Conversation

Notices

1. Embed this notice
   Darnell Clayton :verified: (darnell@one.darnell.one)'s status on Tuesday, 09-Apr-2024 02:48:28 JST Darnell Clayton :verified:

   So #OpenAI illegally transcribed videos from #YouTube, & #Google is not going after them because…hmm…I wonder if Google is doing the same thing‽

   👉🏾 OpenAI transcribed over a million hours of YouTube videos to train GPT-4 - The Verge https://www.theverge.com/2024/4/6/24122915/openai-youtube-transcripts-gpt-4-training-data-google

   • Embed this notice
     Darnell Clayton :verified: (darnell@one.darnell.one)'s status on Tuesday, 09-Apr-2024 03:02:23 JST Darnell Clayton :verified:

     in reply to

     • jonathankoren

     @jonathankoren OpenAI did not receive prior permission to transcribe from Google, & some of that data would include random videos of people with their families talking.

     Also pulling that much data would be expensive!

     I guess it did explain why Google updated their ToS when they did (probably to specifically prevent this from happening again.

   • Embed this notice
     jonathankoren (jonathankoren@sfba.social)'s status on Tuesday, 09-Apr-2024 03:02:24 JST jonathankoren

     in reply to

     @darnell What does “illegally transcribed” mean? By that logic, I was breaking the law by repeatedly listening to songs and writing down the lyrics, or when I wrote down quotes from The Simpsons.

     This is not how the law works, nor should it work.