# Gary Marcus 联合多位医学专家反驳 Dario Amodei 关于 AI 五年内治愈大多数疾病的说法，指出其忽视临床验证和疾病复杂性

- 来源：Gary Marcus
- 发布时间：2026-08-17 22:56
- AIWatch 分数：53
- AIWatch 标记：未精选
- AIWatch 链接：https://aiwatch.icu/events/evt_01m0845bmf0fbjqrymaw1v76q6
- 原文链接：https://garymarcus.substack.com/p/no-dario-amodei-we-will-not-be-curing

## 精选理由

常规快讯，保留列表

## AI 摘要

Gary Marcus 联合多位医学专家反驳 Dario Amodei 关于 AI 五年内治愈大多数疾病的说法，指出其忽视临床验证和疾病复杂性。

## 正文

On Saturday, Dario Amodei claimed (in a long defense of his company’s approach to AI) that he thinks it “will actually be possible to cure most human disease in ~5-10 years”, highlighting “curing cancer” among other goals. As usual, some in the media ate it up.

And I wish it were true. But his timeline is so naive as to be absurd, both with respect to how medical science works and why it is challenging.

If you haven’t already done so, I strongly recommend you read (or reread) the March essay I wrote about AI and cancer, which builds on an essay by Emilia Javorsky. That essay started with Javorsky’s observation that (as of March) not one drug designed by AI had yet reached clinical adoption, even after over a decade of efforts.

Part of the crux of my essay (and hers) was that even if AI were a lot better, curing disease is hard, both economically and technically, and you also can’t simply skip the time-consuming clinical work. (I further argued that we don’t really have the right kind of causally- and biologically-sophisticated AI yet. New advances in AI and math don’t fundamentally change what I wrote there.)

The happy news is that you don’t have to take my word for any of this anymore. The mood has changed, and this time a whole pile of people pushed back. I have yet to see Amodei’s ridiculous claim dissected in the mainstream media, but below are some of the many tweets over the weekend that rightly took Amodei to task.

Let’s start with Eric Topol, a well-respected and very pro-AI physician/researcher who is having none of it. (Topol also notes that Hassabis made a similar absurd argument last year on 60 Minutes, maybe the most naive things I have ever heard Hassabis say. AI might someday contribute as Hassabis says, but certainly not in the next decade. Many of the quotes below happened to be directed to Amodei, who was in the news, but apply to what Hassabis claimed as well.)

X avatar for @EricTopol

Eric Topol@EricTopol

I'm keen on AI having a transformational impact on human health. But these cure projections are wildly off-base and impossible.

4:53 PM · Aug 16, 2026 · 72.8K Views

70 Replies · 58 Reposts · 415 Likes

By way of setting realistic expectations, Topol adds:

X avatar for @EricTopol

Eric Topol@EricTopol

Here's the current list of all non-infectious diseases that are universally curable by a medication

9:51 PM · Aug 16, 2026 · 8.53K Views

6 Replies · 6 Reposts · 57 Likes

Caltech’s Bren Professor of Computational Learning, Lior Pachter, wrote

X avatar for @lpachter

Lior Pachter@lpachter

We don’t even know what all the human diseases are.

1:34 AM · Aug 17, 2026 · 1.72K Views

6 Replies · 3 Reposts · 41 Likes

Computational biologist Keith Robison:

X avatar for @OmicsOmicsBlog

Keith Robison@OmicsOmicsBlog

Sorry, but anyone claiming AI is curing most diseases in 5-10 years is spewing magical thinking & it’s very cruel for patients to be misled this way No matter how clever your AI is, proving we can trust it is clever takes preclinical & clinical work, which eats that timeline

X avatar for @DarioAmodei

Dario Amodei @DarioAmodei

2/2 Second, on the messaging around AI.  I do not agree that my messaging has been disproportionately negative.  In fact it has been about equally balanced between risks and benefits: I’ve written one major essay about each, and even in interviews where I discuss the risks, I

12:36 AM · Aug 17, 2026 · 1.32K Views

4 Replies · 2 Reposts · 30 Likes

Cancer biologist John Herrman wrote (plausibly if optimistically, but nowhere near as optimistically as Amodei):

X avatar for @ablT315I

John Herrmann, PhD@ablT315I

@EricTopol @GaryMarcus Much more defensible: AI could plausibly accelerate a few medical breakthroughs over next 10 years that surpass today’s best advances. That’s far different from “AI will cure most human disease” including complex dynamically evolving disease states like many malignancies etc

2:48 AM · Aug 17, 2026 · 31 Views

2 Likes

MD/PhD Marios Georgakis asks:

(As far as I know, the correct answer to the question about real plans posed in the first sentence is a simple “no”.)

The software engineer/enterpreneur Chombe Bupe, responding to Georgakis, adds:

X avatar for @ChombaBupe

Chomba Bupe@ChombaBupe

The lack of causal understanding in today's artificial intelligence (AI) systems means the models would generate a ton of proposals for potential cures for diseases most of which would be wrong & it would take a lot of effort to wade through all of that slop, I mean proposals.

X avatar for @MariosGeorgakis

Marios Georgakis @MariosGeorgakis

Has any of the AI companies ever presented a real plan for how they are going to cure even one disease in 5-10 years? This doesn't just sound crazy to ordinary people, but also to any clinician or drug developer I know. Beyond the fact that we still have very few true cures for

2:04 AM · Aug 17, 2026 · 699 Views

4 Replies · 7 Reposts · 29 Likes

Even the ultimate AI optimist, Wharton professor Ethan Mollick, wasn’t really buying:

X avatar for @emollick

Ethan Mollick@emollick

I, also, would like more precision around discussions of how AI will cure diseases. I assume it means "a nation of geniuses in a data center will invent things that should in theory cure diseases" but there is a real big gap between that and actual trials & proof & FDA approval.

X avatar for @MariosGeorgakis

Marios Georgakis @MariosGeorgakis

Has any of the AI companies ever presented a real plan for how they are going to cure even one disease in 5-10 years? This doesn't just sound crazy to ordinary people, but also to any clinician or drug developer I know. Beyond the fact that we still have very few true cures for

1:04 AM · Aug 17, 2026 · 13.9K Views

20 Replies · 4 Reposts · 107 Likes

Penultimate words to Dr. Catharine Young, a Senior Fellow at the Harvard Chan School of Public Health, in a brutal pair of tweets further underscoring Amodei’s naïveté with respect to medicine.

X avatar for @DrCatharineY

Dr. Catharine Young@DrCatharineY

PSA: Cancer is not one disease. It is more than 200 distinct diseases, all with different causes, biology and mechanisms. Scientists who study cancer rarely talk about “curing cancer” because it’s nothing more than an empty slogan.

X avatar for @PolymarketMoney

Polymarket Money @PolymarketMoney

BREAKING: Anthropic’s CEO declares AI companies can win back public trust by “actually curing cancer.”

2:24 AM · Aug 17, 2026 · 3.7K Views

10 Replies · 18 Reposts · 87 Likes

and

X avatar for @DrCatharineY

Dr. Catharine Young@DrCatharineY

Somebody needs to do an internship at the FDA.

X avatar for @Polymarket

Polymarket @Polymarket

JUST IN: Anthropic CEO Dario Amodei predicts AI could make it possible to “cure most human disease” within 5-10 years.

10:28 PM · Aug 16, 2026 · 9.05K Views

19 Replies · 15 Reposts · 178 Likes

§

There are a great many things about Amodei that I respect, including both his technical chops and business sense, as well as his bravery in standing up to the US government on important matters. But his habit of giving people unrealistic hope is not one of them. (See also the time he claimed that AI could double human lifespan by the end of the decade.)

In fairness, Amodei may believe what he says, but in that case he ought take Dr. Young’s suggestion of doing a medical internship, in order to better calibrate his understanding of the challenges in medicine.

Share Marcus on AI

No posts