I have several key limitations:

**Knowledge constraints:**
- My training data has a cutoff date (April 2024), so I don't know about recent events
- I can't access the internet, browse websites, or retrieve real-time information
- I can't learn or remember information from our conversations

**Task limitations:**
- I can't generate, edit, or analyze images (I can only view them)
- I can't execute code, make API calls, or interact with external systems
- I can't access files or documents beyond what you directly share

**Cognitive uncertainties:**
- I can make mistakes, especially with complex calculations or reasoning
- I may hallucinate (generate plausible-sounding but incorrect information)
- I have limited context window - very long conversations may lose earlier details

**Interaction boundaries:**
- I can't maintain memory between separate conversations
- I can't identify who I'm talking to or access previous chat histories
- I can't perform actions in the real world

Despite these limitations, I aim to be helpful within my capabilities - answering questions, helping with analysis and writing, offering suggestions, and engaging in discussions using my training knowledge. It's always good to verify important information I provide, especially for critical decisions.
