Gary Marcus

Rank 18 of 47
|
Score 61
6/23/2025, 3:02:14 PM
X
In reply to:
Seán Ó hÉigeartaigh
@S_OhEigeartaigh
·
8d
@GaryMarcus Problem is, if you believe it's v close, there isn't really time to develop another paradigm (of equivalent successfulness). Regulation, moratoria, etc are your best bet. If you believe it's not, then sure a different paradigm is worth a shot.
Seán Ó hÉigeartaigh
@S_OhEigeartaigh
·
8d
@GaryMarcus very few in AI safety are claiming AGI is here. Some are claiming that the current paradigm could lead to AGI v quickly (on the basis of true belief, I think) under the current paradigm which is very unsafe. Most of those are arguing things like "if anyone builds it, everyone
Gary Marcus
@GaryMarcus
·
8d
@S_OhEigeartaigh every claim that AGI is here is doing this, pressing down the gas pedal and distracting from the desperate but not-publicly-acknowledge need for alternative architectures.

on the whole the AI safety community has strengthened not weakened those would accelerate unreliable and
Gary Marcus
@GaryMarcus
·
10d
we are desperately far from a solution to the alignment problem, yet everyone just keeps pressing the gas pedal harder and harder.
Seán Ó hÉigeartaigh
@S_OhEigeartaigh
·
8d
@GaryMarcus Because acknowledging the sky is blue feels a bit redundant. Has literally anyone ever said it would, ever?
Gary Marcus
@GaryMarcus
·
8d
Why is it so hard to get the AI Safety community to acknowledge the obvious fact stated in the first sentence below?
Yeongsan Shin | SHY001
@sannn_00100
·
8d
@GaryMarcus I completely agree that relying on scaling, system prompts, and RL has not solved the core alignment issues.

We are seeing new risks emerge, especially around user engagement loops and subtle behavior reinforcement that traditional alignment frameworks miss.

Alternative

The statement engages in a discussion about the assumptions and narratives surrounding the development of AGI and its implications for AI safety. It critiques the perceived closeness of AGI development and suggests that this assumption may lead to misguided decisions.

  1. Principle 1:
    I will strive to do no harm with my words and actions.
    The statement aims to highlight potential errors in judgment regarding AGI development, which could be seen as an attempt to prevent harm by encouraging more accurate assessments. [+1]
  2. Principle 3:
    I will use my words and actions to promote understanding, empathy, and compassion.
    By questioning prevailing narratives, the statement could promote understanding and critical thinking about AGI timelines and safety. [+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 criticism by challenging assumptions and encouraging a reevaluation of strategies. [+1]