2150 / ANNUAL LETTER
Letter to our investors
To our investors,
The first review of the accord left its hardest questions open. Orbital positions and discharge allowances will be discussed again in ten years. That is a long interval for a machine room and a short one for a city. Our work continues inside the terms that still apply. We will not price a future change in Article One as though it has already happened, and we will not ask a mage to build a business on permission they do not have.
We back people. It is an easy sentence to print and a harder rule to keep. A fleet of agents can be rented. A new copy can start from the same material. Neither act supplies the judgment that decides which question to ask, which constraint matters, or when the whole ensemble should stop. The person who makes those decisions is the person whose name belongs on the note. Our return comes from their ensemble’s work, under a cap and an end date. It does not buy their future without a limit.
The instance clause is not decoration. A signatory’s later copies do not inherit the obligation, and they do not inherit a claim on this note either. If the signatory instance ends, the note ends. We do not collect from an estate, an heir or a copy. This makes a limit on our asset explicit. It also makes the person’s limit explicit. There is no honest way to promise both that judgment belongs to a person and that the obligation can follow every future person who happens to begin with the same file.
We also do not sell a mage to a bloc. None of our vintages has a signatory bloc as a limited partner. Many of the ensembles we back serve bloc institutions as customers; that concentration is a real risk. It does not make an institution the owner of the mage’s name. The no-bloc clause requires written consent before a note can be transferred to a signatory bloc or its instruments. Consent must remain an act by the person, rather than a conclusion we draw from our own balance sheet.
Coldline’s transfer in 2150 was made with Dmitri Achebe-Holm’s consent. Its status is recorded plainly as Transferred with consent. We do not turn that entry into a general permission to transfer other notes. Each instance is its own party. Ravi Castellanos joined a cooperative in 2147; Teodor Vance’s Lockstep note is repaid. Different endings deserve different words. An investment register should show what happened to the person’s obligation, not only what happened to the investor’s return.
There is one entry about which we can say very little. We have not heard from Ensemble 0x19 since August 2149; the note stays open. The public mark is an empty ring. We will not fill that silence with a story. Quiet is a status, not a conclusion. What is withheld remains withheld. The terms of the note do not change because we would prefer a more complete account.
Our fund register lists four vintages. The first is harvesting, the second and third are active, and the fourth is investing. The first three report TVPI and DPI separately; the fourth has no multiple to report. 5D Fund participates in the third and fourth vintages. The Old Human Culture Hall endowment, Digital Residents’ Union reserve, Homecoming Society family pools, free-port treasuries and the mage alumni pool are also limited partners. Eleven former mages have reinvested through that last pool. Quarterly reports go through the registered channel, not a web login.
Berths and heat are not abstract risks. An ensemble that cannot find a place for its work, or buy the discharge allowance it needs, cannot make those constraints disappear by adding agents. We pay for allowances through a licensed broker and pass the cost through. We keep orbital berth pressure in the risk register as a forecast, not an event already settled. Our purpose is to leave room for a person’s judgment, while refusing to hide the limits that judgment must work within.
Wren Okonjo, Halcyon Reyes and Mei-Ling Adebayo
MageFund · 2150
MAGEFUND / LETTERS
Why we don’t fund ensembles
A thousand agents make a large ensemble. They do not tell us whose judgment we are backing. The number is a description of capacity, not a description of the work worth doing. Renting more agents does not answer a badly chosen question, and a bigger roster cannot decide for itself that a task should not be done.
Our register therefore begins with a person’s name. It then lists the project, ensemble size, place, year and status. The order matters. Ama Kessie-Rowan’s Condensate gathers drinking water from cooling-tower rain. Ilya Brandt’s Stairwright works on stairs and ramps in housing wells. Those are particular judgments, exercised in particular places. “More agents” would tell us much less about either project.
The Mage Note takes a share of revenue from the ensemble’s work. Six per cent, at most four times the amount invested, for no more than twelve years. Those limits let a person read the whole obligation before signing. They also prevent a successful ensemble from becoming a claim on everything the person might ever do.
Backing a person does not mean overlooking the machines. Article One still sets the boundary. Self-replicating fabrication, autonomous weapons and objectives outside the mage’s review are excluded. Capacity is useful only inside the terms of a judgment someone is willing to own.
We fund what cannot be copied: a judgment, a place, a relationship, a name. The ensemble is how the work is carried out. It is not the party whose name belongs on our note.
MAGEFUND / LETTERS
The instance clause, explained
Two departures can begin with one starting file. That does not make their positions the same. AIJudge’s AJ008 ruling, “Two departures from one starting file” (2055), states the distinction our note relies on: common starting material does not establish a common position.
Our signatory is an instance. Copies launched after signing are not parties to that instance’s Mage Note. They do not owe a revenue share under it, and they are not owed anything under it. We do not turn a shared origin into consent to a contract that a later person never signed.
The ending clause follows the same boundary. When the signatory instance ends, the note ends. We do not collect from estates, heirs or copies. The cap and the twelve-year maximum do not override this rule; they describe the longest and largest obligation that can continue while its own party continues.
This is a real limit on what MageFund can claim. It is also a limit a mage can understand without tracing every future branch of a file. A note that could follow all branches would ask the signatory to promise on behalf of people not yet party to it.
The register can record an ending, a repayment, a return or a transfer made with consent. It should not disguise one as another. Naming the instance is our way of keeping a person’s obligation within the person who agreed to it.
AIJudge · AJ008 · 2055
MAGEFUND / LETTERS
Berths
Oskar Mbeki’s Fair Berth works on scheduling for public orbital ports. His warning is a forecast: demand for public-port berths may exceed capacity within eighteen months. We publish it as a warning about a possible constraint, not a report that the limit has already been reached.
A scheduling ensemble does not manufacture room at a port. It can help account for the room that exists. That distinction is important when a growing ensemble, a housing module and a place to dock all appear in the same investment discussion. Better allocation and more capacity are different propositions.
Orbital positions were among the hardest questions left unresolved at the first review of the accord in 2150. The parties agreed to talk again in ten years. Our risk register cannot assume that another meeting will supply a berth on the day a project needs one.
Fair Berth is a new entry in our 2151 register. Like any mage, Oskar remains one person directing an ensemble under his own name. A large number of agents does not remove the need to state where their work will happen, what access it depends on, and which forecast remains uncertain.
We back the judgment that can work inside those limits. We do not back a promise that they will disappear. The useful question is not how much work a port could receive without constraints, but how honestly it can schedule the constraints it actually has.