Tomasz Tunguz@tunguzSep 17, 2026, 21:50One of the best and most nuanced posts about AI risk that I’ve come across.打开原帖#511482