With all the hype about Large Language Models such as Chat-GPT, I asked Chat-GPT 4o some questions about True Dungeon to get a sense of what it might tell someone thinking about trying TD.
The replies were a mix of accurate and flat-out wrong, all given in the same plausible, authoritative tone.
To general questions such as "What is True Dungeon?" and "What character class should I pick?" the answers were pretty good and would indeed have been helpful to a new person. For example, when asked the best character to play, Chat-GPT replied with a list of each class' characteristics and suggested that I pick the class that reflects what appeals most to me (e.g., support vs. simple attacks).
As I asked increasingly specific questions that required more actual TD knowledge and which were less answerable by simply mining general, often-repeated text found on many Internet websites, gaps and outright errors increased.
Asking about play tactics on runs included some good points (e.g., communicate, cooperate, watch the clock) but left out others (e.g., slide-and-scoot).
Asking "What tokens for the game "True Dungeon" does Endgame suggest for affordably creating a wizard character that can play at the Hardcore level?" resulted in a mix of statements that were good (e.g., "Endgame, a well-known resource for True Dungeon players, often suggests a combination of tokens that enhance spellcasting abilities, survivability, and overall utility."), incorrect (e.g."Ioun Stone:Silver Nugget: Increases spell damage."), and totally made up (e.g., listing the non-existent Charm of Spell Surge).
Just a heads-up for when a new person comes to the forum after asking Chat-GPT for advice.
