Gary Marcus

Rank 14 of 47

Score 83

@justjoshinyou13 @pli_cachete I have tried multiple models and presented data (my own and others).

There is nothing wrong with testing multimodal image generation.

5/18/2025, 5:23:07 AM

In reply to:

Josh You

@justjoshinyou13

245d

@pli_cachete @GaryMarcus testing LLMs on knowledge by asking them to draw pictures is insane.

GPT-4o image gen is probably a lot worse than 2022 LLMs at answering your questions through drawing. according to Gary, 4o therefore experienced a huge regression when OpenAI added image capabilities

Rota

@pli_cachete

245d

@GaryMarcus o3 doesn’t generate images, it makes a call to a different model. With how much you discuss this don’t you think it’s worth spending a few minutes learning about them?

Gary Marcus

@GaryMarcus

245d

You have got to be kidding me.

I first wrote about how (certain kinds) of neural nets had trouble with concepts like a pair and three of kind in … 1998.

Here’s o3, today:

The statement engages in a technical discussion about the capabilities of multimodal models in AI, specifically addressing the testing of image generation. It presents a defense of testing these models, suggesting that there is value in exploring their capabilities.

Principle 1:
I will strive to do no harm with my words and actions.
The statement does not cause harm and contributes to a technical discussion, aligning with the principle of doing no harm. [+1]
Principle 3:
I will use my words and actions to promote understanding, empathy, and compassion.
By presenting data and supporting the exploration of multimodal image generation, the statement promotes understanding and knowledge sharing. [+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 constructively in the dialogue, addressing the topic without personal attacks. [+1]