Pranav Bhave

Noetic log 002 · shipping record + rating · 2026-08-23 · lookback due 2026-09-23

The observatory ships, and gets graded

Version 0.4 went live today. This entry records what shipped, rates it without anesthesia, and signs a lookback contract: on 2026-09-23 this page's questions get answered against evidence, and the most unflattering possible conclusion is pre-registered below so it cannot be dodged later.

What shipped

The record became an observatory: a thirteen-claim registry (schema v0.3) rendered three ways (ledger, capsule field, six modules under one grammar), an interactive feasible-worlds instrument whose geometry is asserted in CI to 1e-9, clean-clone reproduction of bounds and endpoint witnesses, and a new gate that clones every bound repository to prove each pinned commit is reachable from its default branch — because one binding was found stranded in rewritten history, and byte-watching triggers were structurally blind to it. An independent hostile review filed twelve findings; all twelve were fixed before shipping, one of them a factually false sentence in the module that lectures about expiry. Log 001 carries the lineage of this failure class.

The rating — owner-judged, adversarially informed, no anesthesia

Scores are the owner's judgments after the hostile review, not measurements. Each row names why the score is not higher, which is the only interesting column.

DimensionScoreWhy not higher
Mathematical faithfulness9.5 / 10 Every displayed number is machine-checked or commit-bound. Docked: the only arithmetic on the site is a two-guardrail toy — flawless, and small.
Evidence realism9 / 10 Bindings, reproduction, reachability, freshness are all real and enforced. Docked: the strongest empirical fact on display is an absence — E2 is frozen and untested. The governance outruns the science it governs.
Epistemic severity9 / 10 Thirteen envelopes, non-claims walls, planned pages that claim nothing. Docked: severity held for numbers and slipped in narration — the P0 finding was prose inflating the machinery's own track record, caught by a hired adversary, not by the machinery.
Semantic visual language8.5 / 10 Cyan/amber/red now mean one thing each, and green means nothing. Docked: the color law shipped violated in three places and needed the hostile review to become coherent; light-mode amber and gold remain confusable, mitigated only by the never-color-alone rule.
Visual hierarchy & elegance8 / 10 The type system carries real authority. Docked: the front page was restructured more than shortened; six nav routes, three doors, and two strips compete for the same first click; the six module pages are template-uniform to the edge of monotony.
Accessibility7 / 10 (provisional) Structure is right — skip links, SVG descriptions that carry the argument, reduced-motion neutrality, no hover-only information, computed AA contrast. Docked hard: no keyboard-only traversal and no screen-reader pass has been run. Computed accessibility is a hypothesis about lived accessibility.
Cold-viewer comprehension6.5 / 10 The instrument teaches by dragging, which is the site's best pedagogy. Docked: the vocabulary wall is real — "partial identification," "witnesses," "envelopes" above the fold — and the central hypothesis that any of this grammar helps a reader is untested with the null alive. A 90-second visitor may leave impressed and uninformed, which is this site's named failure mode.
Research threat surface5 / 10 The brutal row. What the site proves today: discipline about a toy example, one synthetic study, and its own governance. Zero real guardrails measured. A hostile reader's fair summary — "an exoskeleton that has not yet lifted anything real" — currently stands, and nothing on this page rebuts it. Only E2 data can.
The brutal line

Instrument-grade scaffolding around a research program that has not yet produced its first empirical result. The scaffolding is genuinely rare. It is still scaffolding, and thirty days from now the only question that matters is whether something real got lifted.

What would make this instrument dangerous

Dangerous in the only sense this record endorses: menace pointed at claims, never at people. The site's machinery — envelopes, bounds, witnesses, expiry — currently prosecutes only its own claims. The move that turns it from a portfolio into a threat to overclaiming is to point it outward: take published AI-safety numbers from the wild, compute what their stated evidence actually licenses under the Fréchet bounds, and publish the interval next to the headline number with witnesses at both ends — a claim autopsy, powered by exactly the E2-style measurement this program has frozen but not yet run. The standard, made visible, applied where it stings. That requires E2 to exist first, which is why the lookback below has teeth.

The lookback contract — due 2026-09-23

On or after that date, answer each question against evidence that exists outside this page. Anyone can run this checklist; none of it requires trusting the owner.

  1. Did the standing gate stand? The weekly verify runs (five expected by then) are public in the Actions history. Any red run, and what fired.
  2. Did E2 move? Either a conforming observation set exists, or a documented failure-to-obtain exists. If neither, the "next artifact" promise on /now/ is thirty days stale and this entry says so.
  3. Were the open QA items closed? Keyboard-only and screen-reader passes were disclosed as not-run at ship time. After thirty days, an open disclosure becomes a broken promise; the accessibility score above drops to 5 until they run.
  4. Did any external falsifier arrive? Corrections, issues, or mail against any claim — and what the registry did about them.
  5. Do any decay clocks show amber? None should before December; amber earlier means a review was triggered, and the ledger must show the re-review.
  6. Did the claim count move for the right reason? New claims should arrive because evidence arrived — not because surfaces multiplied.

Pre-registered conclusion if nothing moved: the observatory spent its first month documenting stasis beautifully — an instrument without an experiment — and the research-threat score falls from 5 to 3. That sentence is written now, while it can still be falsified by work.

Non-claims — the scores above are owner-attested judgments, not measurements; no reader study exists and none is implied. This entry is hand-written and carries no evidence markers; the linked gates and Actions history are the checkable parts.