Anton Vey had been dead for nearly five years when Shah finally heard him explain VUBU.
The recording had been found among the material surrendered during the audit. It was not secret. That was perhaps the strangest thing about it.
Anton had given the talk at a computational sciences conference eighteen months before VUBU launched. There were forty-three people in the audience. Until Davey Mann disappeared, the recording had accumulated only a few thousand views. Then somebody found it yesterday. Over the previous twenty-four hours, it had been watched more than three million times, its least remarkable sentences dissected alongside its most extraordinary ones.
At the time of the conference, VUBU did not exist as a product. There was no company, no consumer service and no angel investment; only Anton, a small research project and a handful of people who understood enough of his work to find it interesting. Some of VUBU’s later investors would eventually discover the recording, but by then Anton spoke a different language. Observation had become prediction. Inaccessible states became predictive modelling. Whatever he believed he had found was translated into the familiar vocabulary of artificial intelligence, and money followed. Nobody appears to have gone back and asked why the man selling them a prediction engine had once spent forty minutes insisting that prediction was precisely what it was not.
The title was:
PREDICTION AS AN ERROR OF PERSPECTIVE
Anton appeared younger than Shah expected.
He was forty-one, narrow-shouldered, with untidy dark hair and the distracted posture of a man whose body had accompanied him somewhere his attention had not. There was no podium. He stood beside a large screen showing a white circle on a black background.
His first words were already underway when the recording began.
“—because prediction is only impressive when observation is impossible.”
Shah stopped the video.
He went back.
Anton began again.
“Prediction is only impressive when observation is impossible. If I tell you it will rain tomorrow, you might ask how I know. If I tell you it is raining in Tokyo, you don't ask how I predicted Tokyo.”
A few people laughed.
Anton did not.
“The distinction is access.”
Shah wrote the word down.
He had been keeping notes for three months, although increasingly he was unsure what they were notes towards.
Anton changed the slide.
A horizontal line appeared.
A dot sat near its centre.
“We have a habit of drawing time like this because this is how we experience it. Behind us. Here. Ahead of us.”
Three more dots appeared.
“The diagram is psychologically accurate. That doesn't make it physically accurate.”
Anton walked across the stage and stood partly in front of the projection.
“If you lived your entire life travelling north, you might develop a very sophisticated vocabulary for north. You might organise your society around north. You might build religions explaining north. None of this would prove south had ceased to exist.”
Shah paused the recording again.
He looked at the audit report open on the second monitor.
BU_COORD.
BU_OFFSET.
BU_SAMPLE.
He resumed.
Anton had not yet built VUBU.
Not publicly, at least.
He referred throughout the talk to an experimental system called V, which he described as a method of extracting visual information from states unavailable to direct observation. Whenever somebody in the audience asked whether this meant prediction, Anton corrected them.
Not impatiently.
Precisely.
“No. Prediction implies derivation.”
Later:
“No. You're adding probability because you're assuming absence.”
And, when one questioner persisted:
“You keep asking me how accurately we can guess what is in the room. I'm telling you I think I found a window.”
There was laughter at that.
Anton smiled.
It was the first time Shah saw him do so.
“I appreciate how that sounds.”
The talk moved into mathematics Shah could not follow. Much of it remained beyond the auditors too. Anton discussed information boundaries, inaccessible states, observational asymmetry and something he called temporal occlusion.
Then, twenty-eight minutes in, the tone changed.
Anton put down the remote.
“I should say something about consequences.”
He looked towards the audience.
“If this works, there will be four stages.”
Shah leaned forward.
“The first is wonder.”
Anton spoke the word almost dismissively.
“This will be the shortest stage. People will ask the large questions. Can we see our deaths? Can we prevent disasters? Can we know who wins wars? Can we look ten years ahead? A hundred? People always begin with eternity.”
The slide behind him disappeared.
“The second stage is practicality.”
A few people were still taking notes.
“This will last longer. Once people discover the system is constrained, they will stop asking what time is and start asking whether they should carry an umbrella.”
More laughter.
Anton waited for it to finish.
“They will use it for deliveries. Traffic. Clothing. Appointments. They will discover that access to another part of time is useful mainly because human beings are late.”
Shah sat very still.
He had read the early VUBU marketing reports.
Weather.
Deliveries.
Commuting.
Household planning.
Anton continued.
“The third stage is boredom.”
Nobody laughed this time.
“Anything miraculous becomes infrastructure if it works reliably enough. Reliability destroys astonishment. Nobody wakes each morning amazed that electricity has returned to the wall.”
He walked slowly across the stage.
“Eventually people will forget what the machine implies. This is necessary. You cannot live permanently in metaphysical astonishment. You still have to buy milk.”
Shah stopped the recording.
Outside the interview room, somebody walked past carrying two coffees.
For several seconds he could hear nothing except the ventilation system.
He pressed play.
Anton was answering a question from the audience.
The question itself was indistinct.
His answer was not.
“No. The fourth stage is fear.”
The room became quiet.
Anton looked down before continuing.
“People say they want to know the future. They don't. They want reassurance from the future.”
He glanced towards the screen behind him, although it was now blank.
“Those are different products.”
Shah watched the sentence again.
Then once more.
People want reassurance from the future.
He thought of Graham Mann standing beneath television lights.
Anyone who knows anything.
Anyone who remembers anything.
Please.
Anton continued.
“A useful future is one which confirms that tomorrow remains ordinary. Your parcel arrives. Your meeting happens. Your garden is wet. You are still there to see these things.”
He paused.
“What happens when the window shows you something you would rather not know?”
Nobody answered.
Anton nodded slightly, as though they had.
“Then you discover why we evolved facing one direction.”
The video continued for another twelve minutes.
It was during those twelve minutes that Shah began to understand why Anton had encrypted BU.
Not because the information was commercially valuable.
Not because it contained customer data.
Anton believed access itself was dangerous.
He described the problem as a firewall.
The term seemed deliberately inelegant.
“Every information system has boundaries. We usually regard the boundary as an inconvenience. Something to bypass.”
He picked up the remote again but did not use it.
“I made that mistake.”
There was a different quality to him now.
The earlier Anton had seemed pleased by the intellectual problem. This one did not.
“I thought the boundary was a technical limitation. Noise. Encryption, if you like. I became interested in decoding it.”
Someone asked a question.
Anton listened.
“No. Not encryption placed there by a person.”
Another question.
He hesitated.
“I don't know what placed it there.”
The room became very quiet.
Anton looked almost annoyed with himself.
“I am going to use language I dislike.”
He took a drink of water.
“There appears to be intelligence in the structure.”
Shah stopped the recording.
He sat back.
The sentence remained on the screen beneath Anton's frozen face.
There appears to be intelligence in the structure.
Shah had spent most of his career learning not to enlarge mysteries merely because they resisted explanation.
A missing child did not become supernatural because nobody found him.
An image did not become a message because nobody understood how it had been made.
Coincidence survived investigation remarkably well.
So did madness.
He pressed play.
“I don't mean a person. I don't mean a god. I don't know what I mean.”
Anton seemed irritated now, not with the audience but with the inadequacy of the available nouns.
“But there are limits to what can be extracted. At first I thought those limits were defects in my method. The better the method became, the more deliberate the limits appeared.”
Another question came from somewhere near the back.
Anton listened.
“Yes. A firewall is the least bad analogy.”
The questioner spoke again.
This time Anton laughed.
“No. I haven't hacked the universe.”
A longer laugh from the room.
Anton waited.
“I decoded a very small part of something that may be there precisely to prevent decoding.”
The laughter disappeared.
“That distinction has begun to trouble me.”
Shah thought of the note.
PUBLIC SAFETY REQUIREMENT.
DO NOT REMOVE.
Anton went on.
“The mistake engineers make is assuming that because a barrier can be overcome, overcoming it constitutes progress.”
He looked briefly towards the camera.
“It does not.”
Another slide appeared.
It contained no equations.
Only a sentence.
ACCESS IS NOT PERMISSION.
Anton turned to look at it.
“I don't know whether the boundary can respond. I don't know whether response is even the correct word. But if we continue increasing resolution, eventually I expect the system to produce something that makes continued observation undesirable.”
A woman in the audience asked what he meant.
Anton considered his answer for several seconds.
“A warning would be the human word.”
Shah felt something move coldly through him.
He stopped the video.
On the desk beside him lay a printed copy of the Mann image.
The poster.
Davey's face.
MISSING.
For months everybody had treated the image as though it belonged to Davey's disappearance.
Cause and consequence had become psychologically inseparable.
VUBU had shown the poster.
Then Davey disappeared.
Therefore the two events belonged to one another.
The hacker theories assumed it.
The conspiracy theories assumed it.
The newspapers assumed it.
The police, despite knowing better, had begun by assuming it too.
Shah looked at Davey's photograph.
A warning.
He resumed the video.
Anton was near the end of the talk now.
“What happens after the fourth stage?” someone asked.
Anton smiled again.
“We stop.”
“Why?”
“Because fear is more persuasive than wonder.”
The audience laughed softly.
Anton didn't.
“Wonder makes you open the door. Practicality makes you install the door in every house. Boredom makes you forget the door is there.”
He looked towards the blank screen.
“Fear makes you close it.”
The talk ended shortly afterwards.
There was applause.
Anton thanked the audience and began gathering his papers.
Then somebody asked one final question.
The conference microphone had already been switched off. The voice was distant, almost lost beneath the movement of chairs.
Anton looked up.
The person repeated it.
This time Shah could make out the words.
“If it isn't prediction, what is it doing?”
Anton stopped gathering his papers.
He appeared to consider whether the question required another explanation.
Then he shook his head.
“No. That's the mistake.”
The room quietened again.
“Everyone keeps imagining the machine here, looking forward to there.”
He pointed somewhere beyond the stage.
“But that's only because that's how you experience it.”
A pause.
“The machine isn't looking forward.”
Anton picked up his papers.
“It's looking around.”
The recording ended.
Shah remained where he was.
For months, they had investigated the VUBU image and the disappearance as though one must explain the other.
Perhaps that had been the mistake.
There was still no evidence that VUBU had anything to do with what happened to Davey Mann.
There was only a nine-year-old boy who had left home for school one morning and never arrived.
And somewhere, Shah thought, there was still an answer to that.