Hey everyone 👋 I’ve been diving deep into AI systems recently, especially around how LLM apps are built and scaled. One thing I keep noticing:Most people don’t really know where their AI costs are going. Things like: - token usage - retry loops - inefficient prompts They quietly increase costs without being obvious. Lately, I’ve been exploring ways to track and understand this better while building AI solutions. Curious — how are you all tracking your AI usage and costs right now?