Tonight, on your special edition of SubredditDrama we will debate on the block size limit and watch the arguments unfold. Grab your popcorn and welcome to Sub Reddit Drama!
*cue intro music
First, some context for the uninitiated. Bitcoin
is a decentralized cryptocurrency, which processes transactions in blocks
. Typically, these take about 10 minutes to process, and this is how transaction verification works.
Currently, the limit on the size of each block is 1 megabyte, to prevent people from create gigantic blocks that would take a lot of data to store in the long run. Because of this, only so many transactions can be processed in a block.
Recently, there has been a proposal to increase this size to 20 MB. This would make the maximum size of each block larger, which could make the network able to process more transactions in a shorter amount of time.
Many people have their opinions, and everyone knows that duty calls
Now, for the drama...
First, someone expresses their support for the block size increase
. This starts a few subthreads of drama. A debate is started
over what the current transaction limit is. Arguments are had over whether Bitcoin is scalable
also come out of that.
In another subthread, a few want to remove the limit altogether
, while some want to let the free market
sort it out. Arguments over whether Lightning
would help or increase centralization also occur.
After the break, we'll put the butter in your bowl. Okay, we're back, dramanauts, and we have more drama from another (related) thread. On Twitter, Nick Szabo made this tweet. It has been linked on /Bitcoin, and this is good, because people are arguing already. Here is this thread on whether Satoshi is a god. A small slapfight occurs over whether Gavin Anderseen is in charge.
In a third thread, someone says that most of the developers disagree about a larger block size.
Additionally, in this thread, another user points out that some arguments were based on economic fallacies. Someone else in a different subthread points out that reddit's hivemind was saying "because Satoshi said so". Users disagree.
Some argue over Gavin's activity. Arguments over how this will affect Tor, a way to privately connect, by tunneling through a few nodes (I believe) also arise.
This is getting argued to the death, and I'm running out of things to say about it.
I think this is all the drama I can find for now, but the mailing list has additional drama. This is definitely a developing story though.
We'll see you next time on Sub Reddit Drama!
Good afternoon! A few thoughts for dinner tonight:
Finally, some progress
It looks like there is finally some progress being made to address network efficiency. A "Technical discussion of Gavin's O(1) block propagation proposal" is going on at http://www.reddit.com/Bitcoin/comments/2d7ofh/technical_discussion_of_gavins_o1_block/
. It seems as if I am not the only person who believes this development is very significant.
I hadn't appreciated this for the genius that it is. I saw the parts about reducing bandwidth for uploaded blocks and thought that was a good enough reason to implement the proposal. But it turns out that there is an interesting twist that can come out of this. If you don't have to upload the transactions a second time in every block that is mined, then why upload a list at all? In fact, why not simply make the default to include every transaction that is available, and then only submit differences from that list? The actual implementation is more complicated, but the idea changes the entire concept of the network.
While everyone else viewed the task as figuring out how to deal with increasing block sizes, Andresen came along and entirely bypassed the problem. It's a great example of overcoming existing thinking and coming up with entirely new assumptions. These are the sorts of ideas that will cause bitcoin to rise out of its current slump and resume its march towards widespread acceptance. lifeboatz
thinks that it will take 6 months for implementation of this proposal, but I'd propose that experience is so important that he could implement it in one month. lifeboatz is correct in suggesting that testing will take much longer. If implementation requires x time, then testing is probably 6x, and political discussion to reach agreement is likely 18x.
After implementation, I think that what Andresen has done will be viewed as a significant breakthrough in computer science. He has demonstrated a way, when there are two copies of the same data across a network, to agree on a subset of the data with trivial bandwidth. If implemented successfully, the algorithm can be used for many other applications - like online video gaming, for example.
Right now, everyone assumes that the problem has always been how blocks can be propagated across the network efficiently, and the key to that was always to raise transaction fees high enough to cut out spam. What transactions are spam has been debated by a lot of people and is a sticking point for many of these arguments about how to make blocks bigger. But once you make it the default to include all transactions nodes have received, then there isn't much danger in including "spam." The "spam" transactions need to be received by every client anyway, even if just to determine they are spam and not include them in blocks. Spam costs money to send.
I think it's worth demonstrating how "spam" (which can also be called microtransactions) is a non-issue, even now, let alone in the future.
Let's suppose that we have 100MB blocks, which are filled entirely with transactions that have a fee of 10 satoshi, which is worth about $0.00006. You could send nearly 200 of these microtransactions before you would lose even a single cent to fees. Presume the transactions take up 256 bytes each. The total fees earned in this block would be about $23.40. A hard drive costs about 3.77 cents per gigabyte today, which makes 100MB worth about 0.37 cents. If there are 7000 nodes in the bitcoin network, then the total cost of all the 700 gigabytes in copies is about $25.90. Therefore, with today's technology, the blockchain storage for such transactions can be paid for with fees of $0.00006 using gavinandresen
's new method.
Of course, we can expect storage to decrease in cost, so in two years, blocks could be 200MB in size, and in four, 400MB, and so on. In just 10 years, 10-minute blocks can contain 3.2GB in "spam" and the transaction fees would be equal to the cost of the storage. Note that bandwidth costs are not relevant to these calculations because there is nothing we can do to stop people from using our download bandwidth, even if we don't want to receive the data. Bandwidth is what is called a "sunk cost."
There are some problems with these calculations: for example, node operators don't earn any fees; miners do. But since large banks and hedge funds are going to be the miners of the future, and these are the same people who have an incentive to run fast nodes, they will have a shot at winning these fees by mining.
Some mistaken ideas about decentralization
Expanding the network means giving up on the idea that everyone will be able to run a node at home in their basement without using specialized equipment. Even if Gavin's algorithm were implemented under the conditions where all transactions are included by default, it would still be possible for someone with just a few thousand dollars to set up a node. On the other hand, there will be a point where people who have laptops they want to use to install Bitcoin Core will be unable to do so.
A decentralized network is immune to interference, while a centralized network is more efficient. Bitcoin will probably end up somewhere between the two. There will be many nodes run by companies with powerful computers. The lower number of nodes will make the network faster, and as long as the number doesn't go too low, it still cannot be shut down by anyone. The goal should be to create the maximum amount of centralization that can be attained without introducing a risk of takeover.
A note about VISA
People keep mentioning VISA's transaction volume as something that bitcoin needs to be able to achieve. That is limited thinking. Bitcoin has the potential to make types of transactions possible that VISA cannot handle, like microtransactions for only a few cents. Saying that VISA has 1,000 transactions per second is irrelevant. The bitcoin network needs to be able to handle 100,000 transactions per second or more, because there will be new use cases that VISA doesn't handle, and which aren't possible with VISA.
Altcoins moving to avoid ASICs?
There are a lot of altcoins appearing nowadays that believe they can be "more secure" by using an algorithm that does not have ASICs (yet). Their creators believe that, by forcing miners to use CPUs or GPUs, then anyone will be able to mine, making the network more secure. This idea is flawed for many reasons.
First, using general purpose CPUs to secure a network is dangerous, because there are so many of them that can be repurposed for a brief period at minimal cost. For example, someone could rent a supercomputer for a few hours, and destroy every single altcoin that doesn't have ASICs protecting it. After the attack, the supercomputer is put back to its other work, so the only cost is the rental fee. The same person can rent the supercomputer a second time and perform the attack again, should the creators of the coins reissue their blockchains. If ASICs were required to destroy these coins, the attackers would have to spend a lot more to buy up all the ASICs, the price of which would skyrocket as supply dwindled (note that a pool with different miners cannot execute such an attack, because miners will leave the pool, so one person needs to buy up all the equipment). Then, his or her investment would be worthless after someone produced newer, more efficient ASICs. Even if it were possible to make an "ASIC-resistant" coin, then it is too easy for people to buy up a batch of CPUs and destroy hundreds of coins with 50 different algorithms.
But that assumes that it is possible to design an "ASIC-resistant" coin, which it is not. It is always more efficient to implement something in hardware than it is to do it in software. Even if the ASICs are only 1.2 times as fast as the CPUs, people will still buy them because they can make more money. Any coin that succeeds will have ASICs produced to mine it. My prediction is that within a year, we will start to see ASICs for other algorithms, like scrypt-n, scrypt-jane, x11, and so on.
The other way that new coins try to avoid ASICs, which aren't a bad thing in the first place, is to be 100% proof-of-stake. Since proof-of-stake causes the rich to get richer, a proof-of-stake economy reduces the incentive to buy anything and concentrates power. To earn money in a proof-of-stake system, you have to leave your node online all the time. But since doing that costs electricity and hardware, you need to be earning enough money to justify the odds of running the node. Since many POS coins pay 1% interest per year, and a typical computer and router and cable modem might consume 100W, I computed that at 7.79 cents per Kwh it costs $73/yr to run the node (but it might also cost $100/yr in hardware costs, for a $300 initial outlay depreciated over 3 years). To make $173 in one year at 1% interest, a POS miner needs to hold and not spend $17300 just to break even! That's an absurd figure - who would invest so much money in an altcoin that could go down the tubes when the much safer stock market has made an average of 7% for the past 100 years?
There is no innovation in the altcoin world when it comes to algorithms or mining. Bitcoin's SHA-256 mining is still as good as anything else out there, and it has the benefit of simplicity.
- Days until the initial New York comment period ends: 28
Planning Your Bitcoin Mining Operation – Block Operations – Operating a Bitcoin mining facility can be profitable, but you need to treat it like a business. I operate a small Bitcoin mining facility, and there is a lot more that goes into it than just plugging in the equipment and letting it run. Seite 1 der Diskussion 'OILEXCO - vom Explorer zum Produzenten' vom 01.06.2006 im w:o-Forum 'Öl- und Gasstocks'. The energy used to mine bitcoin has long caused debate over whether it’s a wasteful process. As the arguments have rumbled on, some people have been focused on mining coins with renewable energy. The block size debate has not been resolved. We cannot have a situation where unsophisticated investors are investing in a digital commodity that could literally change overnight and disrupt the market value significantly — I’ll happily debate this in a separate post but a hard fork right now would be horrible for Bitcoin, IMHO. The new amendments have not been open to public debate and I ... On tonight’s episode of “The Crypto Show,” our guest is Julia Tourianski of www.bravetheworld.com and the Youtube channel “bravetheworld.” Before we talk to her, however, Danny updates us on his recent arrival to San Francisco, where he has already met up with Julia. Despite some automotive troubles, Danny has been able to stay on track with his trip, and he gives us a quick review ...
Whether the block size should be increased to 20MB has created more controversy than any other question in Bitcoin's recent history. For some, it is an urgent and necessary step in Bitcoin's ... In this talk in Berlin, Andreas looks at the inner structure of bitcoin and how high-level financial and trust applications are composed from smaller element... The highly-charged debates around both Ethereum's potential hardfork and Bitcoin's block size appear different on the surface, but are caused by the same des... Roger Ver Debates Charlie Lee (FULL DEBATE) - What is Bitcoin? - Duration: 1:26:29. ... Must watch talk on mining, block size, and more - Duration: 55:04. The Bitcoin Foundation 19,926 views. 55 ... Stephanie, Andreas and Adam speak first with Bitcoin Foundation Chief Scientist Gavin Andresen about the Bitcoin Block Size debate, where it came from, why it matters and what he thinks we should ...