Gary Marcus

Rank 13 of 47
|
Score 83
4/30/2025, 12:53:47 AM
X
In reply to:
0.005 Seconds (3/694)
@seconds_0
·
258d
@GaryMarcus Thats why I am looking for enhanced definitions and elaborations on constraints or measurement criteria. For 6, does it have to one shot 10k lines, or does reliably output mean it can write lines with tests, see the tests fail, and automatically iterate until it works?
Gary Marcus
@GaryMarcus
·
258d
@seconds_0 I’ve heard a number of false reports like this before. If someone can do these systematically and reliably, I’ll be interested.
0.005 Seconds (3/694)
@seconds_0
·
258d
@GaryMarcus I am looking at your Brundage Tasks and am curious how you evaluate these because I am almost certain Gemini 2.5 could do 2, today

For write obituaries or bios, what information are they given? Just access to the weights? Do they need to do original research

6 I have done today
Gary Marcus
@GaryMarcus
·
258d
@seconds_0 see https://garymarcus.substack.com/… and https://garymarcus.substack.com/…

I have made many specific falsifiable predictions and offered to put money on them.

then again i don’t hate AI, and people who think I do misunderstand my position.
0.005 Seconds (3/694)
@seconds_0
·
258d
These types of claims are never made falsifiable. In fact, most AI haters refuse to make falsifiable claims, in writing, about what AI won't be able to do or won't progresss.

Anyone making these claims without a definition can safely be ignored.
dax
@thdxr
·
259d
LLMs will never get meaningfully better than they are now

The statement is part of a broader discussion on AI capabilities and evaluation criteria, which constitutes public discourse as it engages with public issues related to technology and its development.

  1. Principle 1:
    I will strive to do no harm with my words and actions.
    The statement does not cause harm and is a neutral expression of opinion on AI debugging, adhering to the principle. [+1]
  2. Principle 3:
    I will use my words and actions to promote understanding, empathy, and compassion.
    The statement does not actively promote understanding, empathy, or compassion, but it does not detract from it either.
  3. 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 dialogue by referencing a study to support its point, without resorting to personal attacks. [+1]