Gary Marcus

Rank 9 of 47

|

Score 112

+2

@flowersslop @captain_marrvel that’s actually a really interesting measure. [link?]

GPT-4 vs GPT 3 would be +what 90%+?
GPT-3 vs GPT 2 also 90%?
GPT-2 vs original GPT, also 90%?
what will be GPT-5 vs GPT-4 be?

7/21/2025, 4:59:53 PM

In reply to:

Flowers ☾ ❂

@flowersslop

·

3d

@captain_marrvel @GaryMarcus 4.5 only wins against 4o in 60% of cases in oais own internal preference benchmark

MarrvelSystems

@captain_marrvel

·

3d

@flowersslop @GaryMarcus GPT-5 is literally o4 and 4.5, why the hell y’all comparing it to 4o which is 10x worse than OpenAI current o series models?

Flowers ☾ ❂

@flowersslop

·

3d

@GaryMarcus I have stricter expectations. I will do a blind test to see if I prefer GPT-5 over GPT-4o at least 8/10 times, if not, that would be a crazy fail.

Because it should actually be 10/10 times.

Gary Marcus

@GaryMarcus

·

3d

last year people were claiming GPT-5 was going to be some monstrously large whale.

now? they are hoping for a 10-20% improvement.

Chris

@chatgpt21

·

3d

@krapstarr @GaryMarcus 10-20%

The statement engages in a technical discussion about the performance improvements between different versions of GPT models. It reflects curiosity and interest in the progression of AI technology, specifically comparing the performance metrics of various iterations of GPT models. The tone is analytical and speculative, focusing on potential improvements in AI capabilities.

Principle 1:
I will strive to do no harm with my words and actions.
The statement does no harm and is focused on a technical comparison, adhering to the principle of doing no harm. [+1]
Principle 3:
I will use my words and actions to promote understanding, empathy, and compassion.
The statement promotes understanding by engaging in a discussion about AI advancements, aligning with the principle of promoting understanding and knowledge. [+1]