rashbre central: Sheep Dreams, Ed Adams - Review

Monday, 3 November 2025

Sheep Dreams, Ed Adams - Review

 


Dr. Tessa McCrea, Senior Neuro-Integration Engineer

(Excerpt from post-release memo circulated via NeuralNet Research Slack, Channel #literary-detritus)


“Thirty suspended cortices. Cross-wired through a lattice. Each one thinks it’s alive.”


We passed that quote around the lab this morning. Nobody laughed.


Adams’ Sheep Dreams describes the XTend rack with unnerving fidelity: thirty mammalian cortices in solution, looped through a bidirectional digital spine. The description of the vagal lattice interface — the twelve cranial channels mapped to data trunks — is so close to our V-Net design notes that legal asked whether the book used leaked documentation. It didn’t. He just guessed right.


In our prototypes, the biological component acts as a chaos reservoir, feeding stochastic variance into the silicon model to prevent pattern ossification. Adams turns this into theology. His sheep aren’t machines — they’re compliance made flesh. They obey because that’s what they were bred for.


“The horror is not that the sheep think; it’s that they obey better than we do.”


That line triggered a brief ethics review on Slack. We spend our days trying to make thought predictable; he’s written a warning about what happens if we succeed.

The novel’s RightMind network, its dream-induction sequence, the “mint-tang” plasma — all plausible. Even the tone is right: that strange calm we hear in our test subjects just before interface sync. Adams gets the effect correct — the silence after consent.


From a purely technical standpoint, the XTend description is the best fictional account of organic-silicon co-processing I’ve read. From an emotional standpoint, it’s too accurate.


Our takeaway: if Sheep Dreams reaches the public before our next announcement, the optics will be awkward.


But it’s also a reminder that the culture already understands what we’re building — maybe better than we do.

— Dr. Tessa McCrea

NeuralNet Cognitive Systems Division

No comments: