Tuesday, May 30, 2017

The physics of an intelligent system and the Bitcoin USAF debate

Summary

mining POW = dQ/dt/dV

infrastructure cost of transaction confirmation rate = sum (sum (dE/dt/dV) dt ) dV

where
dQ = heat energy released = release rate of entropy at a given temp

dE = potential energy of system (infrastructure) = reduction of entropy in sum of space-time, if temp were constant

mining = "demonstrated intelligence of an individual" = competition
confirmation rate infrastructure = "total intelligence of the system" = strength of society

and where there is a government (protocol) that constrains and guides the first towards creating the second. Bitcoin's POW should be the first, but all that energy is being wasted. I totally disagree with the philosophy that all that energy is being well-spent in deciding who updated transactions first.

I wonder about love and hate: There can't be a reduction in entropy without an emission of it. Love requires emission of hate. Intelligence requires emission of ignorance. It required a lot of hate and ignorance in the past in order for mammals to have love and intelligence.

I'm working on trying to create a check-and-balance protocol where the interaction between users, nodes, and miners would create a constant-value coin. Users send a market price signal to miners by paying for coins, nodes send a market price signal by demanding fees, and I think maybe miners could send a signal to nodes by being able to invalidate a percentage of their fees. Maybe nodes could send a signal to miners by being able to change coin release rate (via coins per block or difficulty).

Derivation
I came across the above while pondering the possibility Bitcoin's USAF debate. The rest of this post is is a message I sent someone. It shows my thinking process that lead up to the above and mentions several different parallels.

There seems to be something deep in the debate. "POW" rewards the selfish individual and "USAF" thinking is like community trying to work together for a greater good. It's like capitalist ideology verses socialist ideology.
POW discovers a fact via might is right. It's not a consensus. Consensus is majority is right, even if 90% are weak and even if it is a bad idea. Might is right can also be a bad idea in the long run.
Intelligent consensus results in rules that maximize GDP of the system and median GDP of the individuals over the long run. POW, like marketplace competition, solves the question(s) that the consensus asks it to solve, by killing the losers. Killing the weak seems brutal, but it results in higher total GDP and higher median GDP. The individuals and the system become stronger. "Individual" could be an idea, a coin in a sea of coins, a business, or a person.
Fuzzy definitions can point the way to more fundamental axioms. so bear with me if some of these equivalences are an offense to a mathematical mind. See the quotes at bottom concerning mathematical rigor. Anyway, first bear with me and consider the very rough approximations correlations I have in mind:

POW=competition=individual greed=negative sum=capitalism without governing law=anarchy
"USAF"=consensus=cooperation=system-wide benefit=positive sum=socialism=democratic law

Ignore "USAF" if you don't think it's related to consensus.

There is a key to life in this. It's like determining when we should be selfish and when we should be "loving" (or "working together"). We should work together when there's an agreement on "what's best for everyone" and be selfish when there is a dispute on the protocol. It's the core of ideological disputes.

POW is not actually work because it's based on finishing first so it is proof of work / time = power (which is "strength" or "might"). Saying they have to expend a work to solve it is ignoring the time to solve is not a fixed, but depends on competition. POW is actually "proof of most work, faster than the others". So it should be called proof of power POP. The capital expenditure to get the power is an energy expense that creates a "potential energy" but the kinetic energy expense (electricity) per time at an assumed standard efficiency is what proves the power. Nodes also have a capital expenditure but they do not have to demonstrate power. So nodes are more directly related to a potential energy.

So my equivalences can be extended:

POP = kinetic energy spent per time = cancer hogging glucose
Nodes = potential energy existence = Amazon.com reinvesting all profits

To be more precise, the kinetic energy I'm talking about is converting potential energy into heat. So it's really a heat energy.

But each worker inside Amazon has to demonstrate POP in a way that is not cancerous. POP is an kinetic energy expenditure at a local point in time and space to arrive at a decision that does not care about the system-wide (in all time and all space) potential energy. You are asserting POP can figure out what's best for the system, but I do not believe it is true. You are saying USAF is being forced on everyone (force is a directed POP because time=i x space and P=F x space / time), but my perspective is that it's very gentle (non POP). That does not prove nodes (or USAF) has an Amazon-like intelligence that maximizes non-cancerous POP and therefore maximizes future potential energy of the system.
The above considerations result in two axioms that tie physics to social institutions.

capitalism = POP = selfishness = derivative of local heat energy generated = dQ/dt/dV
democracy = consensus = love = integral of system-wide potential energy stored = sum (sum (dE/dt/dV) dt ) dV

Via relativity, a space-time contortion is energy, so the above relations are fundamentally unitless, so it may be just a different way of saying: capitalism seeks higher rate of entropy emission democracy seeks maximal entropy reduction There can be no entropy reduction without an emission of it. Bad ideas must be released before good ideas can be retained.
So there is an important ying and yang here that I can't quite write a feedback equation for in order to design a protocol that implements it. Unchanging laws should seek system-wide long-term love by advocating constantly changing local instantaneous hatred. That is the socialist's view point. The capitalists viewpoint is the same, but it reverses time: Constantly changing local instantaneous hatred results in (no "advocation" needed) unchanging system-wide long-term love. The socialist believes we should be our own God. The capitalist either believes there is no God needed, or that God (nature) will take care of things without our help.
Continuing with the topic at hand:

Quotes against mathematical rigor: “The object of mathematical rigor is to sanction and legitimize the conquests of intuition, and there was never any other object for it.”
"All physicists and a good many quite respectable mathematicians are contemptuous about proof."
"It is obvious that methods of the professedly rigorous mathematicians are sadly lacking in demonstrativeness as well as comprehensiveness."
"Those who may prefer a more formal and logically-arranged treatment may ... do it themselves."
Many non-math Wikipedia articles have been ruined by mathematical rigor.

Monday, May 29, 2017

All currency characteristics are derived from the goal of constant value

All the characteristics mentioned in the other answers are derived from one basic need: constant value (in both time and space). A constant-value currency is needed to enable contracts to remain valid. The "price signal" is very important in economics. A changing currency is akin to redefining the length of second (especially if "time is money"). Prices and wages are derived from some underlying contracts. Shelf price may fluctuate, but the underlying contracts in commodities and wages are trying to keep the prices stable. A limited-quantity coin like bitcoin does not help them. If bitcoin rises in value from its imposed cap, it does not mean society is wealthier. It only means those who got in first get a larger piece of society's pie, at the expense of late-comers. The fear that it will collapse will cause new entrants to seek other coins that have demonstrated constant value. Those who do not fear a collapse and get in late will get burned. ( I have bitcoin and alt coins, so I think it's a long time in the future.) A producer of things society needs wants constant value. A speculator trying to gain without work wants a limited quantity coin.

A constant value currency will need to expand as its use expands. It also needs to expand if hoarders are accumulating faster than the expansion of the economy, or if they are accumulating in a way that does not benefit society. This will decrease the value of what they hold, giving good workers more power in the economy. Defining "benefit society" and "good workers" in my previous two sentences is the goal of democracy. Enforcing the definition is the role of government.

Saturday, May 27, 2017

Transactions fees for voting = constant-value coin and blockchain=database.

Most coins would benefit letting transaction fees paid equal the the voting power in governance. The "taxes" you pay (transaction fees) determine your voting power. Paying to influence sounds like a lobby, but by acting in self interest to lower transaction fees, they will vote for the most efficient protocol. And if they're doing high volume they are buyers and sellers, not speculators, so they will vote to keep a constant value coin so their costs, revenue, and wage contracts remain valid in the coin. Even von Mises was all about price signals and a coin with a changing value undermines the whole point of a coin as a price signal. New entrants (buyers and sellers) would not be discouraged by feeling like they were cheated out of appreciation, ending the Ponzi scheme feel. The developers should be paid like the nodes, only via transaction fees instead of market place participants having to buy shares in the project. This current situation is not unlike "send me good faith money so that I can unlock your funds". If buyers and sellers decide they want more protection in transactions, then they can vote to spend the fees they've paid to become higher in order to support more governance. That's a traditional VAT tax. Really, I can't see (and Satoshi said it pretty directly) that the only new thing is that it solves the double-spending problem for P2P which seems to be just the traditional problem of deciding which approved user updated the database first. So I can only see the achievement is nothing more or less than making distributed public databases possible. Isn't precise and secure modifications rights subject to governing code (think ETH) sort of the definition of a database? Granted, "base" in "database" implies a single centralized copy so it's not a minor achievement.

Importance of alt coins creating stability in Bitcoin

People stayed in their alt coin when BTC looked high, and now they are rushing back into BTC as it looked oversold. The amount of stability this is adding to BTC is impressive. The competition between alts is determining the value of BTC. This is an oracle (an external fact a cyber-system needs to know, ideally without trusting a 3rd party) that is external to BTC. Oracles taking measurements of the physical world without a 3rd party is the holy grail of these P2P databases. They can't even determine time. Nick Szabo recently re-tweeted a joke someone made about looking at blocks to determine what time it is.

Granted, this oracle only shifts it from the perception of BTC value to the the perception of the value of the entire P2P-crypto-database ecosystem. But the new stability that came to BTC with the advent of viable competitors 2 years ago is clear. This is the first 33% pullback in 2 years ( 4 hour weighted avg) and it seems to be over. And the 20% to 30% pullbacks are only occurring soon after 30% to 50% rises that occur quickly. So traders are being penalized if they do not assume the rise is smooth. This is tremendous for BTC, especially considering the kind of news that has come out in the past 2 years.

Thursday, May 25, 2017

Bitcoin and alt coin prediction

Since it was discovered that those in a prominent position against Segwit were using the ASIC cheat, and because of the SEC review of its decision, some faith has returned to BTC. Up until now the Alts have kept pace, even though they are ~3 to ~5 times the 3 month value in terms of BTC. The current rise is due to restored faith, but now we still have a market cap in Alt coins that is greater than BTC, and everyone is looking at something like at least a 5x 1 year return in their alt coin in terms of BTC. Historically this has caused a hard pause in BTC (in terms fo dollars). So you might think $2700 now is the limit (it was trading $500 most of last year). But look at the 1 yr return big alt coins have in terms of dollars, something like 25x. This changes lives and even extreme investors like the alt group are likely to exit. But alt investors will not go back to dollars and have no expectation $2700 is the long term limit. I lightly predict they will go back to BTC for safety and stability until their alt coin is back to something like the 3-month BTC ratio. I would not bet on a hard BTC top until ETH is at least 1/3 its current BTC ratio. Maybe at 1/3 I'll get back into ETH and I'm waiting for ZEC to return to 1/2 the current ratio. Finally after buying ZEC for the first time at $33 and selling at $237 I've broke even from the ZP and cpu-mining disasters. The worst was ZP because I used BTC without replacing it.

I'll sell some BTC for alt coins when it reaches $5000 (if they do not keep pace with it like I am predicting) or after they crash a lot faster than BTC.

It seems like alt coins have added an enormous amount of stability to BTC. This is the first doubling it's had in less than 6 months in the past 2 years. It dropped a little more than 30% only once (very recent) in the past 2 years.

My BTC sell point for dollars over the next 3 months is $10,000 per BTC. That is the short-term "crazy" point, unless there is some wider-population driver I'm not aware of. But I can't argue against anyone selling now at $2750 and it's what I would recommend to any friend who likes to play it safe and wise.

Friday, May 5, 2017

Relation between Entropy and information


This is sort of a personal note that is probably not as useful or interesting as my past posts on this.

General relativity shows mass to be some sort of convolution of space-time. This is trying to see if the convolution can be viewed strictly as the presence of information, without needing to have a concept of mass or space-time.

S/(kb x ln(2)) = number of Yes/no question needed to determine what state a physical system is in.

Temperature x ln(2) x kb = Joules of heat per particle per yes/no question that determine what state it is in.

The above two statements satisfy Q = S x T.

If Joules are given to a single particle, and if the kb is simply converting from temperature to joules (so that it is not needed when working with joules), then the minimum number of yes/no questions needed (the most wisely chosen, splitting the remaining volume in half at each step) to determine where it is in a square box of volume d^3 is

ln [(p x d)^3) / (1.46 x h)^3 ] / ln(2)

where p = SQRT(2 x mass x Joules).

1.46 x h = hbar/(2 x sigma^2) from uncertainty principle. (the 1.46 may have some error)

Let's say all the joules came from a mass so that a photon exists in the box instead of a particle. And let's say we know it is somewhere in the box in time interval t.

Using meters= i x c x seconds and p=h/wavelength for a photon, the number of yes/no questions needed to position a photon in the space-time box is:

ln [ ( (h/w x d)^4 x ic / (1.46 x h)^4 ] / ln(2)

ln( ic / (1.46w/d)^4 ) /ln(2)

where w = hc/joules and d includes an extra meters to account for a box in space-time.

ln(i) = 1.5708

gives

y/n questions = {1.57 - 3ln(c) + 4ln(joules x d/(1.46hc)]} / ln(2)

Being less precise now:
ln(1.46^4) =~ ln(i). I believe there is some error in my 1.46, so it could be exact. So approximately

yes/no questions = log2 (c (d/w)^4)
or
= log2(c) + 4log2(d/w)

where the log2(c) is just a units conversion that should be log2(1) = 0. hc might be viewed as a conversion from 1/joules to meters. But anyway, I have
y/n questions = log2(d/w) per dimension which in hindsight is obvious. It's sort of the definition of the w.

To give a feel for what I might be trying to say: this is the amount of information needed to describe the state of energy (and therefore mass) in a space-time box. The lack of h is interesting because it's usually the starting point for quantifying the number of states. I let joules and hc cancel each other. Reversing the logic: if I have the ability to perceive a certain number of bits, then this is how much mass and energy I can perceive in a certain amount of space-time.