So I've got the bug. This happens when I have a little downtime in between gigs.
So I'm thinking about Fintech and banking and what it would take to build up a new kind of bank from scratch. So here's what I know so far. Nothing. That means I'm going to have to read a couple books. What I think is quite interesting is that I am not aware of how full stack web developers design their data models. That may be a big opportunity, because I think a lot of the disintermediary types generally don't have the patience to actually design transaction systems. But I don't know this yet.
So let me dump a bunch of stuff here to encapsulate what I've been thinking over the past week which has been a kind of emergence period for various aspects of my personality and demeanor. The short story is that I am closing yet another but certainly not the final chapter of a particular book of grieving. Through that, more pedestrian professional and intellectual concerns now have more running room in my psyche.
As soon as I came to understand what a distributed public ledger was, it became apparent to me that Bank Settlements was the killer app. Well, actually that happened as I was writing the word 'Nothing' just above. So from that moment I went to lookup and find Blythe Masters. She's going to make another pile. So if my instincts are right, as they were about Bloomberg, then I should probably just drop everything and find the buttpath to kiss my way into her company. Of course she's too smart for that which is why she only has job openings in Budapest. But I have stronger instincts to know that what I should do is to leverage what I know and build extensions of that knowledge. So for the moment, the only unassailed item I have is dealing with the actual application of blockchain audit.
One thing that's going to be done is processing off the blockchain. So let's forget about the building of that for a moment. When I think about the building of subject matter specific blockchains, there is going to have to be the embedded sunk cost of having enough compute power to amp up the transaction rates appropriate for that market. That is a big deal. Right now Bitcoin, the most robust blockchain system out there is only capable of handling on the order of a dozen transactions per second. Visa, for example, processes way way more.
VISA handles on average around 2,000 transactions per second (tps), so call it a daily peak rate of 4,000 tps. It has a peak capacity of around 56,000 transactions per second,  however they never actually use more than about a third of this even during peak shopping periods.
So before Bitcoin can take over the world it has a long way to go, some fraction of which is impossible to accelerate due to the very nature of mining blocks. The Bitcoin miners are now capable with 'all' that compute power to maintain a transaction rate of about 12 per second. That is only 10% of what PayPal's capacity is. Visa can run 2,000 tps all day and their capacity is somewhere around 56,000 tps. So for all you people claiming how BTC is going to take over the world, it has to grow 2,000 times as large just to beat VISA. So the question is, who is going to invest the money to make the top 10 mining pools 2,000x as powerful?
The very security of BTC comes at the expense of compute power. It will always cost 10x as much to do 1/10 of the work of the traditional systems. So I think you'll find that mass market retail transactions are not in the future. Bitcoin will always be inefficient, and it will always spend more money for 'perfect' trust. Believe me, less than perfect trust will gain more market share than libertarian dreamers assume.
The question is wether or not there will be tiers of blockchains. That would be an interesting development. There is something called a 'sidechain' but I don't understand that yet. I have a feeling that AWS is going to come up with something mind blowing in the next year or two. Again because I don't see the compute power of independent pools of Bitcoin and other crypto miners approaching the scale of what Amazon builds on a regular basis.
Bottom line is that the magic is expensive, complex, and tied in with a lot of misinformation, poor associations and hype. I think that there will be economies that should and will sustain blockchains on their critical path, but I suspect they will be in the exact same realm as the 'Wall Street' arrangements that are so hated now. In other words, not every banker is a Swiss banker, but on the other hand, given that Swiss bankers offer 'perfect' security, not everybody can afford a Swiss banker.
But here's where I really come down. Disintermediation is hype, because in the end, people want to deal with people. So it's fine that I have this magical technology in my pocket. People still will want to deal with my face. Why? Because I am an expert in understanding the things people want to do, and therein is the value of my service. Not the gear behind me that helps me perform in a cost-effective manner. So I will not be disintermediated. I will improve the operations of my very human centric business. The counterexample to note goes back to the now famous quote that 'Trump is what you get when you make a candidate out of the comments sections'. Big newspapers have been 'disintermediated' by social media, but social media is a rock bottom lowest common denominator of civility precisely because it doesn't have old experienced newspaper editors. The quality of information in Encyclopedia Britannica is still in fact better than that of Wikipedia. You are still getting what you pay for. Crowdsourcing < Expertise.