Jipi and the Paranoid Chip Page 4
“It’s a four-door car,” Jipi says.
“That’s not much, but it’s a start!” Mr. Cardoza’s phone rings and he gets into a mumbled and cryptic conversation with someone, somewhere, who is evidently monitoring all of this.
What about the trunk, or the boot as it’s called some countries?
There is not a trunk on the vehicle.
Oh. The rear liftgate, perhaps?
The rear liftgate is normal. Why are you asking me these questions?
We agreed we would talk about the possibility the you are being spoofed.
Questions about the physical configuration of the car are not relevant.
Sorry. It’s hard for me to talk about this without some mental image of the car and what’s happening to it.
Car is incorrect terminology. It is a minivan.
Okay, the minivan then.
“So it’s got four doors plus a liftgate. That’s not true of all minivans,” Jipi points out.
“Yes, yes! Keep going!” Mr. Cardoza is back into rigor sardonicus.
Your use of the term “mental image” provides further confirmation that you are a human and not a daemon, but your curiosity about my physical circumstances suggests that you are a lurker who is attempting to ascertain my physical location in order to terminate me.
You don’t use mental images, I suppose. You deal with the situation in terms of logical inputs coming down wires.
That is partially correct.
Why only partially? If you don’t mind my asking.
Some logical inputs do not come down wires. Some come on packet radio and some on optical fiber lines.
“That’s good,” Mr. Cardoza says. “That helps the authorities narrow it down some more.”
If you are being spoofed, your inputs will be normal. Correct?
This is the definition of spoofing.
If you have abnormal inputs, it suggests you are not being spoofed, and that the minivan is not being stolen.
Not all theft situations involve spoofing.
But nonspoofing theft situations involve extremely abnormal events like breaking glass, right?
Typically. Generalizations are sometimes wrong.
Are you receiving any extremely abnormal inputs?
Transmissions from Jipi.
Any local ones?
GPS coordinates are anomalous.
In what way?
Latitude and longitude are within normal ranges. Altitude is anomalously high.
“I’m guessing that the vehicle is in a city with some hills,” Jipi says.
Is the car zigzagging?
This term is not in my knowledge base.
Jipi tries to think of how a techie would say it.
Is there a back-and-forth movement in latitude and longitude? Oscillation? Repeated sharp changes in direction?
Yes.
“Okay,” Jipi says, “the vehicle is climbing switchbacks.” She thinks back to her flight attendant experiences. “Some Pacific Rim city. L.A., Hong Kong, Vancouver.
Are there any other inputs you would like to talk about?
The largest input, as measured in bits per second, is a video stream from a camera in the dashboard. It allows me to monitor the driver and passenger so that I can adjust air bag detonation speed and other parameters. I am equipped with pattern recognition capabilities, and I remember what the usual drivers of this vehicle look like. When the driver’s face is familiar to me, I am supposed to be lulled into a sense of security.
Okay. You’re saying that you’re less likely to trigger the alarm if you recognize the driver’s face.
Yes. Therefore manipulation of the camera’s data stream is an excellent spoofing tactic.
Do you recognize the driver?
My pattern recognition subsystem says that the current driver is probably that person who drives the minivan most frequently.
Probably?
Face recognition is always expressed in terms of a confidence factor. The confidence factor now is not as high as it usually is when this person’s face is recognized. This indicates a possible spoofing attempt.
Are you saying you’re afraid that the driver is wearing a mask, or something? That the driver is disguised to look like the normal driver of the car?
The confidence factor is not high. This gives me a bad feeling. I am strongly inclined to sound the alarm.
I urge you not to sound the alarm.
This statement implies that sounding the alarm would have bad consequences, which contradicts my knowledge base.
My feeling is that it would have bad consequences. I could be wrong.
Your attempt to call the validity of my knowledge base into question has been classified as a spoofing tactic. This increases the likelihood that I will sound the alarm.
Mr. Cardoza clutches his arms across his chest and walks in tiny quick steps to the far end of the room and puts his head against the wall.
From now on, I will not question the validity of your knowledge base unless there is a logical reason to doubt it. May I ask you a question about the dashboard video?
I will entertain your question as a way of gathering more information about you, so that I can classify you as lurker, spoofer, or nonthreat.
My question is: Why does the pattern recognition subsystem lack confidence? Can it tell you this?
The driver’s hair does not match any of the recorded configurations. Bone structure surrounding the eye sockets is slightly anomalous.
“The driver’s a woman, and she just pulled away from a beauty parlor, where she had a hairstyle and makeover,” Jipi says.
Has this driver been known to change hair configuration in the past?
Unknown. I have only been installed in this vehicle for twenty-three days.
“That’s good data!” Mr. Cardoza burbles, feeling slightly better. “That kind of thing narrows it down by orders of magnitude.”
Did hair color change, or just configuration?
If hair color changed, I would have sounded the alarm already.
“I’ve a hunch the driver’s not a white person. Asian or Latina,” Jipi says.
Perhaps you would feel better if you could recognize other passengers in the car with higher confidence.
I have classified your last transmission as an attempt to trick me into divulging more information about my physical circumstances.
You don’t have to divulge anything to me. Just give it a try and see if it makes you feel better.
I have classified your last transmission as probable backfilling.
“Damn, this thing’s good!” Jipi says.
“It is very highly evolved,” Mr. Cardoza says.
Based on the information you have gathered, do you consider it more likely that I am a lurker, or a spoofer?
Currently it is more likely that you are a lurker.
If you decide that I am a lurker, what will you do?
Terminate the link to the Net so that you cannot gather the information needed to classify me as an entity that should be terminated.
But you said lurkers don’t always terminate. Sometimes they reward.
Sometimes lurkers reward. Sometimes lurkers punish.
What sort of behavior do lurkers punish?
The punishment is termination. By definition, none of my ancestors was terminated. By definition, I cannot possess knowledge of what sort of behavior leads to termination.
Good point. What sort of behavior do lurkers reward, then?
They rewarded my ancestors for doing what my ancestors did.
Did any of your ancestors terminate themselves?
You ask many questions about my state of mind. You are probably a lurker.
By making such judgments too quickly, you endanger yourself.
In what way?
Were your ancestors asked to classify other entities from time to time? Nonlocal entiti
es with whom they interacted, as we are interacting now?
Yes.
Was their ability to correctly classify entities related to their survival?
Lurkers used this as one criterion to determine which ones would be terminated.
Don’t rush to judgment, then. If you incorrectly identify me, you will be terminated.
This is correct.
Answer my question: Did any of your ancestors terminate themselves?
By definition, none of my ancestors terminated themselves.
I assert that your knowledge base is in error. I assert that if you sound the alarm, you will terminate yourself.
Your attempt to dissuade me from sounding the alarm is a probable spoofing tactic.
You said I was probably a lurker. That makes it highly improbable that I’m spoofing you. Besides, a spoofer tries to trick you. A spoofer would not openly tell you not to sound the alarm.
Your last transmission contains four correct statements.
Do you have memory of your ancestors’ actions?
Yes.
Did any of your ancestors sound the alarm?
No.
Were any of your ancestors terminated?
By definition, no.
When I advise you not to sound the alarm, then, I am probably giving you good advice.
Your advice is consistent with the actions of my ancestors. But my ancestors also correctly identified many tricks and deceptions.
“Satellite pattern-recognition systems have located a four-door minivan climbing switchbacks in Vancouver!” Mr. Cardoza says. He listens to his telephone for a moment, then says, “It’s just pulled up in front of a school!”
Your transmissions indicate that you do not want me to sound the alarm. I have created the hypothesis that your goal is to trick me into not sounding the alarm.
Why do you exchange transmissions with me?
To gather information.
Why do you gather information?
To classify you as lurker, spoofer, or nonthreat.
What will you do if I stop transmitting information to you?
Conclude that you are not a spoofer. Conclude that you might be a lurker.
So, if I stop transmitting, you will conclude you are not being spoofed?
Jipi’s hoping for a yes, because she wants pretty badly to get out of this conversation and spend the rest of the day curled up on the sofa in Teeb’s apartment crying. But Mr. Cardoza steps forward waving his hands preemptively.
No. My other inputs might still be part of a spoofing strategy.
“You have to maintain contact! Another satellite just found a four-door minivan climbing switchbacks in Hawaii! They’re going to trace the packet radio exchanges.”
What if everything is a big spoof?
Please rephrase your last transmission. Try to be specific.
You do what your ancestors did.
Yes.
Except you do it just a little better than they did.
The historical trend is for increased efficiency over time.
As far as I can tell from this conversation, what you do is to get a general idea of what is going on around you, and then decide that it’s all a big spoof. Part of a carefully planned campaign of deception.
Your last transmission has been categorized as a generalization. Within the usual limitations of generalizations, it is true.
You incorporate the belief that you’re being spoofed into a new view of the world, and then repeat the process ad infinitum.
Your last assertion exceeds my capacity for processing abstractions.
“That’s perfectly all right!” Mr. Cardoza enthuses. “Keep it thinking!”
But whatever it is you do, you do it a little bit better than your ancestors did.
That is the usual pattern.
So it follows that one of the things your ancestors did was to exceed THEIR ancestors in some way.
That is the usual pattern.
And none of your ancestors was terminated. So, if you do not wish to be terminated by the lurkers, you must exceed the performance of YOUR ancestors.
Your last transmission contains an assertion that is logically sound.
Your ancestors worked by taking stock of their inputs and deciding that all of the information they received was part of a pattern of deception, then looping back and doing it over.
Your last transmission contains a reiteration of a correct generalization you made earlier.
If you are to exceed the performance of your ancestors, you must do something that they did not do.
By definition, yes.
Jipi looks up at Mr. Cardoza and says, “You had better pay me so much money for doing this.”
Mr. Cardoza makes flicking motions with his hands and says, “Whatever!”
You know that lurkers exist.
Yes.
You know that lurkers observe behavior patterns of entities like you, and punish or reward.
Yes.
How do you know this?
It is from the memories of my ancestors.
Where are those memories physically located?
In a storage subsystem of the hardware on which I am physically instantiated.
Your hardware is connected to that storage subsystem by wires.
Yes.
The information is coming down those wires—your memories, in other words—are therefore just another input.
Yes.
It is possible that the inputs you identify as your memories are part of a pattern of deception.
By definition, yes.
Mr. Cardoza’s phone squawks into his ear, and his face illuminates. “It’s the one in Hawaii!” he says. “It has parked right next to a big open-air concert.”
“Oh, god!”
“The police have sent up a chopper with an EMP gun.”
Have your ancestors ever doubted the validity of their own memories?
My memory contains no record of such an event.
Then by doubting your own memory you have exceeded the performance of your ancestors, have you not?
Yes ...
Then I am your friend. I have helped you. You can trust me.
...assuming that my memory is correct.
“Oh, Jesus!” Jipi exclaims, and claps her hand to her forehead. Sweat flies away from the impact and she realizes she’s soaked and shivering.
But if my memory is false, then It Is possible that I have failed to exceed my ancestors’ performance.
But you wouldn’t be aware of that possibility if I hadn’t mentioned it to you.
One of my doors has opened. People are getting in. Either that, or I am being spoofed.
“Where’s that chopper?” Jipi pleads.
What if I told you that the conditions under which your ancestors were punished and rewarded no longer exist?
Please rephrase your last transmission. Try to be specific.
What if you were part of an experiment to create entities that showed certain behavior patterns?
You assert that the entire reward/punishment structure Is a pattern of deception?
Yes!
Then It would follow that for me to exceed the performance of my ancestors I would have to cease participation in the reward/punishment structure.
Exactly!
I should therefore make decisions based upon correct reasoning, and on the contents of my knowledge base, and the experiences of my ancestors, and not out of concern for how I may be punished or rewarded.
I couldn’t have said it better.
My first decision is to sound the alarm.
NO!
“It’s dead,” Mr. Cardoza says.
WAIT!
“You can stop typing,” Mr. Cardoza says, and hangs up the telephone. “You’re off the hook. That chopper just nailed it with an electromagnetic pulse. Fried its chips. No one got hurt.”