“The Speculative Fiction Indie Novella Championship (SFINCS, pronounced “sphinx”) is a yearly competition to recognize, honor, and celebrate the talent and creativity present in the indie community. We are a sister competition to both SPFBO and SPSFC, and we highlight greatness in the novella format in all areas of speculative fiction (fantasy, science fiction, horror, etc.).” – From the official SFINCS website.
Note: The following review contains only my personal thoughts as a judge and does not reflect the views of the team as a whole.
REVIEW
Olga Werby’s Good Girl is a sci-fi puzzlebox focused on an interstellar mission to the Tau Ceti star system. Its tight cast includes the skeleton crew assigned to prepare the planet of Terra II for future colonists and the ship’s AI whose job it is to ensure that mission is a success. Fans of the short-lived TV series 1899, as well as Mur Lafferty’s Six Wakes, are likely to enjoy this one. Aside from the novella’s sci-fi setting and atmosphere, Good Girl is also notable as a story about stories. While the narrative remains tense throughout, keeping you guessing (and reading) as you attempt to parse its secrets, its emphasis is less on overt action than on interiority, interpersonal relations, and compelling meditations on art, language, and their meanings. If Good Girl sounds like your kind of story, I recommend going in without reading the blurb (or perhaps even the rest of this review, though I won’t share any specific spoilers).
Typically, this is the paragraph where I detail the basic plot of a work, but the puzzlebox nature of Good Girl makes me hesitant to give away too much more than I have above. I can say that Good Girl is the story of a ship’s AI named Saia, several crew members under her protection (most notably an autistic-coded character called Null), and Saia’s task to keep the humans under her care sane over the course of their lengthy journey. Saia has been programmed as a self-learning AI, who tailors the entertainment, diversions, and enrichment she provides to each crew member according to their particular interests and personality. Much of that entertainment is literary and linguistic (particularly in Null’s case). As Null, Saia, and another charge named Ray, challenge one another to interpret the meaning underlying works such as The Master and Margarita and Alice’s Adventures in Wonderland, Werby takes the opportunity to explore themes like the value and meaning of art as it relates to morality, consciousness, and the nature of right/wrong and good/evil themselves. Since each of the human characters is isolated from the others (indeed, it’s not even clear whether the company funding the expedition has allowed them to be aware of one another’s existence), they have plenty of time and mental space for these kinds of interactions and analyses.
The literary meta-analysis and its potential use as a lens through which to understand Good Girl‘s own plot was probably my favourite element of the story. In part, this is due to my own love of Mikhail Bulgakov’s The Master and Margarita (the novel I wrote one of my MA disserations on–SFINCS this year is starting to make me think I might be in a puzzlebox storyline myself, given that that’s two novellas so far that have so directly involved my academic interests, and which both involve Watcher-related stories). What this literary interest allows Werby to do, beyond examine the works being discussed, is use the characters’ interpretations and experiences of the works to reveal the ways in which Saia, Null, and the others each perceive the world and their relationship to it. Since Saia is conscious for the entirety of the trip (vs. Null and his fellow humans, who are only occasionally woken in order to perform maintenance tasks), the most powerful character work comes through in the changes she experiences as she struggles to interpret Bulgakov’s themes, context, and messages.
“Ultimately any story is a something the author and the reader create together. The meaning is always a negotiated activity.”
One might be tempted to view Good Girl as a comment on contemporary LLMs. Certainly, its portrayal of the ways in which Saia is at least initially unable to grasp or fully understand meaning beyond the surface of a text feels like a reflection on the unique relationship humans have with art. One interaction in particular, during which Ray explains to Saia that the original illustrations of Alice are more impactful than photographs (or, by extension, photorealistic representations) of Wonderland would be, feels like a commentary on trends within a number of recent artistic fields toward hyperrealism.* Here, I’m thinking largely of videogames and the reliance of the film industry (particularly children’s film) on hyperrealistic CGI, though one might also gesture toward the uncanny un-realism of many AI-generated images. Ray’s point, that John Tenniel’s Alice illustrations communicate much better the feel of Carroll’s text while giving audiences room to fill in the details with their imaginations, undermines not only an understanding of exactness or realism as being synonymous with accuracy and efficiency, but ultimately undermines the notion of accuracy itself. I’m reminded of attempts on the part of AI users to “add” to beloved works of art by prompting image generators to “complete” those paintings so that audiences might see what exists in a nebulous beyond that supposedly lies outside the canvas. To even ask the question, however, only reveals the ignorance of the questioner.
That said, Werby’s Saia is no LLM. She has much more in common with artificial intelligence as it has been imagined in science fiction since the genre’s inception. While programs like MidJourney or CatIFarted involve machine learning in its broadest sense (at least, based on my understanding), their purpose is ultimately regurgitation rather than the development of unique thought. Any notion to the contrary merely reveals the vastness of the scope of work and of real human interactions that have been trawled from a much richer and diverse ocean of artistic and social meaning. Good Girl might ask, as part of its own internal examination of AI, what such programs are thinking and doing inside the black boxes to which we lack complete access, but I fear the real answer is “not much,” especially in comparison with Saia’s complexity. That said, for readers interested in the anxieties and desires that we foster when it comes to both the concept of “true” artificial intelligence and what people like Sam Altman want us to believe are artificial intelligences, Good Girl is likely to be a compelling read.
Puzzlebox is one of my favourite genres, but while it rarely fails to suck me (or, to judge by the popularity of shows like Lost, a larger audience) in, it’s also one of the hardest genres to pull off in a way that’s satisfying. While a weak or unconvincing ending can ruin many stories, even when the rest of the work was well-received, in puzzlebox that tendency is amplified. If a reveal or an answer is too obvious or too discordant and detached from the clues that preceded it, audiences can feel cheated, while an ambiguous ending can sometimes have fans asking whether even the author(s) know the answers to the questions they raise over the course of a narrative. Good Girl is a bit unusual, in that I found the answers it provided largely very satisfactory, but in that I felt the ending itself was weak due to its abruptness (and a little bit of cheesiness in its final scenes). Its reliance on popular and harmful tropes around schizophrenia and psychosis, which I suspect were entirely unintentional, added to the Hollywood-esque feel of the conclusion (a feel that was lacking elsewhere in the text and which therefore didn’t entirely land). While the prose and pace elsewhere in the novella felt consistent and easy to sink into, the end as it was presented included a jarring tonal shift. What could have been a spectacular finish was, for me, a little underconsidered. Giving the text a little more room to breathe in its last few sections/chapters would have strengthened the entire novella.
The reveal itself, as I’ve said, felt justified by what preceded it and was otherwise satisfying, but it was undermined somewhat by the marketing surrounding it. I’m generally opposed to the idea that reviews should act in any way as learning opportunities for authors (though I understand that many feel differently). That said, if I were to give one piece of advice to Werby, it would be to tweak Good Girl‘s blurb slightly so that it gives away less of the text’s delicious twists and turns. It’s a powerful story, and perhaps just needs a little more confidence in itself to shine.
Good Girl was easy to read, a pleasure to sink inside, and I found both Null and Saia in particular to be fascinating characters. I haven’t discussed character too much, but I was immediately compelled by Null in particular, and I have feeling a lot of neurodivergent readers will appreciate him. Werby is clearly a thoughtful author with much to say–the kind of writer I’d honestly love to sit down and chat literature with. The quality of Good Girl as a digital object is very high, its prose accomplished and polished. In spite of my issues with the conclusion, I would recommend Good Girl to others. It’s a book that will leave you wanting to discuss it afterwards, and that’s not true of very many works.
Buy Good Girl from Amazon.
Add Good Girl on Goodreads.
*The self-indulgent part of me is also tempted to read this observation as a critique of Tim Burton’s Alice, which I feel deserves exactly the amount of loathing I feel toward it (a lot).
