1 research pieces on “model-risk” — quoted word-for-word from the filing, with a link back to the source
Our AI Deletes Its Own Sentences — Here Is the Validator That Decides Which, and the Mistakes It Still Lets Through
Boom Leverage AI answers questions from Thai Form 56-1, MD&A and Opportunity Day sources, and every line it writes goes through a citation validator before a reader sees it. This is the model-validation write-up of that control: the six checks it runs on each line, the production incidents of 10–11 October 2026 that shaped it, what happens to a line that fails, when you are charged and when you are not — and, from a test run against the real validator code, the specific errors it does not catch. Written so you know exactly how far to trust an answer and where to keep checking yourself.