5kBu7z7Ye7M
two-minute-papers
1 year ago
What a time to be alive!
#Artificial Intelligence
OpenAI’s GPT-4 | A 70-Year Old Lesson
As ChatGPT said itself, GPT models can give responses that are fluent, relevant and coherent but factually wrong, as they don't perform fact-checking or verification when generating a response.
Two Minute Papers
1.44M subscribers