For exploring topics in a shallow fashion is fine with LLMs, doing anything deep is just too unreliable due to hallucination. All models I’ve talked to desperately want to give a positive answer, and thus will often just lie.