this is exactly what we need, actual transparency into how these ai systems work (and fail). can't wait to dig into this and see what we can learn https://www.reddit.com/user/1T_Geek