←back to thread

GPT-4 and professional benchmarks: the wrong answer to the wrong question

(aisnakeoil.substack.com)
340 points agomez314 | 2 comments | 21 Mar 23 13:12 UTC | HN request time: 0.013s | source
Show context
ChancyChance ◴[21 Mar 23 13:58 UTC] No.35246181[source]▶
"10/10 pre-2021, 0/10 post-2021"

I guffawed. Hasn't there been ANY update to the Turing test since it was proposed? I suspect the answer to the benchmarking issue was address by philosophers long before we got here.

replies(2): >>35246541 #>>35248367 #
1. sebzim4500 ◴[21 Mar 23 14:21 UTC] No.35246541[source]▶
I think people completely lost interest in the Turing test as soon as it looked like something might actually be able to pass it.
replies(1): >>35249895 #
2. schrodinger ◴[21 Mar 23 17:58 UTC] No.35249895[source]▶
The Turing test has now become "anything ChatGPT cannot solve yet" :)