New piece on the problems with AI training – and the secrecy of GPT-4. @NewYorker on the data, labor & energy issues: "leaving aside the question of AGI…it is clear that large-language engines are creating real harms to all of humanity right now.

→ View original post on X — @katecrawford,