The article states that bitcoin transactions take hours. Is that true?
In theory, you should wait for a transaction to have at least six confirmations on it before truly trusting it. That has always taken an average of an hour from the transaction first going into a block. It's the "first going into a block" part that takes longer now, unless you attach a higher fee to ensure that it goes in quickly.
Not true. A 0-conf transaction is a transaction which has not been confirmed at all, which is orthogonal to its fee.
A transaction with a reasonable fee may not even be confirmed at all, as it only takes a few days to fall out of every mempool and average fees are currently unreasonable, so it is not safe to trust 0-conf transactions currently as they may never make it on-chain.
"as long as the transaction [will be confirmed within several days], you can be pretty sure that it will be confirmed eventually".
While this seems like a technical glitch to some, it is easily addressed using traditional finance techniques or via side chains.
Fees exist to incentivize miners to participate, and so "unreasonably high" fees should result in more miners participating and fees going down. Due to the scale of Bitcoin, adding non-trivially to the mining capacity takes a pretty massive investment, so we should expect a gradual correction.
Overall this is a good thing because it enhances the overall cryptosystem significantly while clearly signaling (via the price) what needs to happen.
While it is tempting to solve this technocratically, I'd argue that institutional stability (even if it seems a bit like calcification) is extremely valuable to the formation of long term incentives and risk-taking for the ecosystem as a whole.
It would be nice to have a proper economic feedback mechanism on this, e.g. miners decide how big to make the block based on how high the fees are getting. However, the costs for big blocks are externalized and so I think miners would make blocks as big as possible to capture as many fees as possible. The cost of storing those blocks in perpetuity costs the whole Bitcoin ecosystem.
I consider this a major crisis for Bitcoin as with current fees, the system is not very useful. The original thinking was that block sizes could be increased a lot and provide transaction rates that match what credit card companies can do. Current Bitcoin core developers and miners can't seem to increase the block size, each blames the other for inaction. Some suggest the core team is mostly controlled by Blockstream and their business model depends on expensive transactions forcing people to use their Lightning network instead. Miners are fine with high fees as long as the Bitcoin price doesn't collapse.
I don't know how this will play out. Perhaps some clever idea will come forth to solve the scaling problems. If so, Bitcoin should continue to be adopted and rise in price. If things say deadlocked, I could see Bitcoin wither away as other crypto-coins do the same job but for lower transaction fees.
However if you consider that miners (or more abstractly, mining rigs themselves) are easily repurposed to mine a BTC fork which could potentially include a larger block size from the start.
So the conflict is between the beneficial network effects of BTC and the transaction cost imposed by arguably "bad" governance (corrupt governance, if the rumor you mention is true).
There are lots of interesting incentives that apply to BTC at scale that are hard to predict when reading the paper.
I think we'll see a very messy world of many competing cryptocurrencies and difficult-to-understand layers of overly centralized governance resulting in layers upon layers of perverted incentives.
But the silver lining is that the blockchain is a nice way to offer the kind of transparency that could make a system like that actually far better than the kind of messes we end up in with fiat currencies :)
I will tell you that, right now, I can just about barely fit the entire blockchain onto my laptop. With a bigger block size, that would no longer be possible. Similarly, the network requirements of transmitting much bigger blocks would hemorrhage users as well. What we need is something like Lightning Network, which is a multiple-level system that doesn't require that all participants download all data from the levels beyond level one. Just naively scaling up level one removes lots of participants from the system and still hits a cap at well below what we'd want Bitcoin to finally be capable of handling.
This makes bitcoin much less useful in my personal opinion.
IMO we really should be pushing for GNU Taler, which is a system that doesn't have proof of work, is confirmed effectively immediately, and has many more privacy properties similar to cash. It works by having a mint sell customers coins (using whatever standard method of payment you like), the customer spends coins on merchants and then the merchants can redeem the coins for "real" money.
In particular, customers cannot be identified (nor can their transaction history) but merchants' income can be identified (for tax reasons). As the system doesn't require proof-of-work the speed problem is massively reduced as well as the transaction cost.
In addition, Taler is _federated_ not fully decentralised. It works basically how cash works, except you don't need to be a new country to set up a new currency. In order for a particular mint's cash to be spent you only need that mint to be around.
I don't get why "decentralised or bust" is the end goal. Federation with individual privacy is still an improvement over centralised credit card companies. Also, like it or not, but Taler has much better privacy guarantees than BitCoin.
This is assuming retail stores will wait for 6 confirmations to get 100% assurance in the blockchain. Many retail stores take a certain level of risk with all plastic/digital transactions currently with credit/debit cards. I doubt many would see much utility in waiting for more than 1 confirmation, considering the difficulty involved in faking one and the lack of real stories of this 'scam' actually happening.
Especially for low-margin sales like coffee, food, or other sub-$100 transactions.
As with any scam of this nature it will likely involve side-step the crypto system somehow (ie, target the mobile apps used for PoS) rather than breaking the blockchain confirmation scheme directly.
We really need Lightning Network, which segwit enables. That allows near-instantaneous (with only network propagation delays) transactions.
The key thing is that bitcoin transactions are not reported to the government -- very helpful if you're desperate to move a large amount of cash out from a country with heavy currency controls like China or Venezuela.
Recently Ethereum has been taking a chunk out of Bitcoin's market share. One of the big reasons why is that it does not currently have this problem of a massive transaction backlog requiring large transaction fees.
And the transaction pressure then wasn't even what Bitcoin normally handles. (I'm not sure if that says more about the state of Bitcoin or Ethereum however.) Anyway, if you need transaction capacity, look elsewhere.
Most cryptocurrencies work like that, especially since they were all more or less inspired by Bitcoin. The capacity to which Ethereum and Bitcoin can scale can differ by a constant factor at most. (Where the latter is likely at an advantage, due to smaller transaction sizes and simpler scripts.)
Scaling Bitcoin to a global payment network for everyone's daily transactions is probably not realistic. But it could very well act as the clearing backend for such a system. (See, for example, Lightning Network.)
http://www.livebitcoinnews.com/ico-investor-pays-us1500-wort...
https://etherscan.io/tx/0x0ca73d29dd82b48eba41331ed5d33ab836...
was a bit surprised myself
Other PoW-based systems have similar recommendations. You can have more frequent blocks, but the economic security offered by each confirmation will be lower.
There are other nifty networks called payment networks which promises instantly secure transactions, the most well known being Lightning. They offer that by identifying cheating and offering both parties to get their money out in case the other party does. They are still pretty experimental but the results are promising.
The simplistic explanation is that the Bitcoin network effectively incorporates a hard upper limit on the number of transactions per unit time. (Every new transaction must be recorded in the blockchain, and the blockchain only grows so fast.) So as the currency becomes more popular and demand for transactions increases, either your waiting time to get the transaction confirmed on the blockchain increases, or your transaction fee has to increase.