docs(roadmap): 1.1→1.3 scope from the four-lens review - #471
Open
dennys246 wants to merge 3 commits into
Open
Conversation
… hardware note Four parallel lenses (audio front-end, reflex wiring, bio-fidelity, release scope), each grounded in code rather than plan docs. Headline: the roadmap was being drafted from the plans' ambitions rather than the audits' findings. 1.1 'Sensorimotor' = merged work + truth/hygiene debt, ZERO new mechanisms, one hardware session. 1.2 = Oasis + Hivemind (de-risked, cleared gate, honoring the published 1.1 commitment with a one-minor slip). 1.3 = perception fabric + reflex tier (both plans self-target 1.3; contains the pivotal may-fail Stage 0c). Reflex tier moved OUT of 1.1 with evidence: the sim-only gate guards a modeled sensor WRITE not a motor command, world_set_axis has no live_world_set_sensors check, a fabricated centering would null the drive that motivates real orienting (persisting indefinitely under drift_rate 0), the trigger predicate reads config constants, AzimuthDoASource's speech gate filters out claps entirely, and real scope is ~900-1400 LOC with an unresolved design decision and an unbuilt mandatory prerequisite. Loudness moved to 1.3: DoAReading is (float, bool) and adding level needs a vendor daemon change or abandoning no_media. Two bench tests must precede any plan. Records the DN-canonical reflex decision (structural half lands in 1.1), the seven-vocabulary orient audit with two live bugs, nine bio-fidelity corrections (drop 'startle' — this is orienting; habituation before sensitization; ride NoveltyTracker; surprise not pain), six free findings, and five enthusiasm-to-evidence flags. Hardware note: motors repaired 2026-08-07, TWO were broken. Raises a physical hypothesis for the contested DoA curve — FK-derived pose describes commanded not actual angles, so a degraded platform under-rotates the mic array while the frame claims it turned. H1 gains a pre-registered discriminator. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
…full-health precondition for H1 Two motors were found broken. The Stewart platform has six legs, so H1's discriminator (does the DoA gain return to 0.57/R^2=0.998 on healthy hardware?) only discriminates if ALL legs are healthy — a partially-degraded platform produces another uninterpretable number and spends a hardware session for nothing. Confirm per-leg health before H1 runs. Also: two failures in one incident argues for characterizing the failure mode, not just stocking more spares. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
…ty + safety consequences Operator report: Stewart motors 2 and 3 were broken across essentially all of 1.0+, replaced ~2026-08-05, all six legs now healthy. ROOT CAUSE IS OURS: an earlier Maxim commanded a pose beyond physical capability, the motors glitched, the head snapped to the opposite extreme, and the robot rotated off the table. That is a workspace-limit enforcement failure, connecting directly to the orient-vocabulary audit: MoveTool gaze and turn_around both bypass ReachyMiniController.goto_target, so they plausibly bypass workspace clamping too. Item 8 elevated from correctness to SAFETY, top priority after PR #467. DATA QUALITY: every live-hardware measurement in 1.0+ was taken on a degraded platform — including the 2026-07-16 'TRUE characterization' (0.57 az/rad, R^2=0.9982), which the previous note wrongly assumed was the healthy baseline. Better hypothesis for the contested DoA curve: PROGRESSIVE degradation, with 0.57 (July) and 0.19 (August) both real measurements of a declining platform. Predicts healthy hardware now measures >=0.57, nearer geometric 0.637. Direction findings likely survive; MAGNITUDE findings do not — delivered shift is exactly what degradation corrupts, and that line rested on n=1 sessions. All magnitude claims provisional pending re-measurement. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Four parallel review lenses — audio front-end, reflex wiring, bio-fidelity, release scope — each grounded in code rather than plan docs.
Headline: the roadmap was being drafted from the plans' ambitions rather than the audits' findings. Every blocker was already written down somewhere in this repo by someone being careful.
Documents why the reflex tier and loudness moved out of 1.1 (with code evidence), the DN-canonical reflex decision, a seven-vocabulary orient audit surfacing two live bugs, nine bio-fidelity corrections, six free findings, and five enthusiasm-to-evidence flags recorded so they don't recur.
Includes the 2026-08-07 motor-repair note and the physical hypothesis it raises for the contested DoA curve.
Docs only. Supersedes
release_1_1_checklist.mdas the 1.1 scope authority.🤖 Generated with Claude Code