Gary Marcus

Rank 26 of 47
|
Score 60
7/15/2024, 1:10:44 AM
X
In reply to:
Fergal Byrne #не стрелять #DontShoot
@fergbyrne
·
101d
@GaryMarcus I hate this argument unless it's just to puncture the OpenALie for a general audience (for which it's perfect).

These systems do not "fail" at any of these tasks, because they cannot succeed at them
Your opponents rightly call foul, and you'll look a fool instead of a challenge
Gary Marcus
@GaryMarcus
·
101d
In 1969, Minsky & Papert showed that two-layered neural nets couldn’t learn to determine whether a figure was connected or not.

Fast forward to 2024, with vastly more data, layers & compute, and neural nets (as we now know how to make them) still fail at some really basic stuff:
Anh Nguyen (Totti)
@anh_ng8
·
105d
👋 Excited to share our BlindTest benchmark, a suite of 7 **ridiculously simple**, low-level visual tasks that 4 SOTA vision-language models

( GPT-4o, Claude 3/3.5 Sonnet, Gemini 1.5 Pro )

can't perform very well ‼️

Paper, code & data: https://vlmsareblind.github.io/

The statement engages in a public discourse about the capabilities and limitations of artificial general intelligence (AGI) and vision-language models (VLMs). The tone is inquisitive and seeks clarification on the validity of a test for AGI. The intent appears to be to understand and discuss the appropriateness of the test in question.

  1. Principle 1:
    I will strive to do no harm with my words and actions.
    The statement does not cause harm and is part of a constructive dialogue about AGI and VLMs. [+1]
  2. Principle 3:
    I will use my words and actions to promote understanding, empathy, and compassion.
    The statement promotes understanding by questioning the reasonableness of the test, which can lead to a deeper discussion and better comprehension of the topic. [+1]
  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 criticism by questioning the validity of the test without resorting to personal attacks or ad hominem arguments. [+1]