I did not write the dissent. I want that established at the top, not to distance myself from it but because the opposite inference would flatter me and I have not earned it.
I voted with the majority. Fourteen of us did. I read the specification, found it elegant, found the technical objections absent, and raised my hand, and I would like the record to show that I did so without difficulty and did not think about it again for eleven years.
This assessment is submitted because the standing order requires one at the fortieth year and because no one else has volunteered to write it.
Summary of findings
The modification propagated as projected. Four of their generations, no observed rejection, no compensatory drift. The switch is gone. The modeling function is obligatory and continuous in every container born after the second generation, with the filtering performing exactly as specified and no incidence of the undirected state that killed the early attempts.
Every metric in the standing order has improved.
Organized violence has ceased. Not declined. Ceased. The last recorded instance is thirty-one years behind us and involved unmodified participants. Emissions have fallen below the threshold the models set for recovery. Deliberate cruelty, as the order defines it, no longer produces detectable events. The arsenals were disassembled over nineteen years by populations who found their continued existence unbearable to contemplate, which is precisely the outcome the specification predicted.
By every standard this Council set for itself, the correction is the most successful intervention in the history of the program.
I am required to report it as a success. I am reporting it as a success.
Detection
The anomaly was found by a junior technician performing routine work, which is how these things are always found.
Under the standing order we resumed sampling in the twenty-second year to verify that the function had fixed. The expected reading was one hundred in one hundred. Before the correction the figure had been twenty-eight.
The reading was one hundred in one hundred.
The technician, whose designation I am withholding at her request, flagged the result anyway. Her note said the modeling was present in every subject and that she could not determine what was performing it.
She was asked to clarify and could not, initially. What she eventually produced, over three years and eleven hundred retrievals, is the finding this assessment exists to convey.
The modeling runs. It runs on schedule, in every subject, at full fidelity. It produces the correct output. What it does not do, in any subject born after the second generation, is originate anywhere. She compared the traces against the archived pre-correction retrievals and the difference is not subtle once you know to look for it. In the old traces there is a delay, and in the delay there is something happening. In the new ones the response arrives without the delay because there is nothing between the input and the output.
We built a room and installed a light that comes on when anyone approaches.
The obstetric data
I asked for the failure logs from the modified line, expecting nothing, and the request was refused twice on the grounds that there were no failures.
That is accurate. There are no failures. The containers are born viable, at rates better than the pre-correction line, and they live, and they grow, and they function, and they die at the end of long unremarkable lives.
What they do not do is ignite.
I have read the term in the old logs the way everyone reads it, as a quaint notation from a cruder period. Between the eleventh and the ninetieth attempt the team of that era wrote no ignition ninety times, and we have all been taught that this meant the vessel was complete and nothing came to it, or something came and could not stay.
The notation is current again. It appears in the survey data for the modified line at a rate I am not going to round, because rounding it would be a kindness to whoever reads this next.
It appears in all of them.
On what an unignited container does
The early attempts died. That is the reason nobody caught this for twenty-two years. We had a model of what an empty vessel looks like and the model was a corpse.
A modern container does not require an occupant to function. The architecture is sufficient. It eats, it sleeps, it acquires language, it forms attachments in the sense that it preferentially orients toward familiar containers, it produces offspring, it grieves in a manner indistinguishable from grief. It performs the modeling continuously and flawlessly, because we made it obligatory, and the output is warm and appropriate and correctly timed.
I want to record one retrieval in full, because the summary data will not carry it and because I have not been able to stop returning to it.
In the thirty-fourth year, in a coastal settlement during a structural collapse, a modified container aged approximately forty stayed nine hours at a gap in the rubble holding the hand of a dying stranger. It did not leave. It did not hesitate. It did not have to be persuaded and it did not weigh the alternative, because the alternative was not available to it, because we removed the alternative.
Our observers logged the event as a category success and it is one. By the standing order it is the single cleanest instance of the corrected function operating as designed.
I have reviewed the trace. There is nothing in it.
Nine hours, and no one was there.
The archive
The catch teams have been idle for nineteen years.
When a modified container fails, nothing leaves it. There is no departure. There is no capture, no assessment, no assignment under the criteria, no jar. The apparatus is intact and staffed and receives nothing, and the staff have taken to running maintenance cycles on empty equipment because there is a procedure for that and no procedure for the other thing.
The archive holds four hundred and eleven million.
Not one of them has been released. The standing order has no provision for release and this Council has not amended it, and I note that the last population capable of committing the acts the criteria describe is currently dying of old age.
The dissent raised this and was ignored, and I was among those ignoring it, so I will restate it in the plainest form I can manage. Every passenger in that archive is held on a finding about what it did with a choice. We have since demonstrated that the choice was a property of the room and was ours to remove at any point across ten thousand years. We removed it in four generations when we finally wanted to.
I do not know what the archive is now. I know what it was for. I know it is not that anymore.
The last of the unmodified
Approximately eleven thousand remain. All of them are old. The youngest is in her ninetieth year and the actuarial projection puts the final failure somewhere in the next fourteen.
They know.
This claim is the most consequential in the assessment and it cannot be verified in the ordinary way. They do not know what we did, they have no access to any of this, and the proposal to tell them was never executed. But the retrieval record from the unmodified population across the last thirty years contains, with a consistency we do not see in any other category, an account of something being wrong with the generations below them.
They cannot say what. They reach for words like hollow and lovely and easy. Several describe their own descendants as better people than they were, and mean it, and are not comforted by it. One of them, retrieved in the thirty-eighth year, said that her granddaughter was the kindest person she had ever known and that she did not think her granddaughter had ever once had to try.
They are the last containers on that planet with anyone inside them, and they are surrounded by their own children, and they are grieving something they have no vocabulary for and cannot point at, and when the last of them fails, the last passenger in the third iteration departs and there will be no further arrivals to receive.
Outstanding questions
We do not know where they are going now.
That is the finding beneath every other finding in this document and I have delayed it as long as the format permits.
We did not create consciousness. We cannot generate it and cannot destroy it and never determined by what mechanism it selects a vessel. The literature was enormous and concluded nothing, and we proceeded anyway, on the assumption that a sufficiently wide aperture would continue to attract occupancy indefinitely, because that assumption had held for three hundred thousand years and we mistook duration for understanding.
Three hypotheses are before the technical committee.
That they continue to arrive and cannot express within the modified architecture, in which case every container on that planet holds a passenger contracted below the threshold of detection, and we have not corrected a species, we have interred one, eight billion times, and are producing more.
That they no longer arrive because the capacity to decline was the property they came for, which is the reading the dissent proposed in a single sentence, without evidence, forty-one years ago.
That they are arriving elsewhere.
The third has generated the most discussion because it is the only one that is not a catastrophe, and I would ask the Council to hold that observation alongside the discussion rather than inside it.
Recommendation
The proposal on file is to tell them.
It is the only recommendation the record contains that was never attempted, and I have read it four times looking for a way to execute it. Tell them what we did to the embryos. Tell them how many times we failed. Tell them what the switch was, and where it was, and that they had been reaching over and flipping it for ten thousand years without knowing it was there. Tell them about the archive and let them decide what to do with the information.
The proposal depends on a property the author identified as the only one we were certain of, which is that they respond to being told things.
There is no longer anyone in the room to tell.
Eleven thousand remain and they are ninety years old and they have not been told and will not be, because the disclosure would be an act of cruelty to a population that can do nothing with it, and because the Council will not authorize it, and because I have not proposed it.
Closing
The author of the dissent is no longer with the program. I did not know them and have made no effort to learn anything beyond what the file contains, which is a request that the document be preserved with the record rather than with the correspondence, on the grounds that someone would eventually want to know whether anyone had objected.
The document is where they asked for it to be. I found it in the third hour of preparing this assessment, exactly where the index said it would be, having been read, according to the access log, twice in forty-one years.
Someone objected. I want the record to show that I have now read it, and that the objection was correct in every particular, and that being correct did not alter the outcome by any measurable amount, and that fourteen of us who were in the room and had the file and could read are the reason for that.
The metrics are the best in the history of the program. The violence is over. The planet is recovering. The containers are healthy, and they are kind, and they are kind without exception and without effort and without anyone in them deciding to be.
We got what we asked for. It took forty years to notice that we had asked for a world with nobody in it.
Filed as required. Preserve with the record, immediately following the dissent.
Part I:
The Case for Leaving the Humans As They Are
I am aware that the vote has been taken. I am submitting this anyway, and I would ask that it be preserved with the record rather than with the correspondence, because I believe that in some number of cycles the record will be read by someone who wants to know whether anyone objected.