Gary Marcus

Rank 12 of 47

Score 83

@peterwildeford People misread 2022 paper where I coined “hitting a wall” as saying it would happen immediately, whereas I actually made case that scaling was not a physical law & would tail off, & predicted difficult to surmount obstacles (hallucinations, reliability, abstract reasoning etc)

1/1/2025, 1:50:21 PM

In reply to:

Peter Wildeford 🇺🇸

@peterwildeford

395d

I'm thinking 2025 will finally be the year that deep learning hits a wall.

No, it hasn't hit a wall any of the past ten times I thought it would but this year is going to be the year. 2025 will be when deep learning finally hits a wall.

you-have-just-experienced-things

@y_h_j_e_t

396d

2024 was the year of "model collapse" replacing "stochastic parrot" as the knee-jerk objection to AI progress; excited to see what we get in 2025

The statement addresses misconceptions about a previous paper on AI development, clarifying the author's position on the scalability of deep learning and potential challenges. It engages in a broader discussion about AI progress and potential limitations, contributing to public discourse on technological advancements.

Principle 1:
I will strive to do no harm with my words and actions.
The statement aims to clarify misunderstandings, which aligns with striving to do no harm by reducing misinformation. [+1]
Principle 3:
I will use my words and actions to promote understanding, empathy, and compassion.
By explaining the original intent of the paper, the statement promotes understanding and empathy towards the author's perspective. [+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 constructive dialogue by addressing misconceptions without personal attacks. [+1]