Found 71 bookmarks
Custom sorting
(1) Rohan Paul on X: "GPT 5 Rumored Benchmark through Copilot. SimpleBench is a roughly 200-question multiple-choice benchmark that targets spatio-temporal, social, and adversarial reasoning. A 90% scroe is quite insane here, as it represent a human-level common-sense reasoning equivalent. https://t.co/X01PC3CDwC" / X
(1) Rohan Paul on X: "GPT 5 Rumored Benchmark through Copilot. SimpleBench is a roughly 200-question multiple-choice benchmark that targets spatio-temporal, social, and adversarial reasoning. A 90% scroe is quite insane here, as it represent a human-level common-sense reasoning equivalent. https://t.co/X01PC3CDwC" / X
#GPT-5#Benchmark#Copilot
·x.com·
(1) Rohan Paul on X: "GPT 5 Rumored Benchmark through Copilot. SimpleBench is a roughly 200-question multiple-choice benchmark that targets spatio-temporal, social, and adversarial reasoning. A 90% scroe is quite insane here, as it represent a human-level common-sense reasoning equivalent. https://t.co/X01PC3CDwC" / X
First Look: Exploring OpenAI o1 in GitHub Copilot
First Look: Exploring OpenAI o1 in GitHub Copilot
(Microsoft Copilot does offer a notebook interface for writing in one box and seeing output in another. more like the development platforms. It suggests a brief summarization of the main points at the start of prompts for the serial interfaces.)
#Copilot#GitHub#OpenAI#Demos
·github.blog·
First Look: Exploring OpenAI o1 in GitHub Copilot