Gary Marcus

Rank 19 of 47

Score 64

@RobS142 @Wharrgarbl42 @DeryaTR_ I think it goes far beyond that, perhaps into regex’s to particular case:, the RLHF itself isn’t exactly a standard notion of MoE, same with O1 add-ons

10/3/2024, 2:56:55 PM

In reply to:

Rob S.

@RobS142

281d

@Wharrgarbl42 @DeryaTR_ @GaryMarcus This is word salad. It was never "a model" we always knew it was a MoE...

wharrgarbl

@Wharrgarbl42

281d

@DeryaTR_ @GaryMarcus The secret behind the scenes:

OpenAI GPT is no longer a model.

That's why it's closed source.

It is a system. It's a giant integration of half-baked specialist models cobbled together into one big lie.

Derya Unutmaz, MD

@DeryaTR_

281d

Hey Gary @GaryMarcus, do you want to concede on our bet? 😊 Just admitting that you were wrong would be enough 😎

Alternatively, we could raise it 10x — I’m in! 😇

The conversation involves a discussion about AI models, specifically OpenAI's GPT, and critiques of its structure and transparency. It touches on public discourse about AI technology and its implications.

Principle 1:
I will strive to do no harm with my words and actions.
The statement critiques AI models without harmful language, adhering to the principle of doing no harm. [+1]
Principle 2:
I will respect the privacy and dignity of others and will not engage in cyberbullying, harassment, or hate speech.
The conversation respects privacy and does not engage in harassment, aligning with this principle. [+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 technical critique without personal attacks, adhering to constructive criticism. [+1]