WTF? You imagine some world where human-driving a car will be banned, but human-driving a motorcycle won't be? And you also think that illegally keeping your high-beamers on is somehow justified because of stupid drivers that won't notice them?

I go one step further and have all the hashing done client-side!

Oh, I know, right? Just the other day I was skimming over our plaintext password file, and it was just filled with crap like that! I mean, they think they have these brilliant passwords, and yet they also complain about their accounts being cracked all the time!

Right. Increasing population plus constant money supply equals decreasing money-to-population ratio. No shit. Didn't need an elaborate example to communicate that.

I was just expressing the standard definition of deflation, which doesn't make reference to population ratios.

You think that a declining money-to-population ratio is "as bad as" deflation? Fine, but it's orthogonal to the point.

Sure, I'll try to give you the intuition behind it.

Let's say I have a belief ("theory") that tries to describe/predict how the world really is ("reality"). My theory is sometimes going to be right, and sometimes going to be wrong. Sometimes it will predict an event to happen more often than it really does; sometimes less.

The KL divergence is a measure of how far my theory differs from reality.

The specific way that it quantifies it is basically by asking, "If I go around expecting the reality to be just like in my theory, then how often, and how badly will I be surprised?"

"Surprise" is itself quantified in such a way that if you assign a *low* probability to an event that *does* happen (or vice versa), that counts as "high" surprise. Likewise, if you predict a high probability for something that does happen, you get a low surprise-value. (The math involves taking the log of the inverse expected probability of the event.)

So the KL divergence goes through the world's "probability distribution", and at each point (i.e. experience all events at least once, but with differing frequencies depending on how probable the event is) it compares to my theory's probability distribution, finds how surprised I am, and sums over the whole thing.

The result is a measure of how bad my theory is (i.e., theories that are often wrong get a high KL divergence from the world).

Make any more sense now?

As for the joke: It's saying that information theory is the theory, and any deviation of the world from it is the KL divergence.

There are cryptosystems based on the knapsack problem, which is also NP-complete. (The basic idea is that your private key is a super-increasing set like 1,2,4,8... -- a special case for which the knapsack problem is easy -- but you disguise it via modular arithmetic as a plain ol' hard-as-fuck knapsack knapsack problem)

But like with (what you say of) the IBM/TSP cryptosystem, they always seem to get broken -- not universally breakable, but commonly-breakable enough to fail crypto standards. I can't remember the reason why they keep failing though.

Er, not quite. When you lose the private key to a bitcoin public key that has received (net) bitcoins, those bitcoins are gone. You would have to somehow infer the private key from the public key, which is by design impossible to do in feasible time.

There is no mechanism for getting the network to "re-credit" you your lost coins. No amount of whining about how this or that HD crashed or computer was hacked will allow you to regain access to the credits. At the very least, it would require a full fork and migration of the entire blockchain (i.e. global ledger), with full cooperation, to cover up your little oopsie ... which isn't going to happen.

As always, of course, if you have the private key(s) backed up somewhere, you can go right on re-spending like nothing happened.

The definition of deflation is *decreasing* money supply, not stagnant money supply.

It's just that our current world is so hopped up on pro-inflation propaganda that even a failure to sufficiently pump up the money supply in perpetuity is, to some, "the same thing as deflation".

Pardon if this is pedantic, but that overstates it, as you next sentence suggests. Bitcoin doesn't remove third-parties, as the network *is* the third party. Rather, what it removes is the need for a specific third party (the "network in general" serves this function), or a 3rd party that needs to know specific details about you.

Fair enough, if you mean that an article about Bitcoin should give some background on how it works, which will probably involve something about its money supply growth.

But the OP wasn't merely suggesting that: he said it should basically give weight to one particular criticism, viz. that the bound on money supply growth will be a "deflation problem". My point (drenched in sarcasm) was that if it did that, they'd have to cover so many criticisms they'd never get around to discussing the "new" in "news".