"What does it really cost to build and run this?" is a very sensible question to ask yourself about your nascent project #ArtificialIntelligence project involving #LargeLanguageModels (#LLM) or #GenerativeAI more broadly. More than ever, the answer to this question might depend hugely on the fine details of what you want to achieve. In this video, I talk about how the guardrails you put on your implementation (data privacy, vendor stability, and quality control) can quickly add complexity that will put pressure on your bottom line.