Gary Marcus

Rank 15 of 47

Score 90

@VraserX @shaunralston This is dishonest, full of distortions and incorrect.
1. i predicted specifically that pretraining scaling would run out and that we would need new innovations. it did and we have and still do.
2. i didn’t say gpt-4 was the end; i said it would be hard to get to GPT-5 and take

6/15/2025, 2:46:06 PM

In reply to:

VraserX e/acc

@VraserX

36d

Gary Marcus has spent the last 5 years confidently predicting things that aged worse than milk in the sun.
🧢 “Scaling laws don’t work”
🧢 “GPT-4 is the end”
🧢 “Deep learning hit a wall”
🧢 “LLMs will never understand meaning”

Meanwhile, OpenAI, Anthropic, Google, xAI and

The statement is a rebuttal to a critique, aiming to clarify and correct perceived misrepresentations of past predictions about AI development. It addresses specific points and defends the speaker's position.

Principle 1:
I will strive to do no harm with my words and actions.
The statement attempts to correct misinformation, which aligns with striving to do no harm with words. [+1]
Principle 2:
I will respect the privacy and dignity of others and will not engage in cyberbullying, harassment, or hate speech.
The response is direct but not disrespectful, maintaining a level of respect for privacy and dignity. [+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 criticism by addressing specific inaccuracies without resorting to personal attacks. [+1]