Gary Marcus

Rank 20 of 47
|
Score 95
4/23/2024, 3:32:32 PM
X
In reply to:
Connor Axiotes
@connoraxiotes
·
262d
@GaryMarcus You might be right. Although I think what might constitute a bigger part of GPT-5 is the reasoning capabilities rumoured (Q* stuff) to have spooked Ilya.

Turbo doesn't seem to have a scary enough element that caused alarm to some OpenAI safety staff.
Gary Marcus
@GaryMarcus
·
262d
@connoraxiotes my guess:
Gary Marcus
@GaryMarcus
·
431d
🤔 The news is the price and speed, not radically new capabilities

Given the cutoff date, and what Altman said when we were at Senate in May … my gut tells me that GPT-4 Turbo is a relabeling of what was supposed to be called GPT-5, and not enough a cognitive breakthrough merit…
Connor Axiotes
@connoraxiotes
·
262d
@GaryMarcus Could this just be that OpenAI are the best at training these models, and when GPT-5 comes out *that* is the time to see if diminishing returns are setting in?
Gary Marcus
@GaryMarcus
·
262d
13.5 months post GPT-4:

• Models that are not as good as GPT-4 have become widespread.

• Models that are roughly as good GPT-4 have become common.

• Models that are significantly better than GPT-4 don’t exist.

If that doesn’t hint at diminishing returns and potential…
Filip Piekniewski🌻 🐘:@filippie509@techhub.social
@filippie509
·
262d
If the boatload of LLMs available already wasn’t enough for you, here is another one.

Plenty supply, just still don’t see any serious application. Nose bleed valuations. What could go wrong?

The statement engages in a discussion about the potential diminishing returns of scaling AI models and the need for new mechanisms to enhance their capabilities. It is a part of a broader conversation about the development and implications of AI technologies like GPT-5, touching on issues of technological progress and safety concerns.

  1. Principle 1:
    I will strive to do no harm with my words and actions.
    The statement does no harm and contributes to a technical discussion about AI development. [+1]
  2. Principle 2:
    I will respect the privacy and dignity of others and will not engage in cyberbullying, harassment, or hate speech.
    It respects the dignity of others and maintains a professional tone without any personal attacks. [+1]
  3. Principle 3:
    I will use my words and actions to promote understanding, empathy, and compassion.
    The comment promotes understanding by discussing the limitations and future directions of AI technology. [+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.
    It engages constructively by adding to the ongoing dialogue about AI capabilities and their implications. [+1]
  5. Principle 5:
    I will acknowledge and correct my mistakes.
    The statement is open to correction and further discussion, as it presents an opinion as part of a debate. [+1]
  6. Principle 6:
    I will use my influence for the betterment of society.
    It uses influence responsibly by sharing insights on AI development, which could inform public understanding and policy. [+1]
  7. Principle 7:
    I will uphold the principles of free speech and use my platform responsibly and with integrity.
    The statement upholds free speech by contributing thoughtfully to a public technical discussion. [+1]