Underestimate blockchain at your peril [LWN.net]

Underestimate blockchain at your peril

Posted Jan 3, 2018 15:20 UTC (Wed) by excors (subscriber, #95769)
In reply to: Underestimate blockchain at your peril by anselm
Parent article: Welcome to 2018

Like the recent bug in an Ethereum contract that allowed some random person to 'accidentally' destroy ~$150M of other people's money. There is no authority that affected users can appeal to for compensation, because the (buggy) contract is considered to be authoritative, so that money is gone forever. (In theory it could be fixed by forking the blockchain and then convincing every Ethereum user in the world to switch over to the new fork, but that sounds very difficult for non-trivial networks where users have competing interests.)

The inability to retrospectively fix mistakes seems like a fundamental flaw in these decentralised systems.

to post comments

Underestimate blockchain at your peril

Posted Jan 11, 2018 5:27 UTC (Thu) by Garak (guest, #99377) [Link] (20 responses)

too big to fail. bailout-forks. diversification. survival of the fittest. evolutionary forces.

The power seems to reside in the network. Governments still seem to have ultimate dominant authority over the networks. But maybe sneakernets are enough for some interesting use-cases. Maybe someone will come up with thousand mile range wifi mesh networking with interesting bitrate specs. That would be cool.

Just because traditional currency can be useful because of its centralized nature, doesn't mean that cryptocurrencies' lack of those features makes it useless. I.e. what you describe as a fundamental flaw, I'm pretty sure many of the early pioneers (myself sadly not included) saw as a basis feature. They perhaps, for some use-cases, considered Trusting Majority Of Network Power to be preferable to Trusting Government. I can't say I consider the attitude crazy. I think it's wise to understand that things are complex, and use-cases are many. Traditional fiat currency is not going to become obsolete, but it has clearly begun to lose market share to bitcoin/litecoin/etherium/dogecoin/etc. I'm sure that baking the net effectiveness of a bailout-fork into the cryptocurrency from the beginning is something we may see more of. And it will be implemented in lots of different ways, and people will literally vote with their money. Interesting to watch it play out.

Underestimate blockchain at your peril

Posted Jan 11, 2018 15:16 UTC (Thu) by nix (subscriber, #2304) [Link] (19 responses)

Traditional fiat currency is not going to become obsolete, but it has clearly begun to lose market share to bitcoin/litecoin/etherium/dogecoin/etc.

In the most fundamental sense, "what can you buy in a market with it", this is clearly nonsense. Bitcoin, and other cryptocurrencies, are not usable as a currency You can hardly use them to buy anything, and when people do implement buy-in-$cryptocurrency features, they as often as not tear them out again because nobody uses them -- because they are not usable as a currency. They are worthless stores of value: the exchange rate shoots madly up and down because the market is so thin that an exchange rate is being derived from only a few transactions, because it's fairly hard to get money *into* these currencies and next to impossible to get it out again.

This last property is not necessarily a problem: there are plenty of nonconvertible currencies around the world. But combine that with the fact that nobody uses any of these things as currencies, merely as speculative instruments, and with the disastrous problems implied by the nonrepudiatable nature of the thing renders it fundamentally useless as anything resembling a currency (oops, I typoed the wallet ID for that deposit payment, what do you mean it's gone? You can't reverse simple mistakes without convincing everyone using the currency to use a new one?!). The important part of central control of a currency is that rules useful for the majority of users can be enacted without breaking the entire currency. Rules like clawing back mistaken or fraudulent transactions, which is not exactly a problem you can say existing cryptocurrencies haven't been impacted by. (Yes, this means something outside the currency has to get involved to reverse the transaction's effects in the real world, if any.)

The underlying datastructure has many uses (though not the proof-of-work scheme, which is just an excuse to run lots of expensive room heaters at worrisome environmental cost, and doesn't even prevent single large entities from dominating the network, as nearly happened with Chinese miners repeatedly over the last year or two). But as a currency? I don't see this ever threatening even something like the kyat, let alone the yuan or the dollar or the euro. Normal human beings will not tolerate a currency with no way to reverse mistakes and no way to defend against theft of their entire bank account or heck their entire bank.

Underestimate blockchain at your peril

Posted Jan 11, 2018 17:10 UTC (Thu) by anselm (subscriber, #2796) [Link] (3 responses)

Normal human beings will not tolerate a currency with no way to reverse mistakes and no way to defend against theft of their entire bank account or heck their entire bank.

Not to mention that people won't be enthusiastic about a “currency” in which, when you buy a newspaper or a cup of coffee at the train station, the time it takes the transaction to clear is longer than the train ride.

Underestimate blockchain at your peril

Posted Jan 11, 2018 20:36 UTC (Thu) by nix (subscriber, #2304) [Link] (2 responses)

Quite! but that's a problem with Bitcoin in particular. I was trying to point out mostly problems common to all blockchain-based cryptocurrencies similar to Bitcoin, not just implementation problems in Bitcoin.

The related problem for all implementations is that they have a horrible choice:

- either make transaction costs ridiculously high due to the need to lug around a massive database of all transactions ever -- which for practical replacement of real currencies would need to have new blocks arrive at least once a second, I'd guess, so it would be exabytes in size soon enough

- or make everything horribly insecure due to the need to trust someone *else* with carrying out the actual, netted transactions of many other people. And at that point you're back to trusting someone else not to run off with all your money, so, uh, what's the advantage of this over ordinary currency again?

Underestimate blockchain at your peril

Posted Jan 12, 2018 9:47 UTC (Fri) by Garak (guest, #99377) [Link] (1 responses)

Did you not get the memo? I have no idea if these will turn out to be the solution of the problem you highlighted. But it seems to me this sort of addressing of issues that come up is standard FOSS operating procedure. I suspect that within the next few years there will be many variants of and competitors to this solution, and enough of them will work well enough that the problems you highlighted will be considered entirely historical within 5 or 10 years.

https://en.wikipedia.org/wiki/Lightning_Network
https://en.wikipedia.org/wiki/SegWit

Of course I believe there will be half a dozen other major issues that come to light, and of the hundreds of attempted solutions to those most will fail. But enough will survive and thrive. This is all the power of darwinian evolution, capitalism, and FOSS rolled into one. It will be bloody, but to quote Jurassic Park- "Life Finds A Way". "The Desire To Have A Cryptocurrency That Survives Will Find A Way".

Underestimate blockchain at your peril

Posted Jan 12, 2018 16:56 UTC (Fri) by nix (subscriber, #2304) [Link]

I haven't paid attention to SegWit but Lightning Network fundamentally doesn't work because it's a network with no mechanism proposed to ensure that routes actually eventually terminate at the intended destination, nor indeed that they won't contain (ruinous) loops, because every routing node has no understanding of anything about the network topology beyond the next link or two. (Also, everyone who wants to be part of a route has to go into debt for howeverlong it might take transactions to clear, which might be forever due to the previous problem). And that's just the *start* of its huge laundry-list of fundamental flaws.

In any case, netting everything up to get around terrible transaction costs and delays is how you get around a broken system, not a way to advertise a system as ready for prime-time.

Underestimate blockchain at your peril

Posted Jan 11, 2018 18:51 UTC (Thu) by gracinet (guest, #89400) [Link] (4 responses)

Well, real currency or not, SFC just got a donation in bitcoins and says it plans to accept bitcoins for the projects under its umbrella.

https://sfconservancy.org/news/2018/jan/11/pineapple-fund...

Underestimate blockchain at your peril

Posted Jan 11, 2018 20:38 UTC (Thu) by nix (subscriber, #2304) [Link] (3 responses)

But in order to *do* anything with that donation, they'll have to find some way to convert the damn thing back into dollars or some other mainstream currency. Good luck with that. (Will the time spent finding a way to do that, administering the payment method, etc be worth less than the donations that'll arrive via that route? Often the answer turns out to be "no".)

Underestimate blockchain at your peril

Posted Jan 11, 2018 21:43 UTC (Thu) by halla (subscriber, #14185) [Link] (1 responses)

I would even estimate that the time I spent answering people who ask me "why doesn't krita accept bit/doge/whatever/coin" is worth more than the donations that would arrive!

Underestimate blockchain at your peril

Posted Jan 11, 2018 22:38 UTC (Thu) by karkhaz (subscriber, #99844) [Link]

For a small business owner's perspective on accepting Bitcoin, I enjoyed reading Colin Percival's blog post from a few years ago:

http://www.daemonology.net/blog/2014-03-27-tarsnap-bitcoi...

he claims that Stripe's service means that he doesn't actually have to deal with bitcoin, Stripe just sends him dollars. So bitcoin is being used to make a transaction anonymously, but not to store value long-term. This seems to be the best of both worlds...

Underestimate blockchain at your peril

Posted Jan 13, 2018 10:25 UTC (Sat) by smurf (subscriber, #17840) [Link]

What do you mean, "good luck"? Get an account at a reasonable broker, send coins there, sell them, transfer $$$ to own bank account. That's less work (and cheaper, even including transaction cost) than opening a non-$-based account at a bank and getting the money converted into dollars.

Yes, bitcoin transaction cost are insane at the moment, but that's only a problem if you want to get the transfer included in the next block or two. If you're willing to wait, reduce them.

War On Cryptocurrencies

Posted Jan 12, 2018 8:59 UTC (Fri) by Garak (guest, #99377) [Link] (9 responses)

>Traditional fiat currency is not going to become obsolete, but it has clearly begun to lose market share to bitcoin/litecoin/etherium/dogecoin/etc.

In the most fundamental sense, "what can you buy in a market with it", this is clearly nonsense. Bitcoin, and other cryptocurrencies, are not usable as a currency

The news stories keep suggesting the opposite of what you are saying. I admit, I haven't started using it yet. But it sure sounds like other people have. And with the total market value of bitcoins at $240 *Billion* USD? I was a skeptic for a long time as well. I still expect major events, like the NSA or KGB deciding they want to hork the leading cryptocurrencies, and succeeding, because they can. But even if that happens, or just a panic crash, I can't look at the history up to this point and have the same kind of sentiment you do. One could have rattled off a similar litany of criticisms of Linux as and OS in 2000. It was a young OS, and there were many legitimate and broad problems with it. But people with imagination, and understanding of FOSS crowdsourced engineering could *correctly* see beyond the naysayers. As a concept, a FOSS OS has thrived despite historical litanies of issues. After watching bitcoin's (and litecoin, and etherium, and dogecoin's) trading value chart over the last 5 years, I think it's very clear that your sentiment is completely off the mark. Cryptocurrencies are here to stay. The main issue I see is not technical- all these problems will be solved relatively easily over the coming years. But rather political. I vaguely recall a headline/story about Russia outlawing bitcoin in some form, but saying that there will be a government managed cryptocurrency. This itself is a fascinating aspect of the mid-term future, as far as noting different country's prohibitions on cryptocurrency, or other technology. And I don't think Russia's War On Bitcoin will be any less ridiculous to watch than the decades old War On Cannabis.

War On Cryptocurrencies

Posted Jan 12, 2018 9:18 UTC (Fri) by anselm (subscriber, #2796) [Link] (8 responses)

If a cryptocurrency ever becomes acceptable for everyday usage (think “grocery shopping”), it's unlikely to be Bitcoin because Bitcoin can't handle the transaction volume. It is running up against hard limits even now even though hardly anyone is using it. Similar issues apply to other “mined” cryptocurrencies.

I don't think the comparison to Linux makes a lot of sense because at the end of the day Linux is just a software program, and in principle anyone can write a software program, especially if one is doing it in one's spare time and can't go bankrupt doing it. Especially when it was new, Linux wasn't doing anything, on a technical level, that others weren't doing already. Bitcoin, OTOH, is trying to break entirely new ground and whether it is a great idea to waste the energy budget of small European nation states enabling an infrastructure that, for the most part, powers ransomware and drug deals on the Darknet remains to be seen.

War On Cryptocurrencies

Posted Jan 12, 2018 10:10 UTC (Fri) by Garak (guest, #99377) [Link] (7 responses)

If a cryptocurrency ever becomes acceptable for everyday usage (think “grocery shopping”), it's unlikely to be Bitcoin because Bitcoin can't handle the transaction volume.

And what makes you think the bitcoin developers wouldn't copy the features of whatever fork or competitor defeats it technically, and call the resulting thing 'bitcoin'. I agree with you that bitcoin as it is today isn't going to last forever. And odds are the longer term winners won't be called bitcoin. But by the time a bitcoin competitor becomes usable for grocery shopping, what makes you think the bitcoiners won't adopt those features/designs to remain competitive? Again, I agree with you basically. There will be lots of competitors. I'm not thinking in terms of caring what the long term successful ones are named. I'm more concerned with thinking about what it means that there will be long term successful cryptocurrencies. Bitcoin is a name, a current design, implementation, and deployment. To say that bitcoin, as it is today, if it doesn't evolve further, will not thrive, is true, but not insightful.

I don't think the comparison to Linux makes a lot of sense because at the end of the day Linux is just a software program, and in principle anyone can write a software program, especially if one is doing it in one's spare time and can't go bankrupt doing it. Especially when it was new, Linux wasn't doing anything, on a technical level, that others weren't doing already. Bitcoin, OTOH, is trying to break entirely new ground.....

At the end of the day Bitcoin is just a software program, and in principle anyone can fork it, with zero or any number of modifications (just like Linux, FOSS is FOSS, Linux and Bitcoin both are FOSS). And when bitcoin, or any past or future fork of bitcoin or any cryptocurrency starts out, it isn't doing anything on a technical level that others weren't doing already. Bitcoin isn't breaking entirely new ground. People were pondering cryptocurrency long before bitcoin became the first majorly successful implementation and deployment. And that was hardly overnight. Any of the inevitably many new cryptocurrencies will have the same trajectory. Starts as a small project, and to the extent it succeeds in doing anything useful, it achieves and reaches critical mass. The real difference between cryptocurrencies and FOSS projects in general is that the financial benefits that drive development are less indirect. So I think cryptocurrencies will have a more accelerated path of advancement than Linux or other FOSS projects.

War On Cryptocurrencies

Posted Jan 12, 2018 13:34 UTC (Fri) by NAR (subscriber, #1313) [Link] (6 responses)

"what makes you think the bitcoiners won't adopt those features/designs to remain competitive?"

As far as I understand, the core concept of bitcoin is the proof of work. It has to be computationally hard, otherwise forks would remain for a long time. On the other hand, in order to speed up the transactions, the proof of work must be solved quickly. Do you see the contradiction? There are blockchains/cryptocurrencies without proof of work, but they are conceptually different from bitcoin.

At the end of the day Bitcoin is just a software program

Most definitely not. Bitcoin is a database that contains who have how much "money". How do you fork data? Each fork would double the amount of money people have? You'd soon have different amount of money on the different sides of the fork. It would be a nightmare.

War On Cryptocurrencies

Posted Jan 12, 2018 14:22 UTC (Fri) by excors (subscriber, #95769) [Link] (5 responses)

> Each fork would double the amount of money people have?

Why not? The Bitcoin Cash fork appears to be reasonably stable so far, and has increased every Bitcoin-holder's money by about a factor of 1.2 when converted to USD. Obviously that's insane, but cryptocurrencies transcend the petty limits of rationality.

War On Cryptocurrencies

Posted Jan 12, 2018 16:24 UTC (Fri) by NAR (subscriber, #1313) [Link] (2 responses)

Well, the rise of the tulip bulb prices were also insane :-)

War On Cryptocurrencies

Posted Jan 12, 2018 22:07 UTC (Fri) by Garak (guest, #99377) [Link] (1 responses)

was the market cap of tulip bulbs comparative to the market cap of cryptocurrencies? If so, then yes, cause for continued skepticism factor. If not, then it's like comparing apples to oranges, or cryptocurrencies to tulip bulbs as it were.

War On Cryptocurrencies

Posted Jan 12, 2018 22:50 UTC (Fri) by Wol (subscriber, #4433) [Link]

Dunno about the market cap, but, in modern terms just one tulip bulb could be worth about £1/4M. OUCH!!!

(Ten times the salary of a skilled worker, said salary being probably in excess of £25K.)

https://en.wikipedia.org/wiki/Tulip_mania

Cheers,
Wol

War On Cryptocurrencies

Posted Jan 13, 2018 11:30 UTC (Sat) by smurf (subscriber, #17840) [Link] (1 responses)

Well, in a rational world the Bitcoin A plus Bitcoin B price after the split would be the same as before, but there's no sane way to discover their relative utility (such as there is) beforehand. So the market will be volatile until that sort itself out. And since bitcoin volatility easily exceeds 20% over the course of a day or two …

Bitcoins per se aren't the problem IMHO. The idea itself has a lot of merit. And just like in the early days of industrialization there's a lot of cash and hype floating around, and there's stupid ideas like prohibiting dealing with cryprocurrencies instead of regulating the whole business (let's start with applying existing regulations to it) which is about as dumb, or long-term effective, as requiring a man with a red flag to walk in front of a car. You all know *that* story.

The one major problem I see is that proof-of-work leads to an arms race w/ investing in hardware that gets obsolete as soon as it's deployed. To be fair, nobody had any idea how to do a reasonable proof-of-stake system ten years ago. (Actually, proof-of-work is just proof-of-stake expressed in nonredeemable hardware and energy cost.) These days the Ethereum people are at least working on a viable solution. Bitcoin? not so much.

The other problem, which I'm not holding my breath about it getting resolved without another MtGox-like big crash (except that it's going to be worse), is Tether / Bitfinex. https://www.nytimes.com/2017/11/21/technology/bitcoin-bit...

War On Cryptocurrencies

Posted Jan 18, 2018 2:57 UTC (Thu) by Garak (guest, #99377) [Link]

"which is about as dumb, or long-term effective, as requiring a man with a red flag to walk in front of a car. You all know *that* story."

Actually I must have missed that memo, please elaborate.

"The one major problem I see is that proof-of-work leads to an arms race w/ investing in hardware that gets obsolete as soon as it's deployed. To be fair, nobody had any idea how to do a reasonable proof-of-stake system ten years ago."

This sounds perhaps overfocused on the Bitcoin model. Another model I've been considering for more than a decade is something along the lines of bittorrent/network-file-sharing. I'm sure there have been and will be other quite different models than the bitcoin one (which I may not understand- is it basically about making the recursively generated ledger signature easier (less computationally costly) for the many wallet-nodes to compute?, i.e. is that the useful product of the proven work, or is there some other useful product, or none at all?)

NYTimes - "These days, no exchange is bigger than Bitfinex, an opaque operation that provides no information on its website about where it is or who operates the company."

For anecdotal evidence, I recently made my first (sub$100, entertainment/experimentation value) crytpocurrency buy from Coinbase(tm) because I saw it referenced in some cryptocurrency discussion thread, looked it up as being one of the top few, checked it's wikipedia page and saw it was headquartered in California, so figured it was safe enough. Oddly my visa transaction log shows 'FRGN TRANS FEE-COINBASE CH'. Which seems weird. I wouldn't have done business with them if I had thought it involved a financial transaction with a company in a foreign country. However despite these things, and despite a future with an equal proportion of shenanigans and charlottanism to what we have seen with traditional currency markets, I still would bet on the general future of cryptocurrency. With math discovery epochs along the way. Diversification of risk of course being the common sense wisdom.