[FRIAM] marginalist pricing for ideas

[glen]

The ubiquity of the assumption that the universe is knowable has always irritated me. At surface, a shift away from the authority of the Academy to the authority of {Gov, Corp, NGO, Think Tank} might seem like a wash from an anti-Scientism stance. But the dichotomy is mirrored in open versus closed source software, with freemium in between. Given that analogy, the categories are more like {Scholarly, Think Tank, Journo} ↔ {NGO, Academy} ↔ {Gov, Corp}.

The point being that the left (on this spectrum) wears their reasoning on their sleeves, whereas the right obscures their reasoning. (I'm sure some might challenge my assertion that the Academy has only translucent reasoning ... or that *some* Gov like NOAA bares more of its reasoning than other Gov -- all good points.)

The thing that often goes unregistered in this knowability conversation is the cognitive/compute capacity of the knower. It's irrelevant if the universe is knowable, as long as the only things doing the knowing are too tightly scoped to read the universe. Scholarly works that are so dense or so steeped in jargon that one could spend a PhD and 40 years of their life exploring it, are effectively the same as authoritarian/deontological mandates.

Hence, the composition from individual knower to collective knower becomes the fulcrum. Do we put our trust in "agentic" wiki-style composition or transformer training style composition? My answer is neither. They're 2 tools in the same tool kit. The trick is knowing when to use which tool.

But I'm in a privileged position. It's up to those of us with more ability to know to engineer the future for those of us with less ability to know. Is that paternalistic frame also "marginalized pricing"? Or at least some form of it? I have in mind the intrepid investigative journalists who spend all their spare resources digging into some obscure topic and dumping what they learned into the open. They pay the price, but in sweat/sanity more than money. (Granted, the "system" eventually gets them with sponsorships and coercive "grants" ... but prior to their corruption, "price" isn't so clearly flat.)

On 12/9/25 4:23 AM, Santafe wrote:
> So the modern “law of one price” in posted-price markets, also known as “marginalist pricing”, works as follows:
> 
> Everybody pays the same price, and the price is set wherever the market clears, with no further buyers left who are willing to pay more, and no further sellers left who are willing to accept less.
> 
> It gives us such outcomes as rich people using water on golf courses in Scottsdale while poor people may not have drinking water in other Phoenix neighborhoods, and AirBnB renters owning all the property in tourist-heavy neighborhoods while no wage-earners can afford houses there.
> 
> In the era when Private Equity is becoming a kind of singleton in the wealth distribution, PE will own all of everything at the posted price, and nobody else will own any of anything.  We will need surreal numbers to subdivide that delta-function, to characterize the intramural competitions of the PE firms with each other.
> 
> The argument “for” marginalist pricing is that it is the only “fair” system, by the criterion that whoever wants something the most, and is willing to push for it the most, should be the one to get it.  We put aside, of course, all the bad faith and nonsense in that argument, since in a system with vast asymmetries in access, information, risk buffering, legal and civil protections, etc., that reflect as asymmetries in purchasing power, all these measures of “force” are wildly artifacted and don’t correlate well with any sane measure of “need” or “want”.
> 
> The following paper gives a nice statistical breakdown and analysis of what marginalist pricing can look like for ideas (still early days):
> preview.png
> 2512 <https://arxiv.org/pdf/2512.03337>
> PDF Document · 11.8 MB <https://arxiv.org/pdf/2512.03337>
> 
> <https://arxiv.org/pdf/2512.03337>
> 
> One probably doesn’t even need to “bias” the input of the ML system to produce skewed outputs; it can just be driven by the bulk of volume of noise in the available sources.  Whoever “wants” a position most, quantified by the noise level generated, can have the position.
> 
> Gonna get harder and harder to maintain a community of sanity and sense in the world rapidly being built up around us, and hoping to drown us all.
> 
> Eric
-- 
¡sıɹƎ ןıɐH ⊥ ɐןןǝdoɹ ǝ uǝןƃ
ὅτε oi μὲν ἄλλοι κύνες τοὺς ἐχϑροὺς δάκνουσιν, ἐγὰ δὲ τοὺς φίλους, ἵνα σώσω.