 You all have to forgive me for my voice, like this. I can't read my own slides from this distance. This isn't so good. But I can read them up there. OK, I too came from theoretical physics sort of background before I became an entrepreneur, or rather, before I recognized that I was an entrepreneur before I went to university. I'd already founded three companies which might have given perceptive people a clue, but I was totally oblivious to this at all. I thought I was a physicist. So it means I like graphs. And I like presenting things. And a lot of what we are doing is easier to visualize on a two-dimensional axis than on a one-dimensional axis, which is sort of how we kind of talk about things. Difference between big companies and small companies. The difference between complicated things and simple things and stuff like this. So Charles Perot, who is a professor of sociology at Yale, emeritus now, and one of my personal heroes, wrote a book in the 70s called Normal Accidents. And Normal Accidents is about why things blow up. So for years, for more than a decade, whenever anything blew up in the United States, there'd be Charles Perot or one of his students trying to figure out why this thing happened. And so he studies catastrophes and things like this. And he discovered that most catastrophes happen because there's a component failure. Something failed and something bad happened. And it was not in any way unexpected at all. But a significant number of them happened for ways that were mysterious for all people involved. And those are things he calls Normal Accidents or System Accidents. And that is two or more systems in your operation which weren't really intended to interact with each other, did producing results that were really perplexing for the people involved and with catastrophic potential and sometimes catastrophic happenings. So that's what a Normal Accident is. And that's what the book Normal Accidents is about. And it's about this. Now, the way he's divided things up and the important thing here is you notice it's on a graph, but you notice there aren't any numbers on the access. This is not a quantitative study in rigor. This is a qualitative study on how to visualize your problem. So on the top, there's tightly coupled things and loosely coupled things. And on the bottom, there are and that's just the coupling of how the process in whatever you're doing happens. And on the bottom, we have interactions which are linear and expected or complex and unexpected. And then he just plunked in all the things that he'd studied and things like this to put them up and things like this. He did this in the 70s, so power grids are fairly not complex and unexpected. The Americans then deregulated their electricity agency. And now things in the power grid business are a whole lot more unexpected than they were when he made the chart. So, but it's just, I mean, just take a look at these things. You will discover that, I mean, there's a difference between ways universities operate. In a way, a drug manufacturing thing operates. But the important thing is tightly coupled things, mistakes in one area can easily spread to another area, cascade and get worse and worse and worse and worse and worse. So if you have a problem in a tightly coupled area, what you really need is centralized authority, real obedience, you need everybody to know what the procedures are, do them, and fix the problem in a hurry, which means you had better understand what the problem is and how they can interact. And if things are really linear, you do. So, I mean, dams could kill a lot of people, but in general, they aren't all that because the interactions are linear and expected. And so no matter how tightly coupled it is when something goes bad, you do the right thing. On the other hand, we have things like universities and R&D institutes and things like this. The interactions are all unexpected. But if something goes wrong, you have plenty of time to sit and scratch your head and say, oh, I wonder why that happened. Let's go think about it for a while. And so they produce decentralized structures where everybody has their own authority and people with specialized knowledge get to come in and say their own little bits. And they get the sort of structures that they want. And for the peril, of course, the problem is up in the corner, because where things like three-mile-long nuclear plants and stuff like this happen are where things are both tightly coupled and complexly interactive. Because they're tightly coupled, they demand a centralized structure with efficient processes, efficient in terms of command processes, authoritarian processes. And because they're complex interactivity, they demand things to be decentralized and loosely coupled. The demands are completely unreconcilable, and that's what makes these things dangerous. So if you think in terms of this, you sit there and say, you can sort of model most of the world this way. And one of the things I've been doing was modeling news sectors in global marketing and what kind of stuff there were. And I was putting on this and thinking this was terrific. And then in 2009, Charles Perle himself came up and did exactly the same thing and could I have my next slide, please? Oh, there's something there. OK. So in the real world now, someone is generic markets. That's what economists generally say when they say they want market. This is what the market is. Up until 1970, practically every business on the planet was in there, and there weren't any other sector to worry about. Lots of business, lots of competition, nothing's particularly too complicated or hard to understand. And generally, things move then up to cell four, which is concentration. This is when one of the business becomes a monopoly or they get hardly any businesses in there and they can really, really tighten up the coupling, which generally means they can reduce costs and they can lay off people if that also is good and they make more profits and things like this. And this was sort of the model. This is what businesses are supposed to do. But now, we are actually in the business of doing things that where that kind of efficiency isn't what we want. We want to be able to do things more thoroughly. We want to be able to do things that react unexpectedly in things like this. And one of the neat things we got in the web, which not everybody expected at the time, was a way to loosely couple things. So you'd have a mix here and a whole bunch of people would show up and you would be able to get things coupled together. The secret to the loose coupling here is something called modular contracting, which is as opposed to relational contracting. Relational contracting is when you're a big supplier or somebody like Airbus or an auto manufacturer, sure, and you sit there and say, good, I will control my suppliers very, very well and we'll have a really close relationship with them and we'll produce things. I'll train them really, really well to make the sort of points I need. Through the strong relationship with each other, we'll all prosper together. And that can really work. It doesn't generally work all that well in the IT department. I mean, if you're making components, you're making things that you can drop on your foot, this may be where you wanna belong as an entrepreneur. But since we're talking about IT entrepreneurship, I am telling you that almost all of the great developments and everything are in something which economists and sociologists called modular contracting. And modular contracting is you take the complexity, you put it in a package and you do all of the stuff in there and you don't let it spread. Communicate through the outside, through an interface, an API, what have you. And then you let people randomly shuffle these modules together and things like this and you can get all sorts of interesting things at the other end that you never would have expected in the first time because you don't have a linear sequence of how these things have to go in any order. You have a whole series of APIs and anybody's free to toss them together any way you like. So this is what we were talking about and if you're talking about let's improve things for the entrepreneurs, it is necessary to know that what you want to do is to improve things for modular contracting because that's where the IT business is rather than, there's a lot of how to improve your business things that are designed for different kinds of businesses and they won't do you any good. Next thing, oh let's skip that. So, and I have this but I can talk about this for a great deal but I think we would rather have more time for conversations. So what do we want to do to encourage the level three companies we've got? One, we want to open standards and normally I talk about this a lot but I think absolutely everybody here already knows that we want to open standards so I don't have to talk about this. Second, we badly need to recognize that this is what we're doing. A great deal of problem when we're trying to develop this in Europe is that we pretend that we're trying to manufacture shoes and we don't have any idea that this is what we are trying to do so we don't actually support the stuff that's actually creating the new value and an awful lot of money we waste in traditional businesses that would like to become monopolies which is not where new value is being created so it isn't going to do us any good. One of the things that, so we need sort of recognition that this is where it is so go out and spread the word, tell everybody that you know that they need to find out what modular contracting is and that it's a good source of value and that maybe we should be supporting more of this. There are some positive signs. If you've ever been involved in a research program and here are the FB7 research programs and things like this are largely in terms of creating money for entrepreneurs a colossal waste of money and because they don't go to the entrepreneurs. The typical members of an FB7 consortium and things like this are companies like SAP and large universities and things like this and there aren't any entrepreneurs there. The entrepreneurs are not working for the large corporations and the entrepreneurs are not working in the universities. They've all gone off and founded their own business so it's not too surprising that if you don't hand any money to the entrepreneurs you aren't actually doing anything to help the entrepreneurs in this particular fund-raising activity. It may be okay for your research in general but even that I've got some strong doubts with. The Your Stars program on the other hand is actually really really good because it recognizes that there are small research producing SMEs and hands them some money and this is really good stuff. And a new phenomenon we have which I wanted to mention is crowdfunding. A lot of small companies are just go off and get themselves a site somewhere and say here are all these features please give me money. And this is working and one of the things we have to worry about is the tax agencies haven't paid a lot of attention to that and they could actually kill this one dead in the water if they don't understand that what raising money this way is actually very very valuable for the small companies and it is a good thing and we'd like to encourage it. They may just sit there and say we should tax more. Let's see. There are other things that we could help in that. Better bankruptcy rules would help small businesses a lot but so I gotta stop. We still have this, we still have the discussion. There's only one other thing that we really need to do and that is that right now our schools are set up to send out MBAs which are trained to administer large businesses that aren't creating the value that we're talking about and it would not be difficult to tell our business schools train entrepreneurs but nobody seems to be doing that. They seem to think that the skills that it takes to administer a large company are applicable to entrepreneurship and they're not, they're really not.