Archive for January, 2009

Transparency

Posted in Doom with tags , , on January, 2009 by melendwyr

There is a comment, held in suspension, for the post “The Other Shoe”.

I reference the principle of Lex Talionis. I am not obligated to respect the speech of those who do not respect my own.

The Confusion

Posted in Science Fiction, Science!, Things You Should Read on January, 2009 by melendwyr

My inter-library loan request finally came through with a copy of Neal Stephenson’s The Confusion! Huzzah!

I’m about halfway through it now. It seems to me that the bits of modern thinking embedded in the plot are somewhat more obvious than in the previous book. (For example, I’m reasonably certain that 17th-century Armenian didn’t include phrases identical in translation to modern ridiculous business jargon, although I suppose it’s possible.) I also suspect that Leibniz might not have anticipated Gödel numbering quite as thoroughly as he is depicted.

‘Half-Cocked’ Jack, the King of Vagabonds, has just constructed a set of phosphorus bombs to facilitate his breaking through the raiders blocking the trading routes. Will he ever regain his alchemically-prized cargo of denser-than-normal gold (possibly an isotope)? Will Eliza regain her stolen child?

Only one way to find out…

Cleaning House

Posted in Doom on January, 2009 by melendwyr

I found this script tucked away on my hard drive this morning. I’d forgotten it was even there. I doubt it will ever be used, so I’ll save it here.

With its deletion, my computer no longer contains anything LRR-related, except one of the few remaining copies of Graham and Paul singing “Still Alive” from last year’s charity gaming.

I think I feel good about that.

The following would make more sense if you were familiar with the “Evil Inc.” series of videos, but I think you’ll get the concept pretty quickly.

The head of Evil Inc. (Paul) sits at his desk, which bears a small sign reading ‘Evil Inc.’. The intercom buzzes.

[Assistant’s Voice (Graham)]: “Sir, someone to see you.”

[Paul]: “Is it the lunch delivery? Because I ordered that deviled ham on wheat an hour ago.”

[Graham’s Voice]: “No, it’s Agent R. He has the package.”

[Paul]: “Excellent. Bring him in at once!”

Graham enters, followed by James, who’s carrying what looks like a DVD case covered in a thin plastic film.

[Agent R (James)]: “We intercepted the target as he left the research station. He was carrying this.”

[He holds out the case to Paul, who accepts it.]

[Paul]: “Good work, R! With these schematics in our possession, nothing will stand in the way of our plan! I need merely remove this wrapper…”

Paul attempts to pull away a flap of the plastic in order to tear it off, but it won’t rip. His fingernails can’t pry away enough of the flap to get a hold on, and there isn’t enough give to the film to shred it. After fumbling with the case for a bit, he hands it to Graham.

[Paul]: “You open it.”

Graham attempts to rip the shrink-wrap apart with both hands, but fails.

[James]: “Maybe if we both try it…”

Graham and James each grab and side of the plastic and try to pull the film apart with all their strength, but the seal isn’t even stressed. Paul reaches into his desk and pulls out a pair of scissors.

[Paul]: “Here, if I can just get a corner started, we can start from there.”

The three attempt to puncture an edge with the scissors, but it just slides off. Their combined efforts are useless.

[Paul]: “I think it’s time to try more extreme measures.”

The camera cuts to a new shot of Graham and Paul standing over a table. The camera’s looking up into their faces, making it impossible to see what they’re working on. They put on safety goggles, and a montage begins.

Graham pounds with a hammer.
Graham stabs with an icepick.
Graham lowers a buzzsaw.
Graham lowers a chainsaw.
Graham pours out a beaker of steaming, fizzing liquid, resulting in sizzling sounds.
Graham and Paul put on welding masks, with Graham applying a welding torch.

We see the case, sitting on the table. It’s completely undamaged.

[Paul]: “This is unacceptable! We need to try something more destructive. Oh, Agent R?”

James enters the shot.

[James]: “Yes, sir?”

Paul picks up the case and the pair of scissors, handles down, and holds them out to James.

[Paul]: “Take the case and these scissors to the office down the hall. Hurry, there isn’t a moment to lose!”

[James]: “Sir!”

He runs out, holding both items. Paul smirks and looks at his watch, counting out loud.

[Paul]: “5…4…3…2…”

There’s a loud splortching sound, immediately followed by a scream.

*Scene Cut*

James is sitting in a chair, holding the case. A square of gauze has been fixed over his right eye with medical tape. He’s obviously in quite a bit of pain, but nobody cares. Paul grabs the case from him and examines it.

[Paul]: “It’s completely undamaged. Curse you, shrink-wrap! Curse you!”

The shot changes to another, almost identical office. The only differences: the sign reads ‘Malice Ltd.’ and the head of Malice Ltd. (Matt) sits behind the desk. Another assistant (Bill), dressed identically to the first, enters the room.

[Bill]: “They were unable to penetrate the monomolecular film, sir. The case is still unopened.”

[Matt]: “Good, good. Everything is proceeding as I have forseen.”

[Bill]: “Also, we intercepted the target as he entered the building. He was carrying this.”

He holds out a brown paper bag to Matt, who takes it from him and opens it, looking inside.

[Matt]: “Excellent, deviled ham on wheat!”

He begins to eat the sandwich.

The Other Shoe

Posted in Doom with tags , , , on January, 2009 by melendwyr

From this comment thread at OB:

[bored now. comment deleted.]

Comments under that name are no longer permitted. The transition is now complete.

For the record, O Moderately Beloved Readers: has Eliezer yet demonstrated that recursive self-improvement is possible, let alone likely? Has he shown that ‘Friendly AI’ is coherent, let alone desirable? Has he shown that the only alternative to FAI is a nightmare or a null? Has he made any objective progress towards his claimed goals? Has he been able to objectively demonstrate the validity of his assertions about rationality?

You’ll have to ask those questions yourselves, now.

[edited to add] Eliezer’s replacement text is probably a reference to Buffy the Vampire Slayer.

Specifically, when Willow is consumed by rage and despair after the shooting of her lover, she tracks down the people responsible. After capturing one, she toys with him for a time, then says “Bored now”, and kills him by magically ripping off his skin.

Removing the Head, or Destroying the Brain: Qualia, Part III

Posted in GIGO on January, 2009 by melendwyr

As discussed previously, there is a problem with the concept of qualia, roughly analogous to the problems with the concept of the nonphysical.

Thing is, though, it’s an even worse problem. We could, if we wished, call one limited understanding ‘physics’ and everything outside of it ‘non-physical’; there’s nothing forcing us to create a more-inclusive system that encompasses both the limited understanding and things beyond it. There’s really no reason to do that, and it preserves the concept of the non-physical through purely artificial means, but it can be done.

With ‘communicable’ and ‘incommunicable’, though, this cannot be done. If taken rigorously, ‘incommunicable’ does not merely mean that you or I cannot express a thing properly, but that it cannot be expressed at all. This is a very strong claim with very strong consequences. There are profound difficulties with the very idea of incommunicable experiences.

What is the best way to address those problems?

***

Frogs are interesting creatures, most especially because their nervous systems are capable of much greater regeneration than the mammalian equivalents. Things can be done to frogs that effectively cannot be done to human beings.

For example: if you sever a frog’s optic nerve, the nerve can grow together again and function. There is an interesting result if you sever the nerve, rotate the eye, and let it heal itself. If the eye were rotated 180 degrees, top to bottom, when the frog’s tongue leaps out to strike a target, it will be directed 180 degrees away from where the target is located. ‘Up’ and ‘down’ are hardwired into the frog’s nervous system.

Human nervous tissue doesn’t regenerate that well, so the experiment can’t be reproduced precisely with them. But humans can be persuaded to wear goggles that, through configurations of mirrors and/or prisms, turn the incoming visual images upside-down.

While there is an initial period of disorientation and improper movement, in time humans adapt and adjust. They can move and react to things appropriately, in precisely the way that the frog cannot.

When people learn that the image that falls on the retina is naturally oriented 180 degrees from the ‘real’ configuration of the things we’re looking at, they sometimes wonder how we can react appropriately. They intuitively believe that since the lens’ projected images are upside-down, our experiences should be as well. The truth is harder to grasp: our brains are computational systems that process data. The light on our retinas isn’t ‘upside-down’ as far as that processing is concerned.

Frogs reflexes are hardwired in such a way that specific nervous signals are associated with specific directions. Human physical movements aren’t. So how do we manage to associate patterns of input with specific directions?

Look at babies. They clearly get excited and try to respond to things, but all they end up doing is wriggling violently. There’s no built-in system of relationships between their intentions and their movements. When they want to reach towards something, they seem to bombard their motor centers with random signals, resulting in random movement. With enough feedback (visual, proprioceptive, etc.), their brains establish relationships between wanting to accomplish some movement-related goal, and directing the body appropriately.

Humans are at the complete opposite end of the spectrum to frogs: instead of having a built-in system, we generate one out of feedback between two subsystems.

Removing the Head, or Destroying the Brain: Qualia, Part II

Posted in GIGO on January, 2009 by melendwyr

Take a closer look at the Wikipedia discussion on the definition of ‘qualia’.

Certain terms are so poorly defined that this can become an anti-criticism feature. No matter how thoroughly one meaning is shown to be nonsense / invalid / useless, people can just switch to the ‘true’ meaning of the word and go on as before. But there are certain regularities I’ve seen in the way people talk about ‘qualia’, and many of them occur in that discussion. What are the regularities and their implications?

Consider ‘ineffable’. If something is ineffable, there’s very little else we can say about it. One of the few things we can say is that an ineffable thing can’t serve to explain anything. It doesn’t dispel confusion and lack of understanding; labeling something as ‘ineffable’ is just a sign not to think about it. It’s a shorter way of throwing up your hands and giving up.

That’s not necessarily an incorrect response. But it means that you’re giving up.

We must also ask ourselves how the ineffability of the concept can be demonstrated. Are we merely asserting it after trying, and failing, to understand? Or can we rigorously show that understanding is impossible?

‘Incommunicable’ is even more troubling.

A very common error is to think of our minds and brains as singular, indivisible entities. They’re not homogeneous things. They consist of parts. For example, we have two hemispheres, each of which can function independently, although hemispheric specialization limits this. We have systems and subsystems that handle particular types of tasks, and those systems are built up out of individual neurons.

Our minds are nothing but communication. Communication between entities, on many different levels of implementation.

So how do incommunicable experiences enter into this entity composed of communication? We can’t remember such an experience, because ‘remembering’ requires encoding. And if we look at increasingly short-term memory, eventually we are forced to conclude that we can’t ‘remember’ the experience even as it occurs. The only aspect that could actually be processed is a simple signal that ‘qualia’ are being had. There might be different signals for different qualia, but those signals would be communicable — and how would they be tied to ‘different’ qualia, anyway?

Truth and Absolutes

Posted in GIGO on January, 2009 by melendwyr

Those of us trying to understand the crazy world we live in spend a lot of effort sifting through assertions of truth. There even more people spending a lot of effort forwarding their favorite assertions and trying to get people to accept them as true.

At some point, it occurred to me that trying to determine what’s true, in an absolute universal sense, is a waste of time. If you roll blocks covered with letters of the alphabet, you’ll eventually produce syntactically valid assertions. And while it’s pretty unlikely that you’ll stumble upon a true statement by searching randomly, they might be.

But, so what?

We have no particular reason to accept randomly-generated statements as true; nor do we have any reason to assert that they’re false. They’re there, that’s all.

Likewise, we can reach a conclusion through careful analysis of the totality of available evidence and accept it as provisionally true. But later evidence might later show this ‘truth’ to be false?

Again: so what?

Most people have an unhealthy craving for absolute certainty, things they can rely upon to be true without the possibility of doubt. I suspect this is a shortcut to reduce the cognitive overhead of thought. We don’t actually like thinking; research shows that for the vast majority of Homo sapiens, engaging in mental effort is strongly negatively reinforcing. We’ll go far out of our way, and do a lot of extra work, to avoid having to do mental work.

Instead of worrying about what’s true, we should focus on what we can justifiably claim to be true. It’s the justification that matters.