“The woods are just trees” - ?
Short-term payoffs of AI are promising. In the systems we deploy, how can we maintain a favorable benefit-to-risk ratio over the long term, given the threats and systemic flaws that LLMs inevitably introduce?
This site shares ideas and experiences to help guide others on their own journey to secure AI systems. This is a journey rife with surprises, good and bad, in which we must constantly update our mental model and stay wary of oversimplification.
