[cookiecaper]

Totally with you here. The blockchain is an awesome experiment and prototype of what a good decentralized, verifiable system could look like, but it has several significant flaws, some of which are obvious even in the limited context of a worldwide ledger (bitcoin).

In particular, throwing away 47 zetaflops/s of computing power (the current compute power of the bitcoin network [0]) and all of the electricity required to generate it in order to verify the transactions of the tiny quantity of people currently using bitcoin is obscene, and the mechanism that encourages this should be recognized as the unscalable disaster that it is. Basing "Web 3.0" on such mechanisms is, frankly, a ridiculous proposition.

And all of this "decentralization" is thrown out the window by virtually everyone, because to store a full copy of the blockchain on your local disk, you need -- let's see --

    $ du -sh /mnt/apps_syn/dot_bitcoin/
    118G    /mnt/apps_syn/dot_bitcoin/

120 gigabytes of space. So you end up just trusting Coinbase or whoever you have your wallet from anyway, no different than trusting your bank.

If Coinbase had an Ethereum moment, they could fork the chain that all of their wallets use, and with their MtGox-like domination of the market, any resistance to their fork would seriously damage Bitcoin overall. I imagine most miners would go along with it because they don't want their btc to drop to 0.

Consider also that there is virtually no transparency into the actual miner operations. The biggest miners are behind "pools" -- relays -- and though the idea is that pools are made up of tens of thousands of independently-controlled computers, I suspect that if we could see through to which machines were actually doing most of the work, bitcoin would suddenly seem a lot less "decentralized".

This is not to mention that miners behind pools are totally at the pool's mercy, or that 5 pools hold over 50% of the network's hashrate, and that these pool operators have been in one room together to decide on blockchain policy (re blocksize, which miners are incentivized to oppose, since smaller blocksize == higher transaction fees). That is little different than big banks meeting under the auspices of the Federal Reserve to control monetary policy.

It's very sad that people saw bitcoin and thought it was a good idea to expand its flawed model out, instead of accepting it for the forward-thinking prototype it was and thinking about how to improve it before jumping all in.

[0] per https://bitcoinwatch.com

[mbrock]

I work full time on Ethereum development, and I don't have any kind of maximalistic opinions about decentralization, the evils of banks, or whatever.

It's just a very interesting technology with a really great momentum. But everything is flawed under the sun. There is no perfect technology that can completely decentralize authority with no resource demands.

Or maybe Satoshi's reincarnation will come and deliver a new mind-blowing concept that makes blockchains completely deprecated. That would be pretty cool. But until then, you know.

By the way, Ethereum's roadmap includes switching away from proof of work towards proof of stake, which doesn't waste electricity. This model has been used successfully, for example by Bitshares.

It also includes a light client protocol so mobile nodes don't need much disk space or processing power.

[As always, Hacker News is full of critics...]

[cookiecaper]

This isn't a nitpick about how the blockchain isn't perfect. IPv4 is imperfect, HTTP is imperfect, etc. The blockchain is ill-conceived and impractical.

At this point, I only check in on the bitcoin community casually, so I'm not familiar with proof of stake, but I'm glad to hear work is underway to resolve that glaring issue in Ethereum. I'll definitely read up on it. Thanks for the pointer.

>[As always, Hacker News is full of critics...]

Indeed. We come here to discuss, which will involve some criticism. Working on Ethereum is not really an indictment, so don't worry. :)

[mbrock]

That's what I mean, you're not familiar with what's happening, yet you comfortably dismiss the whole idea of blockchain as ill-conceived and impractical.

So both of your major criticisms are being actively addressed by very smart people who plan to have them solved in the near to medium term future.

Does that in any way change your view of blockchain? Maybe it could possibly work? Or, no, it's just a hopeless dumb idea that will surely crash?

[cookiecaper]

>That's what I mean, you're not familiar with what's happening, yet you comfortably dismiss the whole idea of blockchain as ill-conceived and impractical.

It's not "the blockchain" anymore if they've changed the fundamental mechanisms of its operation (e.g., eliminating miners, which the sibling commenter indicated as a goal).

It may be a blockchain, if we want to use "blockchain" as a generic descriptor for a chain of cryptographic signatures, or a blockchain-inspired / blockchain-derived concept, but "the" blockchain, as in the blockchain concept developed for use with bitcoin, is fundamentally dependent on mining for its validity. Change that, and it's not really "blockchain" anymore, at least not as used in the current vernacular. And that's fine -- bitcoin quickly outgrew its britches and needs revision.

>So both of your major criticisms are being actively addressed by very smart people who plan to have them solved in the near to medium term future.

Yes, that's awesome. I'm glad that some people are now deciding that these issues deserve a response instead of continuing headlong into the hype abyss. Why should this mean that people can't complain? If I hadn't, I wouldn't have known that Ethereum is working on a solution. :)

>Does that in any way change your view of blockchain? Maybe it could possibly work? Or, no, it's just a hopeless dumb idea that will surely crash?

Again, I haven't reviewed it yet, but sure, changing the mechanisms of operation could potentially make a thing called "blockchain" workable. "A rose by any other name..." and all that.

[DennisP]

Ethereum is attempting to improve Bitcoin's model by (1) changing to a proof-of-stake protocol to eliminate miners, and (2) adding sharding, so each node stores a fraction of the chain and acts as a light client for all the rest.