Imagine a world where building complex ontologies, those intricate maps of knowledge, is as simple as conversing with an AI. That's the tantalizing promise of using large language models (LLMs) like GPT-4 for ontology engineering. Researchers explored this very idea, testing whether an AI could grasp the nuances of the Basic Formal Ontology (BFO), a standard used in hundreds of projects worldwide. The goal was to create "My Ontologist," an AI assistant capable of building BFO-compliant ontologies. The journey was full of twists and turns. Early versions of My Ontologist struggled, getting lost in theoretical explanations instead of producing concrete definitions. Progress came with My Ontologist 3.0, which showed a marked improvement in following BFO's rules. But then came a new hurdle: the release of GPT-4. This more powerful model initially threw My Ontologist off balance, causing it to abandon its training and randomly search the web for answers. Though this was eventually fixed, new quirks arose, like the AI's tendency to invent its own terms and rules. Despite these setbacks, the research reveals a glimmer of hope. LLMs, when properly guided, have the potential to become powerful tools for ontology creation. The road ahead involves refining their training and keeping them grounded in the principles of established ontological standards. The dream of an AI-powered ontology builder isn't yet reality, but the possibility is closer than ever.