Gary Marcus

Rank 14 of 47

Score 83

@peakcooper @fchollet if they are all training on 100% (I haven’t followed the benchmark), none of them are really comparable to people.

12/21/2024, 9:03:06 PM

In reply to:

Cooper

@peakcooper

370d

@GaryMarcus @fchollet Gary, the previous solutions all trained on 100% of the training set, yet got results like 5%
What gives?

Gary Marcus

@GaryMarcus

370d

Update: It IS true, from @fchollet’s blog. 𝗢𝟯 𝘄𝗮𝘀 𝘁𝗿𝗮𝗶𝗻𝗲𝗱 𝗼𝗻 𝟳𝟱% 𝗼𝗳 𝘁𝗵𝗲 𝗽𝘂𝗯𝗹𝗶𝗰 𝘀𝗲𝘁 𝗳𝗼𝗿 𝗔𝗥𝗖-𝗔𝗚𝗜.

Yesterday’s announcement shows what you can do when teach to the test and spend enormous amounts of money.

How much the demo says about real…

Gary Marcus

@GaryMarcus

370d

Wow. This, if true, raises serious questions about yesterday’s announcement.

The statement engages in a technical discussion about AI training benchmarks and their implications, questioning the comparability of AI models to human performance. It reflects skepticism about the validity of AI training methods and results.

Principle 1:
I will strive to do no harm with my words and actions.
The statement does not cause harm and contributes to a technical discussion, maintaining a neutral tone. [+1]
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 training methods and results, without resorting to personal attacks. [+1]