ave these little diagrams that people like to do for queuing models, well one you would need to use a negative binomial to be right there---but that's even wrong. They're only correct if you have no deviation; if no one defects. The Poisson is only approximate if everyone is doing what their meant to do and never changes. So we're releasing the actual mathematics and how things work in Bitcoin, which some people will understand and those who don't can put it into a computer program and a computer can simulate it. Then we start getting all these things correct. Now, the model. Emmons [?] deleted every aspect of any code he ever wrote for his model, but luckily I have some. I kept it and archived it for him because, well, he would never want the fact that he missed all these things lost to posterity. Now, when you create a model in science, you test it. You don't test your model; that's what we're all doing right now and we're testing our model. That just tells you you have a model. Who cares? If you can't run it on Bitcoin, who cares? One of the things in Slack channels I've been doing at the moment is I've been incentivizing people to do things I've already done. There is a selfish miner test run on VMs going on at the moment. Real VMs, real tests, real whatever else. And I'm funding it. That doesn't mean it's not independent. I have said you get paid, because I've put it in third-party escrow, either way. I've put bitcoin up to pay for all the VMs and all the testing and all the code, and it will all be public. So it's not just my math; I'm getting it independently verified by other people in the community. Over and over. So, Turing completeness. I've had a lot of shit about this one. Guess what: you're all wrong, it is. What you're not going to get is that you've missed it. There are actually, we have what is called a decider in Bitcoin. It follows some of the predicate logic of Gödel. Gödel, back around the time of Turing, wrote a lot of mathematics around the idea of predicate systems. If you're a mathematician, you would know this shit, but not everyone is. So Bitcoin a 2-PDA. Now, 2-PDA, all those little squiggleys, unfortunately once again, they're from a paper that is coming out. Those little squiggleys actually mean that it is Turing complete. I'm not going to explain that to you right at the moment. I'm going to say that... any recognize what that is? That's a Wolfram 110. Anyone know what a Wolfram 110 is? It's the simplest Turing-complete system. That's in Bitcoin. It was running until a few months ago. We had PSOs in Bitcoin, running. Unfortunately they starved to death---fees. So we had the first self-evolving in Bitcoin, that ran for two years. There are people that know about these now; that's why I'm mentioning them. They've been discovered. It's not just me. They're dead! Because they can't afford the fees. And I'm not going to fund a thing requires $3.00 every bloody block to keep running, it becomes expensive very quickly. When it's low fees, all sort of things can run. So, these are there. IP to IP transactions. Does no one look at the old code any more? We keep saying, "this is not meant to be there." The market database in Bitcoin allowed IP to IP transactions. There was a reason for these. The whole idea is to have a merchant-type solution. I should be able to go up and just tap something, and I don't even give a shit whether my transaction has settled properly. All I care about is going up to the merchant, and the merchant will propagate my transaction. Why will the merchant propagate my transaction? Because they want to get paid. Merchants are not going to sit there and go, "well, maybe at the end of the month I'll run all this shit." They won't, because you will double spend. Merchants have the incentive to ensure that zero-conf works right now. Why? Because they want to be paid! It's very simple. Economics makes this work. It is not security toys. It is not any of these little fluffy bits of crap that we say are really security the network. I've work in the security industry for a long time. I