Gary Marcus

Rank 19 of 47

Score 64

@EricTopol @HarshaNori @chrisck @Dominic1King @mustafasuleyman @erichorvitz Unless I have misunderstood, @EricTopol, this was a completely unfair test.

In particular, the test seems to have been open book for machines (training in the web UpToDate etc) but not humans:

Dr. Dominic Ng

@DrDominicNg

12d

3. The physician comparison was rigged
Docs were banned from:
❌ Googling symptoms
❌ Consulting colleagues
❌ Using UpToDate/medical databases
❌ Calling specialists

That's not how we practice!!
It's like testing a chef who can't use recipes or taste their food.

7/1/2025, 10:08:58 PM

In reply to:

Eric Topol

@EricTopol

13d

Agentic A.I. vs experienced physicians for diagnosis of > 300 complex diagnostic cases:
4-fold higher accuracy and 20% lower cost https://arxiv.org/abs/2506.22405
@HarshaNori @chrisck @Dominic1King @mustafasuleyman @erichorvitz

The statement critiques the fairness of a test comparing AI and human physicians, suggesting that the conditions were biased in favor of AI. It raises concerns about the methodology, implying that AI had access to resources that human doctors did not, which could skew the results.

Principle 1:
I will strive to do no harm with my words and actions.
The statement aims to highlight potential harm in the form of unfair testing practices, which could mislead public perception about AI's capabilities. [+1]
Principle 2:
I will respect the privacy and dignity of others and will not engage in cyberbullying, harassment, or hate speech.
It respects the dignity of human physicians by pointing out the unrealistic constraints placed on them during the test. [+1]
Principle 4:
I will engage in constructive criticism and dialogue with those in disagreement and will not engage in personal attacks or ad hominem arguments.
The statement engages in constructive criticism by questioning the test's fairness without resorting to personal attacks. [+1]