Gary Marcus

Rank 14 of 47
|
Score 118
4/23/2024, 5:51:42 PM
X
In reply to:
James Miller
@JimDMiller
·
263d
@TrueAIHound @GaryMarcus I meant to imply that if you define intelligence such that LLMs lack it, then you have defined intelligence such that humans are also not intelligent.
AGIHound
@TrueAIHound
·
263d
@JimDMiller @GaryMarcus I argue that only humans do. This being said, parsing your statement is kind of tricky.
James Miller
@JimDMiller
·
263d
@TrueAIHound @GaryMarcus My view is that either LLMs have an intelligent understanding of the world or human's don't.
AGIHound
@TrueAIHound
·
263d
@JimDMiller @GaryMarcus What if OpenAI has advanced stuff?

They don't. If they did, they would abandon LLMs since LLMs are not intelligent. The intelligence in LLMs was stolen from millions of human beings, the actual intelligence behind the texts. LLMs just compute the stats of the word embeddings.
James Miller
@JimDMiller
·
263d
@GaryMarcus Unless OpenAI quickly matched Google's performance. What if OpenAI has advanced stuff they are not releasing for safety reasons, but would release if anyone released someone better than it had, and Google knows this.
Gary Marcus
@GaryMarcus
·
263d
@JimDMiller there would be such a huge economic gain to being decisively better, lasting fame, enormous market boost etc, that i find this implausible. eg imagine what would happen to Google’s stock price if they blew GPT-4 out of the water?

The statement is part of a Twitter conversation thread discussing the nature of intelligence in humans and large language models (LLMs), such as those developed by OpenAI. The discourse revolves around the philosophical and technical aspects of artificial intelligence, comparing it to human intelligence, and speculates on the motivations behind the development and release strategies of AI technologies by major companies. This conversation constitutes public discourse as it engages in a substantive debate about the nature of intelligence, a topic of public interest and concern in the context of AI development and its societal implications.

  1. Principle 1:
    I will strive to do no harm with my words and actions.
    The statement does not appear to cause harm but rather contributes to a broader understanding of AI and human intelligence. [+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 privacy and dignity, focusing solely on the topic without targeting any individuals negatively. [+1]
  3. Principle 3:
    I will use my words and actions to promote understanding, empathy, and compassion.
    The statement promotes understanding by sharing insights on a complex topic, aiming to enhance comprehension among participants. [+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 dialogue is constructive, focusing on the exchange of ideas rather than personal attacks. [+1]
  5. Principle 6:
    I will use my influence for the betterment of society.
    The statement uses influence to discuss important aspects of AI, potentially contributing to better informed public opinions and policies. [+1]
  6. Principle 7:
    I will uphold the principles of free speech and use my platform responsibly and with integrity.
    The statement uses the platform responsibly, engaging in a meaningful discussion without spreading misinformation or harmful content. [+1]