谷歌当年求着出版商把书交出来做Google Books搜索，现在训练Gemini顺手把版权信息改了改，这操作真的是把“借书不还”做成了商业模式……加州那两个判例一开先例，AI公司训练数据这事基本就是变相追认既成事实，法律永远慢一个版本号，等它更新完这波早就训练完下一代模型了
google literally begged publishers to hand over their books for Google Books search back in the day, now they're quietly tweaking copyright metadata to train Gemini... "borrow and never return" as a business model, legendary behavior tbh
those two CA rulings just opened the floodgates — AI companies training on scraped data is basically getting retroactively blessed at this point. law's always running a version behind, by the time it patches this, they've already trained the next gen model lol
https://x.com/Dr_goosee/status/2077212151555756454
google literally begged publishers to hand over their books for Google Books search back in the day, now they're quietly tweaking copyright metadata to train Gemini... "borrow and never return" as a business model, legendary behavior tbh
those two CA rulings just opened the floodgates — AI companies training on scraped data is basically getting retroactively blessed at this point. law's always running a version behind, by the time it patches this, they've already trained the next gen model lol
https://x.com/Dr_goosee/status/2077212151555756454
Comments
Be the first to add a comment
Log in to comment