Gary Marcus

Rank 13 of 47
|
Score 115
9/14/2024, 3:57:59 AM
X
In reply to:
Gary Marcus
@GaryMarcus
·
361d
@sjgadler @tylercowen thank you i wrote a detailed line by line evaluation of that article two years later, evaluating each claim within the article, within an implicit evaluation of the title. I appreciate you for being a respectful person during @tylercowen’s temporary lapse from professionalism:…
Steven Adler
@sjgadler
·
361d
@GaryMarcus @tylercowen I think the article’s points are much more reasonable than the headline, FWIW - the headline just seems very incorrect. I think many fewer would have taken issue with the article if it had been called e.g. “To achieve general intelligence, symbolic representations might help”
Steven Adler
@sjgadler
·
361d
@GaryMarcus @tylercowen With due respect, the headline claim here - that “deep learning is hitting a wall” - seems quite mistaken, given it was published five days before even GPT-3.5’s release. Perhaps you didn’t write this headline though, as authors often don’t? https://nautil.us/deep-learning-…
Gary Marcus
@GaryMarcus
·
361d
@tylercowen All of this has held this far as well. Just giving few more examples. you have given a total of zero.
Gary Marcus
@GaryMarcus
·
548d
Prediction: By end of 2024 we will see
• 7-10 GPT-4 level models
• No massive advance (no GPT-5, or disappointing GPT-5)
• Price wars
• Very little moat for anyone
• No robust solution to hallucinations
• Modest lasting corporate adoption
• Modest profits, split 7-10 ways
Gary Marcus
@GaryMarcus
·
361d
@tylercowen Still incredibly empty and nonspecific. My predictions of the following have all been correct, and generally doubted when initially made

Tendency towards hallucinations (2001 for deep learning in general, 2019 LLMs)
Tendency towards erratic reasoning (2001 and 2019 respectively)…
tylercowen
@tylercowen
·
361d
@GaryMarcus If someone wanted to understand the last say five to seven years in AI, and how they have unfolded, your earlier writings simply are not where anyone gets sent. No one contests that.

The statement constitutes public discourse as it engages in a substantive discussion about the capabilities and limitations of deep learning, a significant topic in the field of artificial intelligence. The tone is analytical and aims to correct or refine a previous statement for accuracy.

  1. Principle 1:
    I will strive to do no harm with my words and actions.
    The statement does not appear to cause harm and is focused on improving the accuracy of a title, which aligns with the principle of doing no harm. [+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. [+1]
  3. Principle 3:
    I will use my words and actions to promote understanding, empathy, and compassion.
    The statement promotes understanding by suggesting a more accurate title, which can help clarify the discussion for others. [+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 suggesting a more accurate title 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 by contributing to a more accurate and informed discussion. [+1]