Anthropic when Chinese companies use Claude to train their own AI models vs Anothropic when they’re stealing everyone’s data to train Claude

[malwaretech@infosec.exchange]

Anthropic when Chinese companies use Claude to train their own AI models vs Anothropic when they’re stealing everyone’s data to train Claude

[tchauhan@mastodon.mit.edu]

@malwaretech So the Chinese companies actually had to pay to train their models ? That sounds terrible. Wait. No. That sounds almost fair ?

[kevingranade@mastodon.gamedev.place]

@tchauhan @malwaretech yep, it's not even any kind of bypass, just, "we disapprove of what you're using this legitimately obtained access to do".