Raw LLM Responses

>If you want to get some impression of what could be so scary. Read this article: [AI suggested 40,000 new possible chemical weapons in just six hours](https://www.theverge.com/2022/3/17/22983197/ai-new-possible-chemical-weapons-generative-models-vx). > Meh, as a chemist who's actually interacted with ChatGPT I wouldn't even begin to worry. Current AI is useless for chemistry at a higher-level due to its frequent hallucinations. Plus, there's already purpose-built, commercial software for drug discovery that predicts docking interactions exponentially better (e.g. Schroedinger). Typical doom-and-gloom from a source who only knows enough about the subject to be dangerous. It'll be scary eventually, no doubt, but we ain't there just yet.