Gary Marcus

Rank 19 of 47
|
Score 64
6/19/2024, 1:39:20 PM
X
In reply to:
Noah Giansiracusa
@ProfNoahGian
·
380d
Absent this, I just don't buy these arguments by Hinton and various ppl in AI safety that ASI is automatically a species smarter than humans. Again, that would suggest current AI is a species dumber than humans, and to me it's manifestly not. It's a tool, not a species. 7/7
Noah Giansiracusa
@ProfNoahGian
·
380d
Maybe evolution is key? Genetic algorithms are fascinating but so far play almost no role in AI. I'd love to see more discussions of AIs from eg a @RichardDawkins selfish gene perspective, or anything trying to measure how far AI has moved from tool to genetic replicator. 6/7
Noah Giansiracusa
@ProfNoahGian
·
380d
I don't see eg GPT2 as insects, GPT3 as rodents, GPT4 as primates, I see all three as equally non-species like, just more and more capable tools. Not saying AIs won't move in the species-like direction, just saying not much evidence yet or discussions of what that means. 5/7
Noah Giansiracusa
@ProfNoahGian
·
380d
This seems like a huge gap in that line of reasoning. Maybe AIs will become more agentic and in doing so more species-like (maybe...), but that's a separate argument. How can one look at benchmark performance of GPT2 vs GPT3 vs GPT4 and conclude this? 4/7
Noah Giansiracusa
@ProfNoahGian
·
380d
I can't fathom that anyone sees ChatGPT or any other current AIs that way. They very much seem tools not species. We can certainly extrapolate a growth in the capabilities of these tools, but I don't see ppl arguing with data that AIs are getting more species like. 3/7
Noah Giansiracusa
@ProfNoahGian
·
380d
The argument is basically that AI systems keep getting more "intelligent" so once they exceed humans we'll be in this smarter-than-us species situation. BUT doesn't that imply that current AI should be seen as a less-intelligent-than-us species? 2/7

The statement engages in a substantive discussion about the nature of artificial intelligence (AI) and its potential risks, particularly in relation to the arguments made by Geoffrey Hinton and others in the AI safety community. The tone is analytical and critical, aiming to dissect and challenge the ideas presented by ProfNoahGian. The intent is to contribute to the ongoing debate about AI's capabilities and risks, making it a part of public discourse.

  1. Principle 1:
    I will strive to do no harm with my words and actions.
    The statement does not appear to cause harm with words or actions. It focuses on critiquing ideas rather than individuals, adhering to this principle. [+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, as it does not engage in cyberbullying, harassment, or hate speech. It maintains a professional tone. [+1]
  3. Principle 3:
    I will use my words and actions to promote understanding, empathy, and compassion.
    The statement aims to promote understanding by addressing the complexities of AI and its potential risks. It encourages a deeper examination of the issues, aligning with this principle. [+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 and dialogue. It critiques the ideas presented by ProfNoahGian without resorting to personal attacks or ad hominem arguments. [+1]
  5. 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 and uses the platform responsibly. It contributes meaningfully to the public discourse on AI and its implications. [+1]