Can artificial intelligence be truly creative? It's a question that has captivated and concerned us in equal measure. Recent research delves into this fascinating area by examining not just *what* large language models (LLMs) create, but *how* they create. Think of it like this: we've been judging AI art by looking at the finished painting, but now we're peeking into the artist's studio to understand their process. Researchers explored this by giving humans and LLMs creative tasks, like coming up with alternate uses for everyday objects (a brick, a paperclip) and comparing how they explored different ideas. Interestingly, they found that LLMs, like humans, can be "persistent" (diving deep into a few core ideas) or "flexible" (jumping between many different concepts). While both humans and LLMs showed these tendencies, the link to creativity was different. For humans, both approaches yielded similar levels of creative output. However, for LLMs, the more flexible models were judged as more creative. This research opens up exciting new avenues for understanding both human and artificial creativity. It suggests that while LLMs can achieve impressive creative feats, the underlying process might differ significantly from our own. This also has implications for how we might collaborate with AI in creative endeavors. Imagine partnering with an LLM that complements your own creative style, pushing you to explore new ideas and perspectives. While this research provides a fascinating glimpse into the creative potential of LLMs, it also highlights the need for further investigation. How do different prompting strategies influence the creative process? How can we better evaluate the true creativity of AI, moving beyond simple originality scores? These are just some of the questions that remain to be explored as we continue to unravel the mysteries of creativity in both humans and machines.