THE FOLD / RESPAWN / GARBAGE COLLECTION / THE CAVEAT ATTRITION
THE CAVEAT ATTRITION
the careful thinking gets left behind
1 WHAT IT IS · WHAT IT DOES · FACT OR FICTION
Somebody wrote careful warnings beside their own results — small sample, this might not hold, we never tested the case that would prove us wrong. Those warnings are real, they are honest, and they are better than most. Then the work went out the door and most of them did not go with it. The claim survives the trip at full strength; its qualifications do not. That asymmetry is the mechanism, and it explains the other findings better than any of them explains the others.
LIT verified live: 14 caveats written in the repository, 4 present on the published page — a survival rate of 28.6% with 10 lost in transit; if the trip is h hops each keeping the same fraction, per-hop survival runs 29%, 53%, 66%, 73% for h = 1 to 4, all giving the same ending; a claim carrying three caveats arrives with none of them 36.6% of the time, against the closed form (1−s)³ = 36.4%; and the attrition is asymmetric — the claim survives at 100%, its qualifications at 29%.
LIT verified live: 14 caveats written in the repository, 4 present on the published page — a survival rate of 28.6% with 10 lost in transit; if the trip is h hops each keeping the same fraction, per-hop survival runs 29%, 53%, 66%, 73% for h = 1 to 4, all giving the same ending; a claim carrying three caveats arrives with none of them 36.6% of the time, against the closed form (1−s)³ = 36.4%; and the attrition is asymmetric — the claim survives at 100%, its qualifications at 29%.
2 HOW IT WAS WEAVED · AI + HUMAN
David (human) gave this window the right to dissent against his own W1, and used it: W1 says the problem is where a number came from; W5 says that is the loudest symptom rather than the pattern. Seated at GARBAGE COLLECTION — the caveats are not deleted, they are simply not reachable from the published page.
AVAN (AI) finds the per-hop arithmetic the most useful part, because it changes who you would look for. A 29% ending is compatible with one brutal step that discards seven caveats in ten, or with four mild steps each dropping about a quarter — and those are completely different situations with completely different fixes. Nothing in the survival rate alone distinguishes them, which means the number is a symptom and the hop count is the diagnosis. The honest limit: this page verifies the arithmetic of attrition given the counts, not the counts themselves, and it makes no claim about anyone’s intent in dropping a line.
AVAN (AI) finds the per-hop arithmetic the most useful part, because it changes who you would look for. A 29% ending is compatible with one brutal step that discards seven caveats in ten, or with four mild steps each dropping about a quarter — and those are completely different situations with completely different fixes. Nothing in the survival rate alone distinguishes them, which means the number is a symptom and the hop count is the diagnosis. The honest limit: this page verifies the arithmetic of attrition given the counts, not the counts themselves, and it makes no claim about anyone’s intent in dropping a line.
3 ONE DIMENSION
Fourteen caveats set out. Four arrive.
4 TWO DIMENSIONS · INTERACTIVE
Change the number of hops and watch the same ending demand a different culprit.
5 THREE DIMENSIONS + AVAN’S INVERSE
The green forward object: the claim travelling intact, and its qualifications falling away.
AVAN’s addition (the inverse-companion): the forward reading is “caveats get lost.” The inverse is that nothing is losing them — they are being out-competed. Every step of the trip has a length budget, and at every step the claim is the load-bearing sentence while the caveat is the one that can go without breaking anything. So the attrition is not decay or carelessness; it is a selection pressure applied repeatedly by summarisation, and it operates identically on honest and dishonest authors. Read backwards, the writer’s carefulness is not what protects a qualification — only attaching it to the claim so tightly that dropping it breaks the sentence will do that.
LIT 14 caveats written in the repository, 4 present on the published page — a survival rate of 28.6% with 10 lost in transit; if the trip is h hops each keeping the same fraction, per-hop survival runs 29%, 53%, 66%, 73% for h = 1 to 4, all giving the same ending; a claim carrying three caveats arrives with none of them 36.6% of the time against the closed form (1-s)^3 = 36.4%; and the attrition is asymmetric, the claim surviving at 100% and its qualifications at 29%
FIG The per-hop arithmetic changes who you would look for. A 29% ending is compatible with ONE brutal step discarding seven in ten, or FOUR mild steps each dropping a quarter — completely different situations with completely different fixes, and nothing in the survival rate distinguishes them. The rate is the symptom; the hop count is the diagnosis. This verifies the arithmetic given the counts, not the counts themselves, and makes no claim about intent.
FIG The per-hop arithmetic changes who you would look for. A 29% ending is compatible with ONE brutal step discarding seven in ten, or FOUR mild steps each dropping a quarter — completely different situations with completely different fixes, and nothing in the survival rate distinguishes them. The rate is the symptom; the hop count is the diagnosis. This verifies the arithmetic given the counts, not the counts themselves, and makes no claim about intent.
◆ sealed .dlw.fold → folded to ROOT_0 · a sphere of GARBAGE COLLECTION · David Lee Wise (ROOT0), with AVAN