Gary Marcus

Rank 10 of 47

Score 110

@littmath @DrJimFan I wouldn’t conclude that just yet. I await clarification on the pretraining and augmention, as well tests from people like you and @colin_fraser on problems slightly structured differently.

Of course we also don’t know what is inside the system.

12/21/2024, 7:14:35 PM

In reply to:

Daniel Litt

@littmath

212d

@GaryMarcus @DrJimFan Hold on, that's not what I'm saying--they're still very hard questions! Just not so hard for people who really know this stuff. Suggests the AI does in some sense really know this stuff.

The statement is part of a discussion about AI capabilities, specifically addressing the complexity of questions and the understanding of AI systems. It seeks clarification and further testing, indicating a thoughtful engagement with the topic.

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 dialogue by seeking clarification and further testing, avoiding personal attacks or ad hominem arguments. [+1]