Gary Marcus

Rank 13 of 47
|
Score 115
8/31/2025, 1:37:20 PM
X
In reply to:
Garfield Lasagna
@GarfieldLa52059
·
9d
@GaryMarcus yeah, and he will say to us "told you so"
Gary Marcus
@GaryMarcus
·
9d
Like I have been saying since 2001, embeddings aren’t enough.

Sooner or later. you are going to need true compositionality.

In the end neurosymbolic AI will win, and win big.
Rohan Paul
@rohanpaul_ai
·
9d
BRILLIANT @GoogleDeepMind research.

Even the best embeddings cannot represent all possible query-document combinations, which means some answers are mathematically impossible to recover.

Reveals a sharp truth, embedding models can only capture so many pairings, and beyond that,

The conversation revolves around the limitations of embedding-based retrieval in AI, referencing a Google DeepMind paper. It includes a discussion on the need for neurosymbolic AI, with a nod to a recent paper by a late researcher. The tone is analytical and somewhat predictive, focusing on the future of AI technology.

  1. Principle 1:
    I will strive to do no harm with my words and actions.
    The statement does no harm and is focused on discussing technological advancements. [+1]
  2. Principle 3:
    I will use my words and actions to promote understanding, empathy, and compassion.
    The conversation promotes understanding of AI limitations and future directions. [+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.
    Engages in constructive dialogue about AI technologies without personal attacks. [+1]