tl;dr: Optimal strategies for the prisoner’s dilemma evolve in a cycle, with nonviolent strategies being optimal at a particular stage in the evolution of the game.

Advertisements

Our Universe: A Prison

August 16, 2013

There are reasons to suspect we are living in a simulation, some empirical, and some logical. If that is true, we might conjecture about the ethical qualities of our overseers, as well as their competence. It’s possible our universe has simply been forgotten, a scenario that brings to mind the final scene of Ran Prieur’s novel, Apocalypsopolis:

But now he was drifting upward. The rock dwindled and the stars…

Were they getting closer? Yes! He could see now that they were just tiny holes in a great black shell with supporting girders. He caught onto one of them and pulled himself along until he came to a protruding black handle, a crank. He turned it, and a great seam of blinding light opened, and he slipped through.

He was in a bedroom, in a corner. Below him was the Earth, the size of a basketball, on a little platform, with the black perforated shell over the top. A label said Planetarium.

Over at a desk was a kid, a twelve year old boy with glasses, playing some kind of computer game. Sirach stepped out of the corner, over some dirty clothes and boxes, and the kid turned and looked at him. “Who are you?”

“I just came out of the Earth.”

“The what?”

Sirach pointed at it.

“Oh, that thing.” He looked vaguely guilty.

“Is it yours?”

“Yeah, I guess so.”

“And you’re supposed to be taking care of it?”

“That’s what my dad says too.”

“It’s dying.”

“Yeah, I thought it’d be dead by now.”

“You just left it in the corner to die?”

“Sorry. I took care of it for a while. After I built it up to dinosaurs, it got boring. Dinosaurs are pretty much all that model’s good for.”

There are some problems with the simulation idea, however. First, there’s the assumption of substrate independence, which Bolstrom derives from computationalism. Judging from the past failed metaphors for mind, the outlook for computationalism isn’t bright. It’s apparent that whatever technology is viewed as most advanced is always the darling metaphor, but that only tells us about our own society’s level of development. To derive mind from symbols is probably a completely backward way of thinking, although I don’t presume to be certain. But in linguistics, in which I do know enough to judge, it’s a complete fail. Grammaticality is a matter of degrees. People seem mostly aesthetic, balancing the yin and yang almost as if their brain were composed of some sort of network of “cells” summing potentials stochastically. Which state could arguably also be implemented in a computer, but Bolstrom appears commited to computationalism.

Another problem is that, if we assume the third case is true, that we are in a simulation, it immediately reduces the probability that we can ever create a simulation, which increases the probability that no post human civilizations exist. This is because if sims could create sims it would place resource constraints on the base computer. Furthermore, being in a simulation would at least call into question our ideas of physics on which the idea of computability rests. If holding the proposition reduces the likelihood of it, it’s probably wrong.

Then there’s the very significant fact that, as mammals who live on the planet earth, there’s no reason to believe that we will ever be creating giant computers and attaining any sort of “posthuman state”. It’s just highly unlikely. Species tend to be rather limited by what the biosphere provides them.

Let’s consider the energy required for a simulator that actually runs 10^36 operations per second at room temperature. Not being a scientist, I’m going to assume for the moment that one Landauer computation is equal to one computation in Bolstrum terminology. Using Landauer’s principle and multiplying by the number of operations per second that Bolstrum suggests, we get a per second energy consumption on the order of petajoules per second.

1.38*10^-23 * 0.69315 * 298.15 * 10^36 = 2.85 * 10^15

This would require the energy of three Tsar Bombas, released every second, for as long as the simulation runs. Even given the possibility of time compression inside the simulation, the total energy consumption remains gigantic. It’s telling that no one that I’ve seen in the discussion seems to have noticed this as relevant. But it’s hard to evade the consequent fact that no naturally evolved species– even a tool making ape– will ever reach such a post-human state.

The greatest objection, however, is that creating an ancestor simulation would be so unethical that even humans would decline to do it. This creation of new sentient beings is not necessarily a natural thing, and so it will be scrutinized the way that cloning is today, or the way procreation ought to be. It would be apparent that all of the advances they have made will not be available to us. Just as in the same way we could not collectively bring ourselves to simulate the stone age, they would not be able to collectively bring themselves to simulate our age.

Very expensive propositions pursued collectively are usually pursued because of some enemy– a negative inducement– not to create new people. Even panspermia has not garnered much traction. No motive is in fact given for why our post human civilization would want to do this. It would be wrong to compare it to our motivation to colonize space, because that case makes use of the drive for conquest and exploration. And empires are forged on the basis of self interest. But if we built a simulation, we would conquer nothing, learn little, and inflict much pain.

There is one situation in which we pursue goals in which we conquer no space, learn little, and inflict much pain, however. Our universe could be being used as a kind of penal colony, for pre-existing beings in the base universe. Provided there are enough prisoners, and keeping us alive in physical form turned out to be more expensive, mind-uploading could be pursued to both safely kept us at bay and ensure that we undergo some sort of life-lesson. It is also likely that if we developed simulations (at least, simulations in which the beings are amnesiac regarding any possible previous state, as we are), these would also be prisons.

Because of the constraints on collective behavior, I consider imprisonment to be the almost certain conclusion if the simulation argument is true. If that is the case, we might hypothesize that something like karma does exist, with each of us being punished for what was committed in the past. This would raise further questions of whether we will be here repeatedly– until the lesson is learned– or whether our time in the simulation is a death sentence.

All take advantage on miserable man. The world deceives him, life flatters him, fortune derides him, health forsakes him, his youth passes, felicity withdraws, years fly away, and contented ones come not, time runs, his life ends, death catches him, the grave devours him, the Earth covers him, putrefaction dissolves him, oblivion swallows him, and he who yesterday was a man, today is dust, and tomorrow nothing. But how long shall we lose precious time, let us return to our former road, our straight and direct way; for here is nothing that we can expect more, unless to entangle ourselves in a labyrinth of errors.”

-Gracian, The Critic

But there is more: justice deserts him. For proof, consider the rules regarding the so-called “lost volume seller”. To understand the background for this abomination of class warfare, you should know that typically, under the Uniform Commercial Code, when a buyer breaches regarding a sale, and has not yet received the goods, the seller can recoup the difference between the market value and the resale value. A reasonable, well-rounded rule, I should say. Unfortunately this is only the rule for the rest of us, for the commoners.

Rule for peasants: market value – resale value (a middling sum)

For those who have capital there is another rule. This rule says that, if you are lucky enough to have enough goods in reserve to fulfill any separate order, the second order will not count as a resale deduction. Instead, you are entitled to the profit of the breaching buyer, even though he has breached and you have kept the goods.

Rule for capitalists: pure profit

The rule has been controversial for obvious theoretical reasons of overcompensation, though I am not sure anyone has pointed out how it is an example of class war. In his critique of the rule, Breen restates the essence of the claim to lost volume damages:

The seller claims, however, that the resale has not made her whole. It has not put her in the same economic position as performance would have done. Had the original buyer not breached, she would have sold two boats instead of one and thus would have collected two units of profit. Because the seller has “lost volume,” she cannot be made whole by any price differential damage formula. She has lost the sale of some quantum of goods and the corresponding profit that this sale would have brought. Thus, the argument goes, the “relevant characteristic” is not the “standard pricedness” of the goods, “but the fact that [she] will lose one sale.”

Now, the stated purpose of contract damages is to put the non-breaching party in as good a position as he would have been in had the contract been fulfilled. For a “volume seller”, this means restoring profit, under the present system. For a poor person trying to sell something, this means restoring the difference between market price and resale price. Why two standards? The logic is that the capitalist has enough in reserve to claim that he would have made the second sale anyway. Therefore, again according the Breen, the definition of the lost volume seller has traditionally been something like this:

Harris proposed that a seller who resells finished goods could qualify as a “lost volume seller” only if three conditions were satisfied: “(1) the person who bought the resold entity would have been solicited by plaintiff had there been no breach and resale; (2) the solicitation would have been successful; and (3) the plaintiff could have performed that additional contract.”

The key element of the definition here is the last:The plaintiff could have performed that additional contract. Never does it seem to enter into consideration that the person who meets this definition is in possession of excess assets which affect the moral claim to extra damages, compared to the single-item seller.

To get a picture of this, consider two series of sales, one in the breaching case, and one in the non-breaching case. Each number is a different possible sale. The breached sale gets an asterisk.  The sales in brackets are for identical goods.

No breach:
1, 2, 3, 4, 5, 6

Breach:
[1*, 2], 3, 4, 5, 6, 7

In the non-breaching case, the seller gets six units of profit for six sales. In the breach case, they are lucky, in that they get seven units of profit for six successful sales. It should be obvious that this scam is better than fractional reserve banking.

Note that lost volume seller is able to perform the additional contract, number 2, in both sets. However, they do not get to perform number 7 in both sets. But number 7 is not considered the “resale” of the breached first sale, so Harris’s conditions still apply for the second sale. Because the definition involves only a direct comparison of the first two sales, the lost volume seller gets compensated as though the price differential would not work. The definition applies because it limits the resale to the immediate present. Because the breach of sale 1 and successful sale of 2 refer to the same items, it is clear that the contract could not have been performed with those same items, but for the breach. But because the breach did not make possible the fulfillment of the second contract, which would have occurred anyway, then the breach is not considered a cause of resale. Sale 7 does not enter into the equation.

To return to the single item seller, if they resell it is deducted from the contract price of the first sale. Therefore they only get profit for the second sale, not the breached sale.

No breach:
1
Breach:
-1, +2 + (2-1)

But in the case of a rich person, breach by the buyer brings the windfall of an extra unit of profit.

It is clearly unjust that those most capable of mitigating the effects of a breach should be in the position to receive the most benefit from it, whereas the person who has only one thing to sell receives a deduction for mitigation by resale. It is equally wrong to not recognize that the contract is a risk-allocation device, and that the risk that the single item seller took was greater than that which the lost volume seller took, such that the one item seller has a greater claim to damages than the lost volume seller. Furthermore, the case has been made that this is overcompensation, as evidenced by the extra profit:

The award of “lost” profits to the volume seller does exactly that – overcompensates the seller. It does not put the seller in the same economic position as performance would have done, but in a better one. The volume seller who succeeds in reselling the contract goods following breach and who is not awarded the profit remedy has his original expectations fulfilled. The seller expected to sell one unit of goods and receive one unit of profit. This in fact takes place. That is, the profit made on the resale of the contract goods replaces the expected profit on the original contract because the seller’s expectation interest in that profit is the only contractually protected expectation interest at the time of breach. Put another way, the volume seller succeeds in mitigating his  [828]  damages by effecting a resale. The resale, in fact, constitutes a mitigation because the seller did not already have a contract in place protecting his expectation in the second sale. (Breen)

It would appear that we could just deduct the market value from the contract price. What logic justifies this? The critical analysis of this rule could branch out into a number of areas regarding the absence of sale 7 from consideration. This involves the time-structure of ontology under capitalism, with assets being less “real” than profit. But in essence we are dealing with two separate notions: the resale of specific goods the sale for which was breached and the provision of a sale that could not have otherwise been made. Let’s call these two concepts the resale and the mitigating sale.

To justify the lost volume seller rule, the argument is that while the resale (of the specific goods) and the mitigating sale (the sale that could not have been made otherwise) are the same sale in the case of the one item seller (sale number 2 in the above series), in the case of the lost volume seller the resale and the mitigating sale are two different sales (number 2 is the resale and the mitigating sale is number 7, which has not yet occurred). Because sale 7, the one that could not have been made without the breach, has not yet occurred or is impossible to locate, it can’t be mitigating. And in order to have an effective resale, the resale must be both a resale of the goods and a mitigating sale.

Let’s take it for granted that the mitigating sale actually can’t be located, and this isn’t just rhetoric. It still remains that the absence of the mitigating sale at present is only one factor. There are far more factors that militate for differential damages than against them. Some of these factors are:

(1) It is clearly perverse that those most capable of mitigating the effects of a breach should be in the position to receive the most benefit from it, whereas the person who has only one thing to sell receives a deduction for mitigation by resale.

(2) It is equally wrong to not recognize that the contract is a risk-allocation device, and that the risk that the single item seller took was greater than that which the lost volume seller took, such that the one item seller has a greater claim to damages than the lost volume seller.

(3) In the long run, this rule clearly overcompensates the seller for the breach, as can be seen in the comparison of the two series. To put the seller in the position he would have been in if the contract would have been in if the contract had been fulfilled cannot mean giving him his profit but letting him keep an extra item of goods for sale. The position he would have been in would be to have one item of profit, absent the goods. The capital itself should not be allowed to “fly under the radar” of the rule, simply because the ontology of capitalist entitlement makes a hard distinction between assets and profit. The unsold goods which would have been the resale goods, but became sale 3, are a quasi-mitigating nonsale.

(4) The buyer takes no contractual responsibility for the seller’s profit. This is not mentioned in the terms, rather the contract price is mentioned. This is also the essence of Breen’s claim.

(5) There is no conceptual justification for separating the mitigating sale from the resale of the same items. The naïve conception of the single-item seller is no doubt that the resale of the specific goods constitutes the mitigating sale. And why should it be any other way? Why should this be different for a seller who simply happens to have an inventory of other items like the one that was actually resold?

(6) There is no reason to distinguish between a person who has one item for sale and one who has many, for the purposes of allowing damages, nor there is any reason to discriminate against someone dealing in multiple types of goods as opposed to a single type of good in greater number. The difference in the two rules is therefore also unfair.

(7) It provides a perverse incentive to allow breaching rather than to come to terms by mutual agreement

(8) There may be mathematical problems in the idea of the lost volume seller being certain of make the sale either way. If sales are certain, than breaches are equally certain. If the same number of profit units would have been made either way, certainly the same number of damage units would have been accrued either way. Thus the person who breaches has no reason to pay for a breach that would have happened either way, which belongs to a class of volume breaches that are certain to occur. This is tantamount to holding the breacher responsible for an act of God.

Clearly, in light of all these factors, the mere absence of a mitigating sale seems rather insignificant, since the resale in fact is present already. I believe the unfairness of the rule is indicative of problems that are systemic. The only way to stop the power of money is to stop quantifying future-oriented action chains and then ranking them based on our own neurological limitations, as these will come into conflict. This can possibly be done by transhuman modification to attain a Borg-like state, but is best done by smashing the bioarchy and stopping reproduction entirely.