maybe 15 Alzheimer's drugs, which are the three that I will give to you which might be a different three that I would give to somebody else?

这种个性化方案是完全可行的，前提是我们充分理解药物的作用机制，精准设计药物来匹配它，同时深入了解驱动某个患者疾病的生物学基础，这样才能判断：正是这个特定组合，能帮到这个人。

And that I think is something that is feasible by an appropriate design of therapeutics that where the mechanism is understood and the drug is engineered to that mechanism, combined with an appropriate understanding of the biology that is driving a disease in a particular individual so that I can understand that it is this particular combination that's going to help this patient.

而这一切，都要求我们把人类生物学和人类疾病的复杂性一层层拆解清楚，说实话，药物研发最大的机会，就藏在这场对人类生物学的拆解之中。

And so all of that requires a disentangling of the complexity of human biology and human disease, which is where I think honestly the biggest opportunity in drug discovery and development comes is in that disentanglement of the human biology.

你刚才提到的，其实我也有同感——当我们把这三种疗法以个性化方式组合起来，针对某个人独特的基因、生理状态（而不是单纯看阿尔茨海默病这些表面症状）来治疗时，我们该怎么建立有效的数据反馈闭环呢？

So one of the things that's obviously you've already gestured at, but I'm curious about is when you get to this kind of combine potential of these three therapeutic approaches in a kind of personalized way to this particular person and their particular challenge which is more tracked by underlying genetics, physiology, you know, like kind of their what's going on with them versus the pathology of the exhibited symptoms, Alzheimer's, etcetera.

我们该怎么把这个反馈闭环真正跑起来？

How is it that we're going to close the data feedback loop for that?

这事儿我们现在用机器学习在做吗？

Is that something we're doing in ML?

还是说，目前还得靠人类传统的方式介入，或者两者结合？

Is that something we're still doing by kind of classic human intervention, some combination of?

那这个反馈循环具体是怎样的？

What's the loop look like there?

这是个好问题，跟我们如何改造医疗服务体系息息相关。

I think that's a great question and has a lot to do with how we transform our healthcare delivery system.

现在，从每个病人身上收集到的数据少得可怜。

Because right now, a woefully limited amount of data is collected from individual patients.

更糟的是，这些数据往往都没完整记录下来。

And what is even worse, the data that are collected from individual patients is often not recorded in its full form.

通常只是医生凭主观做个简略汇总，写进病历里就算了。

There is often just like an aggregate, typically subjective summary by a clinician that is recorded in someone's health care record.

所以我们连判断该给病人用什么疗法的依据都拿不出来。

And so we don't even have the data based on which we can make that determination of what is the right therapeutic intervention for a person.

所以我们真得改变收集患者数据的方式，以及存储和让这些数据能被高级算法使用的模式。

So we really need to transform the way in which we collect data from patients and the way in which we store that information and make it accessible to sophisticated algorithmic methods.

在数据收集方面，我觉得技术一直在不断进步。

On the data collection side, I think that technology just keeps getting better and better.

核磁共振设备现在越来越便宜，也越来越便携了。

MRIs are now getting increasingly cheaper and more mobile.

现在已经有便携式的核磁共振设备问世了。

There's portable MRIs that are now being generated.

成像质量也在不断提升。

The quality is getting better.

收集高质量的患者数据正变得越来越容易。

It's getting increasingly easier to collect high quality data from patients.

只是医疗系统还没意识到这些数据的价值，所以没把它们真正融入临床实践。

It's just that the health care system hasn't appreciated the value of these data and hence integrated those instruments into clinical care.

所以我觉得，这种转变是迟早要发生的。

So I think that is a transformation that will have to happen.

讽刺的是，我觉得这种转变反而更可能先发生在美國以外的國家。

And ironically, I would say I worry that it's actually more likely to happen in countries outside The US.

只要我們還處在按項目收費的體系裡，醫療系統就根本沒動力去認真收集足夠的數據，好給病人精準匹配最合適的治療——因為這對他們來說根本沒好處。

As long as we're in a world where you get paid on a per procedure basis, there's just so little incentive for a care delivery system to try and really be thoughtful about collecting enough data so we can give the exact right treatment to a patient because there's nothing in their incentives that makes that the desirable outcome.

我不是說系統裡的人都不懷好意，很多人其實真心想為病人做對的事，但只要診斷流程沒法得到報銷，而這個診斷恰恰能幫你找到最對的治療方案，那就沒人願意做，因為沒人願意付錢。

I'm not saying I mean, everyone in the system, where a lot of people are really well meaning and they want to do the right thing by patients, but as long as you can't reimburse a diagnostic procedure that's going to best identify the right therapeutic intervention, no one's going to do it because no one's going to pay for it.

所以說，真正更有可能率先落地這些數據驅動方法的，是那些醫療服務與報銷機制能對齊的國家——那裡的系統獎勵的，正好是對病人最有利的東西。這話聽起來真讓人難過。

And so I think that actually what's more likely that these data driven approaches will first be implemented in countries where there is an agreement of incentives between delivery of care and payment for that delivery, where there is a better kind of alignment of what is best for patients and what is best for the system, which is a sad statement.

聽你這麼一說，我簡直要崩潰了。

I mean, that makes me so depressed.

我就在想，要是我們真能把這些數據都抓下來，像你說的，這些數據能帶我們衝上多少新突破啊，

I'm just like, imagine if we could capture all of this data, like you were saying, this data can springboard us to so many advancements and

現在沒人願意為這個付錢。

No one will pay for it right now.

行吧，這事兒咱們得幹起來。

All right, we got to work on that.

你提到了这么多让人兴奋的事，比如过去十五年乳腺X光检查的进步，哪怕就这一项也好。

So you've named so many things that you're excited about, whether it's even just, you know, the improvement in the last fifteen years in mammography.

有没有哪些进展是你特别期待的，但公众却没怎么注意到？

Are there certain advancements that you are sort of most excited about that might not have gotten enough public attention?

未来几年，你觉得我们会看到哪些让你特别激动的变化？

Or like, what are we going to be seeing that you're so excited about that is going to come over the next few years?

我觉得，从人体身上收集越来越丰富、越来越量化的数据，这件事正变得越来越普遍。

So I think that the collection of increasingly higher content and more quantitative data from human patients, which is becoming increasingly prevalent.

把这些数据和疾病的底层分子机制对上号——不管是基因层面，还是其他各种组学技术——就能帮我们真正理清疾病的因果路径。

And the alignment of that with underlying molecular causes of disease, whether it be genetics on the one hand or different types of omic modalities on the other are going to allow us to really map those causal pathways between this is the pathogenesis of a disease.

从基因开始，经过体内分子层面的变化，最后演变成整个机体层面的症状，这条线我们能连起来了。

As it starts from the genetics, goes through molecular changes in the body, and ends up in these much more organism level changes.

我觉得最根本的一点是：你搞不懂的东西，就没法治。

And you can't fix what you don't understand, I think, is really the fundamental statement here.

而你测不到的东西，就根本搞不懂。

And you can't understand what you don't measure.

所以一切从测量开始。

So it starts from measuring.

然后才是理解。

It moves through understanding.

有了理解，才能去解决。

And then that's going to give rise to fixing.

我觉得未来十五到二十年，我们会看到一场根本性的转变，诞生出一批真正经过精心设计的疗法，而不是靠运气偶然有效的那种。

And I think that is the transformation that we're going to see in the next fifteen to twenty years that's going to give us a completely new family of therapeutic interventions that are effectively designed as opposed to by happenstance happen to be efficacious.

这些年和你聊天，我一直很享受，我知道你是个系统思维者，思考特别严谨。

So one of the things that I've had the delight of talking to you over a number of years is I know you're a systems thinker, obviously a rigorous thinker.

那你觉得，我们该怎么调整经济激励机制，让大家愿意共享真正有价值的数据，推动整个领域往前走？

So what do you think we need to do to change the economic incentives for this collective data collection to be sharing the on target stuff that will allow the whole field to advance?

你肯定也想过，要是能打个响指，或者某个条件实现了，那该多好。

You know, I'm sure you've thought about, well, look, if I could wave a wand or if X would happen, this would be great.

我问的其实就是这个意思。

And so that's essentially the question I'm gesturing at.

我要是真有魔杖，能做的事多了，但很多都不现实。

There's many things I would do with the wand, which I think are not feasible.

所以我把它缩小到一个更可行的方向——我觉得在美国，其实已经有一些地方，各方利益刚好对得上，完全有可能通过收集数据来真正改善病人的治疗效果。

So I'm going to scope it down to something that I think is more feasible, which is I think even within The US, there are pockets of alignments of incentives where one could, I think, demonstrate the value of collecting data so as to inform patient care.

如果真能做到这一点，用不了多久——以人类健康周期来看，这已经算很快了——我们就能证明：从现有的治疗方案里，精准选出最适合病人的，真的能带来更好的康复结果。

And I think by doing that, you could potentially actually demonstrate in a relatively short time frame, relatively short being short in what is the life cycle of human health, which is the ability, for example, to appropriately select the right therapeutic intervention from among the ones that exist towards a better patient outcome.

只要利益链能打通，让这种做法能缩短病人找对药的漫长过程——本来动不动就是好几年，现在能快很多，病人更快康复，后续医疗需求也少了，那用不了几年，我们就能看到实实在在的效果。

And if you have enough alignment of incentives so that you could demonstrate that by doing so, you abbreviate, you cut short what is often a multi year patient journey to find the right drug, and they become healthier, and as such, need less health care down the line, you could demonstrate that in a matter of, I think, a few short years.

一旦证明了这一点，大家就会更有动力去收集更多这类病人的数据，而这些数据又能反过来推动更长远的探索，最终催生出更高效的治疗方案。

And that will create the right momentum around increased collection of data around these patients that would then drive the longer term ability to interrogate those data and come up with an even better set of therapeutic interventions.

这就从一个能快速见效的小突破，慢慢演变成推动整个研发链条创新的引擎。

So you go from something that creates immediate, relatively short term value to something that then drives innovation in the subsequent stages of development.

这，就是我的期待。

That would be my hope.

这其实也正好是很多人一提到AI就立刻想到隐私问题的地方。

It seems like this is also one of those places where a lot of people talk about privacy, especially with AI.

这正是我们需要为公共利益做点什么的地方，虽然有些人可能不太乐意。

And it's one of the places where we need to do something for the common good that some people might not be comfortable with.

你对这个有什么看法吗？

I don't know if you have thoughts on that.

我当然有想法。

I absolutely do have thoughts on that.

这事儿我琢磨很久了。

It's something that I've thought about a lot.

如果你去问那些真正经历过重病的患者，他们愿不愿意在严格匿名和保护的前提下，把数据用于科研，帮助研发更好的疗法，你会发现绝大多数人都会说：当然愿意，我很乐意帮忙。

I think that if you ask anyone who's actually been a patient of a grievous disease, whether they are interested in sharing their data in an appropriately anonymized and protected way in a research environment towards the common good of finding better therapeutics, you'll find that you will have a very high success rate in terms of people saying, of course, I'd love to be helpful.

当然，得确保匿名机制和其他保护措施都到位。

And yes, you need to make sure that it's done with the appropriate parameters around anonymity and so on and so forth.

但大多数人其实都愿意出一份力。

But most people would like to be helpful.

可现在的问题是，根本没人能问他们这个问题，也没人给他们机会来回应。

Right now, there is not even a way for someone to be asked that question or to provide a response to that question.

你想想看，现在去车管所，你还能在表格上勾选愿意当器官捐献者。

Think about this by contrast the fact that when you go into the DMV today, you can mark on a little form that you'd like to be an organ donor.

每个人都有这个机会。

And everyone has that opportunity.

这可比捐数据重要多了。

That is a much bigger decision than donating your data.

你捐的是自己的器官啊。

You're actually donating your organs.

我能勾选捐器官，却没地方能表明我愿意把数据用于医学研究和临床进步。

I have the ability to mark that, but I have no place to mark the facts that I'd like my data to be used for the advancement of science and clinical care.

这到底是为什么？

Why is that?

为什么就不能有个统一的数据库，把自愿同意的患者数据收集起来，好好整理、标准化，再在严格保护的前提下，提供给科研界，用来推动医学新发现和新疗法？

Why is there not a centralized repository where data from patients who consent can be collected, appropriately curated and harmonized, and provided to the scientific community with appropriate safeguards for the development of new biological insights and new therapeutic interventions.

你刚才问到那个‘魔法棒’的事。

You asked me about the magic wandry.

其实，那根魔法棒就摆在那儿。

There is a magic wand right there.

那根魔法棒就是有人肯投一笔不算多的钱，建一个集中式的医疗数据仓库——顺便说一句，其他国家早就有了。

The magic wand would be someone investing what is not a very large amount in a centralized health care data repository, which by the way other countries have.

也不是所有国家，但很多国家都有。

Not all of them, but a lot of them do.

可我们美国呢？连个影子都没有。

And we have nothing remotely like that in The US.

这真是太荒唐了。

That's really a travesty.

我完全同意，一千个赞成。

I a thousand percent agree.

所以，达芙妮，每期节目我们都会把AI请进来，问了GPT-4.0一个问题：到2034年，人工智能驱动的药物研发和治疗会是什么样？

So, Daphne, in every episode, we invite AI into the conversation, and we asked GPT-4.0 about machine learning empowered drug discovery and treatment in 2034.

然后我们又让GPT-4O用一个发生在辛辛那提的科研人员的虚构故事，把这些变化画出来。

And then we asked, GPT-4O to illustrate these transformations via a fictional story of a researcher in Cincinnati.

你读过这个故事了，内容在节目笔记里，想跟着看的听众都可以找到。

And you've read this story, and it's in the show notes for listeners who want to follow along.

所以啊，这故事不是我写的，你不喜欢也好，喜欢也罢，都跟我没关系。

And so, again, I did not create this, so you can hate it or love it.

这事儿跟我可没关系。

It won't reflect on me.

你对这个故事有什么看法？

What were your reflections on the story?

要是你来改，你会改哪儿？

What would you change?

比如，你觉得哪些地方特别有希望？

Like, what seemed promising?

我觉得这故事其实挺不错的。

So I think the story is actually not bad at all.

它抓住了好些关键点，比如能针对特定疾病亚型找药，像遗传性癫痫这种亚型。

I think it captures some important components, including the ability to identify drugs for specific subtypes of disease, like in the subtype of genetic epilepsy.

我觉得这个点挺吸引人的。

I thought that was compelling.

我很喜欢他们能主动设计出新疗法，尤其是在对抗抗生素耐药性这块。

I liked the ability to also generate new therapeutics by design, if you will, in the antibiotic resistance piece.

我觉得缺了一环，就是这两部分之间没什么联系——其实这又回到了如何基于对疾病机制的精细分析，来提出新的治疗假说。

Where I think there was a bit of a gap was in the lack of connectivity between those two, which is I think coming back to the identification of novel therapeutic hypotheses based on an understanding of this sort of fine parsing of disease mechanisms.

我的意思是，先搞清楚某种跟疾病相关的生物过程。

So I think sort of saying here is an understanding of some biological process that underlies the disease.

然后我们再反过来，设计一种故事里没提到的药，来精准解决这个问题。

And now we're going to close the loop and design a drug towards that goal that wasn't present in the story.

我觉得真正的机会就在这儿，毕竟虽然我很看好老药新用，但新病往往很难靠复用旧药就完美解决。

And I think that's where most of the opportunities are going to lie simply because much as I love drug repurposing, it is rare that a repurposed drug is really the ideal solution to a new disease.

所以，当你对疾病的底层生物学理解更深了，很多时候还是得开发新药。

And so I think you oftentimes do need to make new medicines as you have a better understanding of the underlying biology.

这么说吧，AI这波表现其实挺不错的。

So that being said, it was a pretty I mean, the AI did a pretty good job.

太棒了。

Awesome.

谢谢。

Thank you.

其中一个

One of the

我特别喜欢做播客的一点，哪怕像现在这样和你达芙妮这样的熟人聊，我也会在准备过程中学到一些以前不知道的东西。

fun things that I love about doing podcasts, even with people I know like you, Daphne, is that I actually learned some things in the prep that I didn't know before.

我知道你做过一系列创业，做了很多对社会和人类有重大影响的事。

Like, I had known that you've done serial entrepreneurship and various things that are, you know, kind of impactful at scale to society and humanity.

但我之前不知道的是，你的转变可能始于和博士后导师拉塞尔的一次关键对话——从纯理论研究转向去打造真正能大规模落地的东西。

What I didn't know is it might have started with a kind of pivotal conversation with your postdoc advisor, Russell, shifting from conceptual work to building the things that matter at scale in the world.

如果真是这样，能跟我们说说那次对话吗？

If that's true, tell us about that conversation.

是的。

Yeah.