Thanks for this, and I hope it doesn't go completely unnoticed in the Steemit echo chamber. I think these are questions that are really worth asking! I think something that probably compounds the problem is that there seems to be a really deep misunderstanding among crypto nerds about some fundamental facts about decentralization.
I hear way too often that "decentralization is more efficient than centralization." That is, nearly by definition, not true. Decentralization comes with significant costs; see Ethereum, for example. Ethereum has many thousands of GPUs available to it, but the Ethereum Virtual Machine can only process one instruction at a time. That's an extreme consequence of the particular flavor of decentralization that the blockchain forces you into.
Of course, decentralization confers benefits as well, but in choosing the decentralized architecture, you've got to weigh the costs.
Thanks agains!
The decentralization means that we can remove the inefficient intermediaries (banks, lawyers, notaries, clerks, courts...) and replace them with computer hardware.
So ether can only do one instruction a second, try buying a house and see how much beurecracy is involved. Try buying one on the other side of the globe and you will be burried by documents. That is inefficient, not ether.
If that house was listed on a public blockchain you could make the exchange in minutes and pay next to nothing in comission and you could trade that house globally.
You're essentially saying that blockchains are more efficient because they let you dodge regulations? I'll grant you that, and it does make it harder to analyze.
But it's really tricky to disentangle these things, and I'd argue that it's not so much decentralization that enables these things as much as it is automation. A centralized entity could just run a blockchain themselves, right? Why don't they? Because other database structures are better if you have access to the advantages of centralization.
If you're saying that it's the blockchain itself that gives efficiency, then you need to come to grips with the fact that a blockchain is just a very inefficiently-designed database. That's all it is. It's designed inefficiently to allow for decentralization; its inefficiencies are necessary because when you decentralize you lose all kinds of robustness.
And admit it - your example of the house is a straw man. What does "list a house on a blockchain" even mean?
I don't mean dodging regulations. It's possible to keep all standard regulations on the blockchain but without all the paper work using smart contracts.
That way you avoid all the corrupt middle-men like the lawyers and government clerks.
Why do you need it public? So it can be audited.
Why do you need it distributed? So a copy can be obtained and kept on every computer giving it credibility and robustness.
Why do you need decentralisation? To avoid corruption. The little computers doing PoW are securing the network from corrupt humans.
Why is it efficient? Because having offices in downtown, employing tens of thousands of humans that audit, test, check, read, sign documents is not efficient and makes no sense when computers can do it better, faster and a lot more secure.
It's not a strawman example, read about colored coins, they allow listing any asset right on the Bitcoin blockchain. It's possible, problem is that it has no meaning nowadays but if governments decide to list assets on the blockchain than turning all that paper records (or sql databases) to a public record listed right there on the bitcoin blockchain and secured by PoW is easy and each owner will have a private key or a P2SH address proving his ownership with no doubt.
Suggested reading:
http://unenumerated.blogspot.co.il/2017/02/money-blockchains-and-social-scalability.html
"Of course, decentralization confers benefits as well, but in choosing the decentralized architecture, you've got to weigh the costs."
It's tough to compare these, because while you can quantify the technical efficiency costs (power usage, hardware costs, etc.) you cannot quantify the benefits. That's because the benefits are almost entirely confined to "removing the need for trust / trustless system". In theory, a system with fully ethical actors would see minimal gain from the increase in trust. A wild west, snake-oil kind of system would see huge benefits.
How do you quantify the greed of mankind?
Great points, these.
There's an ebb and flow between how Facebook grew up and became big and then became ubiquitous - and did certain things with the implied levels of trust along the way.
Most everyone getting involved didn't fully realize that "it's free and always will be" ended up meaning that, to borrow someone's adage, if you aren't buying the product, they're profiting off your data.
This is a great discussion...
Yes, but necessary, IMO.
There is a theoretical framework for this called the "Price of Anarchy"; it was introduced by some theoretical computer scientists about 20 years ago and it's attracted quite a following among the engineer/computery game theorists out there. Essentially, it tries to measure just that - how do you measure the harm of greed?
For the past couple years I've been studying the price of anarchy in some network routing systems, and I agree with you that it's a hard problem - but I think it's an interesting and important one nonetheless.
Did you see Vitalik's post on Reddit yesterday about his take on the democratization of Bitcoin vs. Ethereum? It was quite interesting - his point was that Satoshi might have done more harm than good by making Bitcoin very democratized, perhaps over-decentralized. Scattered, if you will.
Conversely, while I do agree with all you say about Ethereum, Vitalik's (semi-invisible) hand means that there's the ability to sorta kinda change the rules when they make sense to change. Within reason - see ETH v. ETC.
Appreciate the convo. Following you now...
It also creates a huge failure point, as we saw with the Vitalik death hoax.
Yeah, failure points are a problem...
...then again, like a Martha Stewart, where the person is the brand, the person can do lots to make sure that the brand can continue without them.
That's a good point, and one of the fundamental strengths of decentralization is the whole Hydra effect - cut off one head and two more grow to replace it.