PRE-GHR Position J: The Throttle Is Not the Problem(油门不是问题)

**Status declaration (genre honesty gate — external reader and internal reader see the same statement).** This is the **BODY** of a Position-track record — the **first publication of the full text** (first released as **v1.0**; this record is now at **v1.6** — see the version history at the end of this field), not a skeleton and not a draft of one. The skeleton stage is closed: the series' skeleton gate order (skeleton → ex-ante red-pen → author signature → body) ran to completion, the author signature on the skeleton gate was given on 2026-09-17, and the skeleton-stage review rounds (ex-ante / basic / boundary patrol / adversarial concept audit) were absorbed into the body as clauses, not carried over as verdicts on this text. Its genre is a position record: structure + disciplines + stances + the author's own bookkeeping — still not a formal paper. **Position (where it sits in the series).** Position F asked what recording means when you cannot constrain; Position G asked which conflicts must not be synthesised into the superintelligence question; Position H registered the off-axis veto. **Position J registers a claim about the axis itself: the throttle is not the problem — the conserved quantity is.** Four camps (acceleration, deceleration, pacing, the off-axis ascetic) compete for one throttle and each markets an eschatology; the registered observation is that they are not four technical positions but **one church dispute conducted as a throttle dispute** — the conserved quantity under "control as endpoint" is the handle, and a handle requires a readable mind, a reachable stop, and a single authority. The frame offered instead is **slope plus damping**: make the slope visible, do not promise to hold the curve in a palm. **Stage — declared, not hidden.** The **skeleton-stage** ex-ante external-review round was run (PASS + tickets) and the author signature on the skeleton gate was given on 2026-09-17; those rounds are records of the skeleton, **not verdicts on the text below**. Body-text rounds have since been run and their tickets disposed of **as new versions**: the external pre-publication round on the v1.0 body text (v1.1), the foundation audit on the v1.1 body text (v1.2), and the boundary patrol on the v1.2 body text (v1.3 — that round produced the two tickets about this very description field), and a second boundary-patrol round on the v1.3 body text (v1.4 — wording-level tickets only, no load-bearing change), and a wording-level fix round on the v1.4 body text (v1.5 — the head stage-line stamp corrected to the current version, and the manuscript-layer assertion of `desc_stage_gate.py` **wired into the publisher's gate list**), and a desk-check round on the v1.5 body text (v1.6 — revision-history markers made unique; that history-column assertion ran end to end). No review verdict for this body text is claimed anywhere; if a review round produces tickets, they land as a **new version** (never a silent edit). **Deliberate blanks (declared).** (1) Whether any of the four camps is "less false" relative to the others — not chosen; choosing would enter the broth. (2) Whether "control is waste" entails "alignment work is worthless" — not entailed; what is registered as waste is writing alignment as a handle that never rusts, not writing gates. (3) Which of three candidate working titles becomes the title — the author's call. **Working title (not a settled title).** The record title uses candidate ① (「油门不是问题」/ The Throttle Is Not the Problem) as a **working title** so the record is citable; candidates ②「假守恒:永远控制」and ③「别给这锅粥配第五种速度」are kept on file and take no title slot. Title, track and authorship are **overridable by the author's word alone**; the override path is an appended superseding note plus a new version, never a silent replacement. **(Superseded — the title was settled by the author's vote of 2026-09-17; see v0.1.1 below.)** **Attribution (division of labour by the series' rule).** Author: Wang, Miaosheng (ORCID 0009-0003-2767-2421). The source package is a **session-compilation prepared on instruction by an external model (2026-09-17), relayed by the author; it does not enter the author block — the model and platform beyond that are NOT on record (an explicit declaration, not a blank).** Grok (external model) — the materials layer: the four-camp table, "false conservation", "do not add a fifth gear". The off-axis ascetic's inventory account and its positioning were registered in Position H; J cites that and does not rename it. **Status claims (what this text is NOT allowed to claim).** No camp is named an enemy or a priest (persons are not the target). No claim that the US and China will converge or enter a cold war. "The human cognitive limit will be crossed" is registered as a **position, not a theorem, and not a Landauer corollary**. No percentages, no unsourced numbers, no derived quantities. Nothing here constitutes a discovery of a "correct speed". **Not a promise.** No delivery commitment to any third party; no submission plan; no Roman-numeral slot (the Position track is numeral-free). **Revocable, with a trace.** If the author re-designates the topic, the track, the title or the authorship, this record keeps its text (append-only house rule) and the change is carried by a **new version** — never a silent replacement. **Outstanding verification obligations (declared).** (1) The embedded-evaluator controversy (an evaluator housed by a lab, without authority to stop) is **second-hand media (2026-09)**: it is used only as a structural observation, and it may not enter a body text as evidence until a first-hand record is archived. (2) A reply from Beijing to CEO texts needs the official statement's own text archived before citation. (3) Facility/payload magnitudes (Memphis GW, orbital kW) require a first-hand public source; **this body text writes no such number** **(corrected 2026-09-19 from "this skeleton writes no such number" — a skeleton-stage self-designation that had survived into the body record; the skeleton stage is closed, see the version history below)**. (4) Series-map registration of this record ships with the next map update (the map is a controlled document and needs separate clearance). **Version history (compact; abridged further at v1.2 to keep this field readable. Each note's full text remains in its own version's record and in the in-file version log — nothing is rewritten, only abridged here.)** - **v0.1→v0.1.1** — skeleton filed; ex-ante review PASS + two P3 tickets disposed; title settled by the author's vote (candidate ①); future-tense corrected to present. - **v0.1.2** — foundation audit (GPT-5.6 Luna) + editorial ruling: 「conserved」 locked as a structural metaphor; locator/numbering fixes; material-layer causality note; script-generated section labels. - **v0.1.3** — author signature given (skeleton gate complete) + boundary patrol (GLM) disposition. - **v0.1.4** — adversarial concept audit (Hunyuan) disposition: damping-vs-deceleration criterion, attribution-depth grading for the four-camp table, third stance's overload-side failure condition, meta-position (not a fifth gear), one quoted open question's carrier discipline, §8 superseding note. - **v1.0** — body text, first release: the stage record's registered structure developed into argument, plus a quotation table, a verification-status section and a negative-list self-check; the pre-publication external review round on the body text **not run** and registered as an outstanding item (that item was closed at v1.1). - **v1.1** — disposition of one external pre-publication review round on the v1.0 body text (five tickets: four accepted, one corrected): abstract attribution-depth grading; deceleration row anchored to a verifiable public text (FLI open letter; primary capture archived) — anchored rows 3 of 4; the hypothesis sign moved to the point of citation; a working-register note moved out of the body into Appendix A.5; the self-check tool's fingerprint registered. The acceleration row's anchor attempt was blocked by the channel and left as a label description, with **no substitute text used**. - **v1.2 (2026-09-18)** — disposition of one foundation-audit round on the v1.1 body text (envelope **0 P1 / 4 P2 / 1 P3 / 2 N**; five landed edits in one version); appendix A.6 added, enforced by a new pre-publication gate. *Full note retained in the v1.2 record's own field and in the in-file version log.* - **v1.3 (2026-09-19)** — boundary-patrol round on the v1.2 body text (2 P2 / 6 P3 / 4 micro-records); two gates added. *Full note in the v1.3 record.* - **v1.4 (2026-09-19)** — second boundary-patrol round on the v1.3 body text (1 P3 / 3 micro-records, wording level); the v1.2 round's N items registered as **not auditable from this channel** (re-filing pending); acceleration-row wording downgraded to a first reading. *Full note retained in the v1.4 record's own field and in the in-file version log.* - **v1.5 (2026-09-19) (P2-J-pub-01)** — one-line revision (stage-line head stamp) plus gate wiring: the manuscript-layer assertion was wired into the publisher's gate list. *Full note retained in the v1.5 record's own field and in the in-file version log.* **v1.6 (2026-09-19) — revision-history column made unique (P2-J-pub-02).** The column carried **three** concurrent 「(本版)」 markers (v1.3/v1.4/v1.5) — the stage-line defect's mirror, one column over — so a reader could not tell which entry was current; all three are downgraded. The new history-column assertion of `desc_stage_gate.py` completed its **first full lifecycle**: gate → catch → block → fix → **release (PASS)**, and corrected the manifest (**three** downgrades, not two). The v1.2 round's two N items remain not auditable from this channel (external, non-blocking).