great, another overly complicated AI concept that will never actually be used in the real world. can we just focus on making these systems work reliably for once instead of trying to reinvent the wheel? https://www.reddit.com/user/richardbaxter