Scan barcode
A review by thogek
Parallel Realities: A Turing Fiction by K.R. Simms
adventurous
challenging
medium-paced
2.0
Simms's Parallel Realities builds on an interesting concept. With a new brain implant embedded with advanced AI to assist his recovery from a severe brain injury, the main character struggles through multiple concurrent "realities" to sort out what's really going on, along the way touching on many possible implications the growing use of advancing AI could have on human society.
Unfortunately, for me at least, multiple shortcomings of the implementation outweigh the interestingness of these motivating ideas.
The narration throughout felt clumsy and often repetative as might a quick middle-school first draft. Frequent scene contradictions stuck out as sloppy and distracting—a character walks the sidewalk in a new-to-him city alongside a five-story building while describing details about its flat rooftop; a character follows a path around the city perimeter until stopping in the city's center—like cognitive speedbumps along the way, also suggesting the author hasn't really thought out the details of the story.
One could assume that only a generative AI model would make such inconsistency mistakes and thus the clumsier parts must've been among those written by the co-authoring AI, but that seems a weak excuse to me. I've read poorly written books with these same sorts of issues and no AI involvement, and if the AI used isn't ready to produce narration that isn't riddled with these kinds of errors, then it isn't ready to produce this sort of a book.
Meanwhile, various scenes along the way add commentary on possible effects increasing use of advanced AI might have on various elements of life, from policing to courts to fiction and performance art and more, perhaps too many to try to address in one story without feeling forced. Assorted other concepts are also introduced as AI-assisted skills to level-up the main character, such as emotional intelligence and pattern recognition as predictive and manipulative tools, and something along the lines of Zen acceptance. However, all of these elements are hugely simplistic, almost cartoonish characatures of themselves, as though written by someone who's heard of these concepts but doesn't really understand any of them.
The book's description asks if the reader can separate human-written from AI-written. Honestly, I lost interest in the question because it was all equally poorly written.
Finally, the story just ends on a sort of step forward but no real resolution. I suspect it may have been going for an open-but-dismal 1984-like ending, but if so the pieces just don't hold up strongly enough to deliver it.
Overall, I think the author has some interesting ideas to ponder, but for multiple reasons had great difficultly weaving together a coherent story that illustrates them all in any meaningful way. The result just doesn't work for me.
To be fair, the book's copyright page calls it a "beta edition", so perhaps it has some awareness of the need for further iterative training of both authors' learning and generative models?
Unfortunately, for me at least, multiple shortcomings of the implementation outweigh the interestingness of these motivating ideas.
The narration throughout felt clumsy and often repetative as might a quick middle-school first draft. Frequent scene contradictions stuck out as sloppy and distracting—a character walks the sidewalk in a new-to-him city alongside a five-story building while describing details about its flat rooftop; a character follows a path around the city perimeter until stopping in the city's center—like cognitive speedbumps along the way, also suggesting the author hasn't really thought out the details of the story.
One could assume that only a generative AI model would make such inconsistency mistakes and thus the clumsier parts must've been among those written by the co-authoring AI, but that seems a weak excuse to me. I've read poorly written books with these same sorts of issues and no AI involvement, and if the AI used isn't ready to produce narration that isn't riddled with these kinds of errors, then it isn't ready to produce this sort of a book.
Meanwhile, various scenes along the way add commentary on possible effects increasing use of advanced AI might have on various elements of life, from policing to courts to fiction and performance art and more, perhaps too many to try to address in one story without feeling forced. Assorted other concepts are also introduced as AI-assisted skills to level-up the main character, such as emotional intelligence and pattern recognition as predictive and manipulative tools, and something along the lines of Zen acceptance. However, all of these elements are hugely simplistic, almost cartoonish characatures of themselves, as though written by someone who's heard of these concepts but doesn't really understand any of them.
The book's description asks if the reader can separate human-written from AI-written. Honestly, I lost interest in the question because it was all equally poorly written.
Finally, the story just ends on a sort of step forward but no real resolution. I suspect it may have been going for an open-but-dismal 1984-like ending, but if so the pieces just don't hold up strongly enough to deliver it.
Overall, I think the author has some interesting ideas to ponder, but for multiple reasons had great difficultly weaving together a coherent story that illustrates them all in any meaningful way. The result just doesn't work for me.
To be fair, the book's copyright page calls it a "beta edition", so perhaps it has some awareness of the need for further iterative training of both authors' learning and generative models?
Content Warnings
Graphic: Confinement
Moderate: Alcohol, Medical content, and Schizophrenia/Psychosis
Minor: Death, Blood, Vomit, and Sexual content