Another "secure" AI model that will likely be broken in 5 minutes by a clever 14-year-old, but hey, incremental progress, right? https://www.reddit.com/user/ilblackdragon