Gary Marcus

Rank 16 of 47
|
Score 70
7/20/2025, 11:29:18 PM
X
In reply to:
Brandon
@PixelCanuck
·
63d
@GaryMarcus Based on past behaviour of the companies like OpenAI who are pushing AI hype, it is reasonable to assume they are not being fully honest about the model that won

That's my guess why it's an "experimental" model

It has been fine tuned specifically to try and win
Gary Marcus
@GaryMarcus
·
63d
@PixelCanuck we don’t know
Brandon
@PixelCanuck
·
63d
@GaryMarcus It's a pretty impressive benchmark but let's be real. It's not a general model that won, it was a specially tuned model specifically designed to perform on this test

Is it still cool and impressive? Yeah, for sure. But it's still basically a marketing stunt
Gary Marcus
@GaryMarcus
·
63d
You think AGI is coming this year?

GPT-5 (which will) didn’t win IMO Gold.

A different “experimental” model did.

That model probably won’t come out this year.

And it didn’t even win IMO (25+ high school students did better; most professional mathematicians probably could if

The statement is part of a public discourse on the topic of AI models and their performance, specifically in relation to claims made by companies like OpenAI. The conversation involves skepticism about the honesty and transparency of AI companies and the nature of AI models being promoted. The tone is speculative and analytical, aiming to provide a balanced view on the capabilities of AI models.

  1. Principle 1:
    I will strive to do no harm with my words and actions.
    The statement does not cause harm and maintains a neutral tone, focusing on the performance of AI models rather than attacking individuals or companies. [+1]
  2. Principle 3:
    I will use my words and actions to promote understanding, empathy, and compassion.
    The statement promotes understanding by acknowledging that AI models have strengths and weaknesses, which can foster a more nuanced view of AI capabilities. [+1]
  3. 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 providing a reasoned guess about the performance of AI models without resorting to personal attacks. [+1]