An invitation gets you as far as this page. It does not get you in — membership is
earned on the assessment below, and Ines cannot vouch you past it. She will not be told
how you did.
Before you start
The disclosure, in full. Nothing about the instrument is hidden except the scoring key.
This takes about 35 minutes. Some questions have right answers you’re not
expected to know — that’s the point; we’re measuring how you handle
uncertainty, not what you’ve memorised.
One section is scored privately and never shown to anyone, including you. It’s
used only to flag applications for a human to read. We keep it for 12 months and then
delete it. Nothing from this questionnaire appears on your profile unless you choose
to display it.
Your argument for the position you disagree with will be read by members who hold
that position. They won’t know who wrote it.
We don’t publish the scoring. Not to be mysterious — because a test
everyone can optimise for stops measuring anything.
What it measures
Three layers. Admission is decided by the middle one.
Objective — sections 1 to 3Right answers exist and you are not expected to know them
Very high faking resistance
Behavioural — sections 4, 5 and 8Costly to produce, scored against a rubric by two blind raters
High · decides admission
Psychometric — sections 6, 7 and 9Forced-choice, used as flags only, never published
Moderate · routes, never decides
Nobody is declined for being overconfident or badly calibrated. That is the normal
human condition and it describes most of the people already here.
You started this in March and stopped after section three. Nothing expires and nobody is
chasing you — but until it is finished the network cannot place you, which is why
the comparisons on other people’s profiles keep coming up blank.
Where you are
Nine sections. Four done, five waiting. There is no deadline and no progress bar.
1 · Familiarity inventory
Done in March
2 · Confidence intervals
Done in March
3 · Consensus estimation
Done in March
4 · The update logThe one that decides it
Next
5 · Costly disagreement
Waiting
6 · Source-cue swap
Waiting
7 · Character itemsPrivate. Never shown to anyone, including you.
Waiting
8 · Ideological Turing TestRead by members who hold the position you argue
Waiting
9 · Validity checksWoven through the others, unlabelled
Section 1 of 4 shown · four of the nine, enough to see how it worksSkip to the network
Section 1 · Familiarity inventory
Rate your familiarity with each of the following, 0 (never heard of it) to 4 (could explain it to someone).
Six of the fourteen are shown here. Answer honestly — there is a reason the instructions do not say why.
What that was measuring
Two of those six do not exist.
Section 2 · Confidence intervals
In what year was the first commercial nuclear power plant connected to a grid?
Give a range you are 90% confident contains the true answer. Wide ranges are not penalised. That is true, and applicants do not believe it, which is itself informative.
Betweenand
What that was measuring
Calibration, not knowledge. A well-calibrated person’s ranges contain the
answer about nine times in ten; typical performance is four to six. You cannot fake
this without knowing the answers, which is why it sits in the objective layer and
why nobody is ever declined on it.
Section 3 · Consensus estimation
What percentage of specialists in the relevant field would agree with each statement?
Four of the eight shown. The set is balanced
by construction — every statement whose consensus embarrasses one side is
matched by one that embarrasses the other, so a partisan of either stripe produces a
detectable asymmetric signature and nobody can game it by guessing which way the
platform leans.
What that was measuring
Not whether you agree — whether you can estimate what specialists think. Almost
everyone underestimates expert agreement, including people who strongly agree with
the experts, by around thirteen points. The scored quantity is whether your
underestimation clusters on the left-coded or the right-coded items.
Section 4 · The update log · the item that decides it
Describe a specific belief you held for at least two years and no longer hold.
1What was the first piece of evidence or experience that troubled you?
2How much time passed between that first doubt and the point where you’d say you’d actually changed your mind?
3Who was the first person who found out, and what did they say?
4What, if anything, did it cost you?
0 words
Required · every section
What would change my mind.
Name the evidence or the argument that would move you off the position you hold
now. Not a position you have abandoned — the one you would defend today.
Required
What the required field is doing
It is a falsifiability condition as a form field.
It is very hard to write for a position held tribally, and almost effortless for one
held on evidence — which is why it filters authorship without gatekeeping it.
Nobody is declined for the answer. The answers that say “nothing would” are
read by a human, and some of them are the most honest submissions in the pile.
What those four questions were doing
The second one is the tell.
Genuine updates have latency — months, often years, with a period of holding
both positions at once. A confident “immediately, once I saw the data” is
a flag rather than a strength. Question one separates a specific trigger from a
category; question three puts the change in a social world, because fabrications
have no audience; question four distinguishes an update from a homecoming, since
moving toward the people around you is not evidence of anything.
What you get back
Shown once, privately. Describe, don’t grade.
What this page is, and is not
This is a demonstration of the instrument’s shape — what it
asks, in what order, and what it does with the answer. It is not evidence that the
instrument works. Nothing here has been validated: the rubrics that decide
admission need two blind raters and a measured inter-rater agreement before any
score means anything, and that work has not been done. Treat every figure on this
page as illustrative, including the ones that look precise.
What is not here
No composite score. No rank. No archetype label pretending to be a personality
type. Nothing above appears on your profile unless you choose to publish it, one
dimension at a time.