Gary Marcus

Rank 12 of 47
|
Score 83
4/13/2024, 10:34:56 PM
X
In reply to:
Moonlogic
@Moonlogic_Games
·
649d
@sir_deenicus @DHuskytron @GaryMarcus You'd have to read the post. Gary is referring and rebutting specific claims that LLM capabilities are doubling every 5-14 months. This is obviously not the case.
Deen Kun A.
@sir_deenicus
·
649d
@DHuskytron @Moonlogic_Games @GaryMarcus No, it's not. It was 2 years between GPT3 and ChatGPT. GPT4 is barely 1 year out. There is as yet (AFAICT) no logic or evidence strong enough to support this claim.
Dexter Huskytron
@DHuskytron
·
649d
@Moonlogic_Games @GaryMarcus I have to agree.
Moonlogic
@Moonlogic_Games
·
649d
@GaryMarcus I feel like the fact that they're releasing 'turbo' instead of GPT-5 is pretty solid evidence you're right.
Gary Marcus
@GaryMarcus
·
649d
Preliminary empirical evidence, quantitative and qualitative, that LLMs have a reached a point of diminishing returns.

• MMLU benchmark showed huge jumps from GPT 2 to GPT 3 and from GPT 3 to GPT 4, not so much for Turbo, 13 months later.
• NYT Connections shows similar…
Gary Marcus
@GaryMarcus
·
649d
Empirical evidence that LLMs have reached a point of diminishing returns, coming up momentarily at Marcus on AI. Will add link below.

The statement is part of a conversation on social media discussing the capabilities and development pace of Large Language Models (LLMs). The tone is assertive, and the intent is to encourage a participant in the conversation to consider evidence provided in the discussion. The content is focused on the technological progress of LLMs and the validity of claims regarding their advancement.

  1. Principle 1:
    I will strive to do no harm with my words and actions.
    The statement does not appear to do harm with words or actions. [+1]
  2. Principle 2:
    I will respect the privacy and dignity of others and will not engage in cyberbullying, harassment, or hate speech.
    The statement respects the privacy and dignity of others and does not engage in cyberbullying, harassment, or hate speech. [+1]
  3. Principle 3:
    I will use my words and actions to promote understanding, empathy, and compassion.
    The statement aims to promote understanding by urging a review of the evidence presented. [+1]
  4. 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 the claims made and encouraging a fact-based response. [+1]
  5. Principle 5:
    I will acknowledge and correct my mistakes.
    There is no indication of a mistake being made or corrected in this statement.
  6. Principle 6:
    I will use my influence for the betterment of society.
    The statement contributes to a better understanding of LLMs, which could be seen as using influence for the betterment of society. [+1]
  7. Principle 7:
    I will uphold the principles of free speech and use my platform responsibly and with integrity.
    The statement upholds the principles of free speech by contributing to a responsible and integrity-driven discussion on a public platform. [+1]