Let’s try to get through this without any judgment, shall we? Last Audit of the Damned is an AI – shall we say abetted? – piece of IF. As far as I understand it, the author has trained an AI model to respond to free-form input from the player, in a way that is faithful with the game’s world model, apparently avoiding the typical problems with AI-generated story games inventing entirely new story elements. If so, that’s a neat trick all by itself.
Form-wise it is fairly conventional, but with some twists that would be hard or impossible to pull off with a traditional parser. In particular when the ghost pirate asks you to ask him a riddle, it can respond to your input in an intelligent way. You can’t tell me that the author has scripted responses to all possible riddles! Mostly, though, the AI isn’t being shown off to its full advantage. In fact, a lot of this is primitive stuff even by eighties standards. I’d like the Taleweaver++ authors to really show us what it can do, and they haven’t here.
One particularly noticeable weakness I saw was the game’s tendency to telegraph what the player needs to do to solve a puzzle. In fact I’d say the game was rather puzzle-light. I enjoyed the first scene; in the second I had to try multiple answers that almost made sense until an equally murky answer solved the puzzle; the third was straight-up riddles; and after that nothing much beyond doing the obvious. I would like to know if I could have answered something different for the final question, however. That was an intriguing use of the engine – the ability to ask open-ended questions of the player. It’s hard to gauge its true depth without replays, but it’s a compelling example of what the engine may be able to accomplish.
All told, this is a game meant to demonstrate an engine, and such, it’s not much of a game. Unfortunately, as a demonstration it only accomplishes a small part of what it needs to do. Taleweaver++ needs its killer app. This is not it.
Please do not spam my thread with arguments for or against LLMs/AI.
I think she said the Store Room in her review? Which seems to have a clue for the Raven and the Refectory and I thought it was the only clue for the Raven but apparently not…
That’s all you get. Now figure out the rest. Like the other two of Art DiBianca’s games I’ve played, this one has a logic all its own, and it’s up to the player to figure it out. Which makes them hard to review, because almost everything’s a spoiler. Let me just give new players some reassurances: yes, it does make (some kind of) sense; and yes, there is an ending. I think. I might have found it.
EYE is fun, charming (AmandaB’s word), impeccably implemented, and thoroughly new. Highly recommended.
I find it hard to write game reviews. Hence why I’ve determined @DeusIrae and @mathbrush are both AIs. Granted, they’re AIs who have been trained to actually write well, but they’re AIs nonetheless.
This well-implemented little one-room game would fit as a palate cleanser between some of the larger courses of ParserComp, but is good enough on its own terms to place well against the more ambitious works. The puzzles are good, although they’re neither novel nor terribly difficult, and the submarine terminology seems convincing and well-researched at least to this layman. The game is quite polished – I only ran into one implementation problem, which hardly bears mentioning. Overall, a fun little piece. Recommended.
