This is a book about taming entropy. About the toll it takes — and what doors it opens.
Tamers of Entropy is an ambitious Lunarpunk fiction project written by Juraj Bednar and the main character of the book (neura)Lisa.
Book 1 begins at a breakneck pace with two hacker siblings in New York City. Kevin and Lisa gain access to thousands of WiFi routers in Manhattan, but things get dangerous when they discover that someone has hacked the routers long before them, and they are listening.
Kevin goes on the run, giving up his computer science studies at NYU, and Lisa foils a capable FBI agent who later becomes a love interest.
The story progresses over several years, and Kevin moves to a remote location in Asia where he finds a partner and begins to build private digital infrastructure for the local community in exchange for peace and tranquility.
Lisa gets deeper into the hacker scene and has a revelation at a hacker conference in Germany, where she discovers the powers of neural transmission and synchronization.
Years later, she teams up with Kevin and his partner to fulfil the potential of her capabilities and avoid the forces that seek to destroy her project.
The fast-paced plot, time jumps, and meaningful twist and message at the end kept me on the edge of my reading seat. I found myself raring to dive back into the world of lunarpunk tech, New York streets, and Laotian jungle.
Both siblings were well-drawn characters, with very real flaws, needs, and desires. The story bounces between them and James, the FBI agent who becomes obsessed with bringing them to justice. I sometimes wished for more scenes linking the larger time gaps, as the multi-character perspective and wider themes of how humans and tech intertwine felt more condensed than many expansive Sci-Fi series.
With Bednar’s background as a freedom maximalist and privacy tech expert, it was easy (enjoyable even) to follow the technical explanations of hardware, software, and interplay between AI and organic matter.
“Not many people know that I have aphantasia — I cannot visualize images in my mind. When I close my eyes and try to picture something, there is simply nothing there. No mental imagery at all. I always thought that the "imagine..." was metaphorical. Yes, I don't even have visual dreams.” Juraj Bednar.
This led the author down two interesting paths:
- Adding multimedia elements to the project — curated designs, packaging the book as ‘evidence’, soundtrack playlists, and creating social profiles for the characters on Nostr.
- Liberal use of LLMs to write descriptions and other visual elements of the book.
While point 1 adds to the scope of the project and reading experience, the second point is most interesting to unpack. Bednar openly admits that AI wrote swathes of the text which appears on the page, even weaving it into the story. Lisa, the savant sibling who creates neural feedback loops with AI, is revealed to be the author of the text (via memory transfer), pointing to why the text feels generated in places.
This is the first piece of ‘AI fiction’ I’ve ever read, and I have mixed feelings. I understand that without the help of AI, the author would not have been able to put this story out into the world. It is undoubtedly his story, not one ‘created’ by simple prompts. The technical descriptions, story arc, and character development are human-led.
However, the repetitions of tropes, typical LLM phrasing, the lack of literary devices, and the simplistic ‘telling’ style of the prose did undermine my enjoyment of the story.
Overall, I found it an engaging and worthwhile read; Tamers of Entropy turned out to be an interesting exercise which prompted thought about how, indeed, we do interact with technology.
Find out more and get the book at tamersofentropy.net
I'm glad you reviewed this. I was considering reading it, but I think that I won't. Not that your review was negative, but I do feel less and less patience with llm generated text. In my fiction reading at least, I think I will avoid it.
I saw a post by Marty Bent about how he's using an agent to write his newsletter, of course "Nothing gets published unless I approve it."
I'm feeling somewhat down about the state of writing these days. I'm worried we are all just going to get overwhelmed by the slop and decide that it's easier to accept it than to fight it.
That is a very real threat.
For many people it happens overnight. AI makes writing 'easy'. Even for professionals, The creep happens little by little, mostly due to clients often having AO psychosis.
I see it happening in design and apps too.
Small communities, web of trust, and legacy gatekeepers are all important to help us find quality.
No moar slooooooop
I wouldn't classify this book as slop. The concept, world, tech details and plot had a lot of real input, and the story has a meta-AI angle too.
Still, I'd say for prose, generated text is a big turn off for me.
I feel the same way. It seems to be a hopeless battle. We may get to the point where the "art" lies in creative prompting.
I will retreat to my cave and write there (even if it is on the walls)
In my most generous moments, I wonder if the illuminators of medieval texts felt that the printing press took all the artistry out of writing. And surely writing a novel was a very different thing in the 19th century than it has been in the 21st century (editing via word processor vs editing handwritten texts).
But I still don't like the llm stuff. There is no thrill quite like reading something new that resonates, ringing in your mind like a clapper in a bell. If it's just luddism that I don't think llms can do this, I suppose I will have to change my mind. But so far, the luddism is winning.
https://twiiit.com/MartyBent/status/2077837361548005443
ugh, yikes