Essay · Companion
Open in Principle, Blind in Practice
Why a funder willing to back neglected work still needs the map — and reproduces the loop without it
A companion to The Loop No One Chose. Start there for the core argument — that a decentralized field is steered by an invisible attention-loop, and the answer is a faithful mirror rather than a controller. This piece answers its sharpest objection.
The argument for a self-observing map invites an obvious objection, and it is worth taking seriously because it is the cleanest test of whether the argument holds at all. If the field’s problem is that important work goes unfunded while crowded work compounds, why build an elaborate observatory? Why not simply have funders who are open to neglected ideas? A grantmaker who deliberately seeks the empty corners would seem to correct the skew directly, no infrastructure required. If that worked, the observatory would be unnecessary, and the whole case for it would collapse into a one-line recommendation: fund the gaps.
It does not work, and the reasons it does not are the same reasons the map is needed in the first place. An open-minded funder is not an alternative to the mirror. They are the party the mirror exists to serve, and on their own, without it, they reproduce the very loop they mean to escape, only running it in reverse.
“Neglected” is a relation, not a property
Start with the deepest problem, which is definitional. Neglectedness is not something a research direction possesses on its own, the way it possesses a difficulty or a timeline. It is a relation between a direction and the field’s current distribution of attention and money. A thing is neglected only relative to how much attention it deserves set against how little it gets. Which means that to fund the neglected thing, you must first know what is underfunded relative to its importance, and that is precisely the legible, current picture the field does not have.
So a funder “open to neglected ideas” but without the map is open in principle and blind in practice. Their good intentions are real, but intentions are not a readout. Lacking one, they fall back on the same instruments everyone else uses: the stale, hand-assembled snapshot, the residual sense of what was hot at the last conference, the gossip of who is working on what. That is the attention proxy the original argument identifies as broken. The open-minded funder, deprived of the map, is not standing outside the loop. They are inside it, reasoning from the same corrupted signal as everyone else, merely with better motives.
This is why the funder and the map are not competing solutions. The funder’s openness is exactly the demand that makes the map worth building. They are not an alternative to the observatory; they are the thing the observatory is a prerequisite for.
The reverse herd
Suppose the funder presses on anyway and commits, as a matter of policy, to chasing whatever looks most neglected. They have not stepped outside the loop. They have only reversed its direction.
Watch what happens when enough capital does this. Neglectedness itself becomes a signal: the attended-to thing, the quality everyone is now scanning for. And the moment a direction gets funded for being neglected, it stops being “neglected.” The signal that drew the money destroys itself by being acted on. A reverse-herd, having corrected one empty corner, finds it no longer empty and moves to the next, which it will likewise fill and abandon. The compounding instability of the original loop becomes an oscillating one: directions get over-corrected, rushed into funding, then dropped the instant they stop reading as the underdog.
Chasing the inverse of a broken proxy is still chasing the proxy.
The map breaks this, but not by telling funders to seek neglect. It breaks it by showing the actual distribution, how large each gap really is, so that corrections can be sized to the real shortfall rather than triggered by the binary signal “this is the neglected thing now.” A direction that is twenty percent underfunded and a direction that is empty deserve different responses, and that distinction is invisible when the only input is whether a line of work currently wears the underdog’s colors.
Neglect is not importance
There is a flaw beneath the policy itself, independent of the herding. “Only fund neglected work” is a prescription, and a bad one, because neglect and importance are not the same thing.
Some directions are neglected because they deserve to be: they are premature, or unpromising, or quiet dead ends the field has correctly walked away from. The diffuse judgment of a field is not pure noise. It gets a great deal wrong (that is the whole premise), but some of what it ignores, it ignores for good reason. An organization chartered to work only on the neglected has, by that charter, committed to a contrarian bet as standing policy. It will fund the unpromising-and-ignored right alongside the important-and-ignored, because its selection criterion cannot tell them apart. Neglect is all it can see, and neglect does not distinguish buried treasure from what was buried for cause.
That is not correcting the loop’s distortion; it is adopting the mirror-image distortion as a mission, swapping a bias toward the crowded for a bias toward the empty, neither of which tracks what actually matters.
You no more prescribe neglect than you prescribe popularity.
The honest move is the one the main argument lands on: make the distribution visible and let many independent judgments decide which neglected things are neglected wrongly, and which are neglected exactly as they should be.
The underdog’s halo
Underneath all three reasons sits a subtler trap, and it deserves a name, because it is the most seductive form of the question. An organization that does not merely fund neglected work but authoritatively designates which neglected directions are worth backing has become the centralized recommender the fixer’s trap warns against, now in its most dangerous costume yet.
It is dangerous precisely because it feels so virtuous. Correcting an imbalance, fighting concentration, backing the overlooked: these wear an unimpeachable halo.
“We fund the gaps” is one short step from “we decide what the gaps are.”
That second sentence is the loop’s most efficient engine with a conscience bolted on. A single actor now sets the field’s contrarian agenda, directing capital by its own lights, and does so under the moral cover of the underdog, which makes it harder to contest than a naked authority would be, because pushing back looks like siding with concentration. The recommender that issues from the fixer’s trap is bad enough when it tells the field to work on what is popular. It is worse when it tells the field to work on what is neglected, because then resisting it feels like a betrayal.
Demand, not substitute
So the relationship between the open-minded funder and the map is the opposite of a contest between two fixes. The funder is good and necessary. They are also the demand side the observatory serves, not a replacement for it.
The map does not tell that funder what to back. It hands them, and every other reader, an accurate picture of where the gaps actually fall and how deep they run, so that an openness to neglect can finally act on reality instead of on vibes. And because the picture is shared, the correction it enables comes from many funders reading the same legible distribution and reaching their own conclusions, not from one neglect-allocator deciding, on the field’s behalf, where the underdogs are. That dispersal is the entire point. It is the difference, once again, between reflexivity that returns judgment to the reader and reflexivity that substitutes for it.
Open-minded funders are who the mirror is for. They are not a substitute for it, and without it they do not escape the loop. They run it backwards, with the underdog’s halo on, which is the one disguise under which the loop is hardest of all to see.
Read next →
How to Build a Mirror Without Holding It
How you actually build the self-observing map — and why the attention loop climbs into the very tools meant to expose it, so the answer is a best-available, endlessly improvable mirror rather than a neutral one.
← The original
The Loop No One Chose
Reflexive dynamics in decentralized nonprofit fields, and why AI safety needs a live, actionable map that reflects the field back to itself instead of telling it what to do.