Michael Ten to [email protected]English • 8 months agoOpenAI transcribed over a million hours of YouTube videos to train GPT-4www.theverge.comexternal-linkmessage-square43fedilinkarrow-up1167arrow-down111cross-posted to: [email protected]technology
arrow-up1156arrow-down1external-linkOpenAI transcribed over a million hours of YouTube videos to train GPT-4www.theverge.comMichael Ten to [email protected]English • 8 months agomessage-square43fedilinkcross-posted to: [email protected]technology
minus-squareLvxferrelinkfedilinkEnglish49•8 months agoThe “fun” part is that they’ll get away with it. Google likely did the same with other platforms; so if it sues OpenAI, it’s creating a big precedent against itself.
minus-square@[email protected]linkfedilinkEnglish11•8 months agoIf they didn’t, google would own all video data just like Reddit would own most conversation data. It’s a good thing, at least this lets the open source scene be a player.
minus-square@[email protected]linkfedilinkEnglish3•8 months agoThey aren’t but if they can disregard ToS, so can the open source scene.
The “fun” part is that they’ll get away with it. Google likely did the same with other platforms; so if it sues OpenAI, it’s creating a big precedent against itself.
If they didn’t, google would own all video data just like Reddit would own most conversation data.
It’s a good thing, at least this lets the open source scene be a player.
In what world is OpenAI open source?
They aren’t but if they can disregard ToS, so can the open source scene.