"No one is going to do a billion-dollar transaction on the Internet."


True, and that is why the equilibrium size of firms and transactions - having evolved prior to the Internet - has been so large! Before the Internet, you had no choice but to have people in one organization talk face-to-face with people in another organization. So, to realize economies of scale, the organizations - and the bandwidth of exchanges carpooled through a single analog relationship - tended to become as big as possible.


And centers of decision-making had to be physically close, meaning centralized rather than distributed, because distributed nodes could not be coordinated via some kind of, like, intra-organizational information price, system to foster altruistic, or non-quid-pro-quo data-sharing incentives. (Actually, I once came up with a groupware paradigm called "token-packet transfer" - analogous to holes moving in the opposite direction from electrons in microchips - to solve this problem


If you increased the information bandwidth of Internet trades, the equilibrium size of transactions would go down, as their precision and frequency would rise. The telephone line was the death of Manhattan. You are saying the information bandwidth of the Internet is too narrow, and I'm agreeing with you! I'm saying the information bandwidth of the Internet can be exploded! There is no information that cannot be exchanged, and no need that cannot be met, in a price-implicit/electronic-exchange way.


If all you're saying is that you don't recognize this, or you don't know how to do it, or the whole concept is totally alien to you - and you need flesh human beings holding hands in order for two entities to dance - I am agreeing with you. There is no longer any need for the economies of scale of a monolithic billion-dollar transaction, period. Adjustments can be broken up, gradual, periodic, fine-tuned, continuous. The economic quantum shrinks with decreased friction.


(I don't even frankly understand why government-collected economic statistics are still released in such awkward burts and lumps! Though I do understand why the symmetric-distribution costs created by Reg FD - with the price for information fixed at zero - would encourage less frequent, larger bursts released by corporations. Also, it is important to note that some goods like airplanes - and not commoditized as to time and place of delivery or maintenance condition - are so large and indivisible, and so few, that they will probably subsidize two actual human beings actually talking for a long time to come.)


"4 bad traders will not make as much as 1 good one."


1 Bad trader with 3 units of information will make as much money as 1 good trader with 1 unit of information. That's what I am saying, that by placing your traders in an information vacuum, your are putting a huge burden on them to be good! Moreover, I am insisting that the quantity and quality of information can be brought up cheaper than the quality of traders - only big brokerages are failing to recgonize this possibility!


You folks just haven't got the foggiest notion how to automate without losing information bandwidth. And that is why you can't make money in any automated businesses, but only in business where the presence of human beings, to collect information, is subsidized by the value they add. In other words, flesh people basically do two things, they engineer, and they talk. In roles where there is no need for engineering, the value they create by talking isn't enough to overcome the friction they create by having to talk, and pay their salaries.


But when you automate, you lose the incidental value they create by talking. I say, automate, but in a way that collects information even better than the combination of transaction/price-implicit and talking-explicit you used to get. And since talking was usually the time axis, you need to collect time axis information which is detached and floats separately from immediate transactions - meaning "Gestures."


"Now they're in trouble and come to you for help."


Sounds like the original strategists could have been better - or at least had better time-axis visibility


"Or lets say that you think corporate bonds are a great bargain and you try to sell that idea to fund managers. They agree with you but having been burned by Enron, they don't want short term default exposure."


Here, you have constructed a situation where the ultimate decision maker - the mob or investor - is at a disagreement with you over near-term default rates. There is no need to apply the kind of manpower and customization you describe. As any software company will tell you, customization on a client-by-client basis is the death!


I would say that given you already have fixed overhead, in terms of the manpower already trained and in place, this is the cheapest way to do this transaction. But I would say the equilibrium way for this transaction to happen, given current technology, is for separate, liquid electronic marketplaces for disaggregated and building-block bets - for various strips in essence - to spring up.


That way, instead of structuring anything - or talking to anyone on the phone - you just enter into some kind of a "default swap" with an anonymous counterparty. As you take the friction costs of people having to talk on the phone out of it, the number of varied and obscure products in which independent exchanges can be cost-justified rises.


"You are not going to get pennied to death if you can provide real value added in your client relationship."


I agree 100%. But I think the point that I want to make is that, where you are getting pennied to death, you should actually be making money there, too! You could create even more value. And your problem is your reliance on individual human beings to create value, and your inability to design machines and systems that don't piss it away.


The way you are structuring it, the more machines you add per person, the more incredible volcano-islands of value in the middle of it the remaining people have to be. It is the topology of your whole information paradigm (and make no mistake, the creation of information, and of coordinated, non-redundant inventory and production decisions in response to it, is the entire creation of value) that makes dumb people less and less cost effective. It's self destruction, a slippery slope.


It's just natural that, since you lack private property in information, or at least in a way that can be automated, the value of anything that collects information woudl go down per unit, and the value of anything that uses or leverages information would go up per unit. And since the only way you can collect it is by people talking to people, you would end up with a smaller and smaller core of people who are absolute freaks when it comes to using information, and at collecting it, in terms of units of information per person.


More simply, your total units of information has gone down, or at least not risen relative to dollar volume, but since people are expensive, your equilibrium units of information per person has risen. And your collection-people per units of data collected as gone down (your units collected per person has risen), to where it costs the least, and gains you the most, for the first incrmenetal units of data collected - and the biggest-bandwidth relationships between the smallest, most freakish people are subsidized by customization-engineering which needs people anyway.


If you could get more information, for cheaper - and not give it away - dumb people would have something to do. If you could raise your incoming information relative to your outgoing information - by creating billing barriers - and relative to your total high-cost/flesh surface contact with other entities - other than by distorting your ratio of brain size to flesh surface - you'd be in business!


Sure, when we only had people, and had no Internet, the scales of things were built around the high costs of people, and of geographic communications friction. And no pricing mechanism for information arose at all, because people mingled, and shared information, as a totally free byproduct of having to talk. Nowadays, the cost of people is priced in, primarily as a measure of their engineering and friendlinees utility, but still not by measuring their commoditized-information utility - since they aren't a good way to assemble voluminous quantities of minuscule fragments anyway.


In other words, time-axis information has historically freeridden along the transaction pipeline, not price implicitly, but by transactants talking. Now, you are saying information freerides through the engineering apparatus. I'm saying, it's time for information to get it's own payment mechanism, so it can pay to inflate its own, independent propagation topology.


And while the per-fragment value of information may be low, the per-fragment acquisition cost can be made lower by automation! The quantity of information needs to be price-regulated, rather than to fluctuate haphazardly as a byproduct of fluctuating exhanges and adjacencies, which factors are regulated by, or are fine-tuned in reponse to, other, independent incentives and needs.


In other words, of course you can't afford to have a bunch of idiots sitting around, seeing as what little information they collect, your machines piss it away anyway, and their data becomes non-unique. At its simplest, it is simply a problem of rgeulating the flow of information, in profitabel acquisition transactions, so that it flows towards and can be retained at high points, and can be turned over profitably.


"more thoughts on it later."


eLROY