Comment by in-silico

LLMs assign non-zero probability to every output (unless you you sampling modifiers like top-k), so if you sampled enough times the model would definitely come up with velcro. It's the same as the proverbial monkey with a typewriter.

The real question is whether it would happen in a practical timeframe.