Keyboard shortcuts

Press or to navigate between chapters

Press S or / to search in the book

Press ? to show this help

Press Esc to hide this help

External review, 2026-09-15

An external review dated 2026-09-15 examined the spec and the first implementation plan against the dataset, the Shiu et al. code, and the cflib source. Its findings and their dispositions follow. The section numbers in the disposition column refer to the design specification, docs/superpowers/specs/2026-09-15-flatline-imago-design.md, not to chapters of this book.

FindingDisposition
Signed LIF with histaminergic photoreceptors silences the laminaAccepted; section 8.6 policy and probe
LIF differs from Shiu et al. (g reset, refractory freeze, strict threshold)Accepted after checking the released code; section 7.1
Hover curriculum solvable by a zero decoderAccepted; section 12.3
No optic-flow diagnosticAccepted; section 8.6 probe
10 Hz camera far below motion-vision bandwidthAccepted as risk and measurement; section 8.5
Haltere round-robin discards type structureAccepted; grouping by type; section 8.4
flight-core mask removes 0.2% of neuronsAccepted; replaced by no-optic-lobe, masks are not a real-time strategy; section 7.6
Training compute unbudgetedAccepted; section 19
Safety envelope inadequate for an open workshopAccepted; section 10.5
Missing success criteria and open questionsAccepted; section 17
Connectome contract errors (unclear label, eight scales, 1,314 descending, side rule, hex target types)Accepted; sections 6.3, 6.5, 6.6
Decision log lacks alternatives and reasonsAccepted; section 4
Spec and plan file names divergeAccepted; section 13.1
Command-axis signs half-specifiedAccepted after checking cflib; section 10.4
Clock, threads, determinism underspecifiedAccepted; sections 7.8 and 10.1
Shuffle does not preserve in-degreeAccepted; the shuffle permutes partners within synapse-count classes, section 6.3
Retina normalization scaled every photoreceptor edgeAccepted; section 8.3
Photoreceptor column should be the strongest single targetRejected; strongest column by summed weight retained with the reason in section 4
Input neurons not bypassed and edges onto them keptAccepted; sections 6.3 and 7.2
Full edge table loaded whole for placementAccepted; plan filters by photoreceptor before materializing
Event propagator loops per brain on MetalAccepted; packed batch; section 7.3
Config and checkpoint schemas missingDeferred to the sub-project plans that create them

The full review is filed at docs/reviews/2026-09-15-grok-review.md in the repository.