LLM Limitations and Knowledge Cutoff Explained

Known Weak Spots

Because an LLM generates text that sounds plausible, it can state wrong things with total confidence. It may also not know about recent events that happened after its training data was collected — this point in time is called its knowledge cutoff.

LLMs can also struggle with exact arithmetic over many steps, precise counting, and tasks that need fresh, private, or real-time data they were never given.

Treat model output like a draft from a fast, well-read intern: useful, but worth checking when it matters.