Hacker News Reader

Comment by codexon | Hacker News Reader

Bitcoin (and all other cryptocoins) are just not usable for mass transactions. Buying anything with bitcoin takes over 10 minutes, often times over an hour for shops that are paranoid and require 3+ confirmations. During the peak bubble time, I remember buying something with bitcoin taking 2 hours!!

This technical problem has not been solved and the "lightning network" just turns exchanges like Coinbase into a bank or Paypal that can be shut down or run away with your money, which basically ruins the entire point of using Bitcoin.

Until Bitcoin can scale without "lightning network", I don't see people using it outside illegal activities. And from what I can see, the core concept of Bitcoin (and other cryptocoins) is fundamentally at odds with fast transactions. If you make confirmations 2x faster, then it just becomes 2x easier to fake a transaction. So paranoid merchants will just require double the confirmations.

View on HN

10 minutes is nothing. A wire transfer can take hours. A bank transfer can take days. A wire transfer might cost you 20 bucks as well. Bitcoin is amazing at transferring large sums of money quickly, internationally, for a low overhead cost. It's extremely useful for a lot of people.

Where in the world would a wire transfer cost 20 bucks? It's practically free here unless you do something that makes the bank do extra work.

International transfers?

Even those are like 9 euros if you send over 1k.

USA: https://www.wellsfargo.com/online-banking/service-fees $30 to send a wire transfer, just within the country. And $15 to receive one too.

Why the heck do people use those kinds banks over there? Seems to me it takes away all the usefulness of one. It's a money bucket where you put money in and take money out. As soon as one of those three aspects is not there, then what is the point?

https://www.nerdwallet.com/blog/banking/wire-transfers-what-... There really isn't any low cost option for wire transfers in the US. ACH is the free transfer option, and it takes 1-3 days. Wire transfers are the "pay a lot for immediate confirmation" option just for the very few transfers it's worth it on. The average person is unlikely to ever make or receive a wire transfer. There are consumer-level alternatives for personal transfers like Venmo: https://venmo.com/about/product/

But why wouldn't you want to send or receive money using your bank? Isn't that what it is for? As a non-US person this seems really strange to me (but so is the concept of cheques and tip-based income).

I'm used to using the bank, the apps they have and de debit cards they provide (and now also Apple Pay) to send and receive money instantly every day, and there are on average ~10 instant transfers daily to both other private accounts or businesses to pay for things.

Perhaps this is an exotic or luxury position, or there is some unseen cost to this, but I'm not aware of that at this point.

A lot of merchants ask for 3+ confirmations, and sometimes 1 confirmation can even take 20 minutes. Paypal and credit cards take seconds.

And $20 for a wire transfer pales in comparison to the fees people charge for converting bitcoin to cash https://localbitcoins.com/

Yes transferring large sums of money internationally is useful, but it is a niche, and bitcoin doesn't even do it that cheaply.

Bitcoin transactions are instant. You get the 0-conf transaction instantly and are able to buy stuff. The 30 minutes for 3 confirmations are basically the 'chargeback' time, which is about 2 weeks with credit cards. In day-to-day business, people usually trust 0-conf transactions, or are able to handle the chargeback.

How does a 0-conf transaction work? I mean, without confirmations the transaction is not on the blockchain. It basically doesn't exist. Is it just a promise that I will pay a bitcoin in the future?

The seller now has the buyer-signed transaction as it will (hopefully) be recorded on the blockchain. Common wallets will also already deduct the paid amount.

It means that it has been sent to the Bitcoin network to be mined in the next block.

Could the sender simply cancel it?

No. Once a valid, signed transaction is _out there_, any miner can (and is generally incentivized to) add it to a block.

The closest thing to cancelling is making a new transaction which makes the first one invalid (e.g. by emptying the source address) and hoping the new one gets added to the block first.

Because transactions can have fees attached to them, you could presumably make it more appealing to mine the second transaction than the first, but I don't know if this is something people do, and as far as I understand it most miners prefer the first transactions they see.

to expand on the "replacing"

there is a mechanism called replace by fee, so it becomes a question of incentives

The payment processors can scan the mempool for transactions that would invalidate the initial transaction, so then it would be a question of lead time. Great for digital services etc, not so great for in life purchases.

So if you absolutely 100% need confirmation, then yeah 10 minutes is what you need to wait, but that again is a question of probability comparative to credit card chargebacks/fees etc.

Even the next block inclusion isn't a guarantee... it just means that you have at current prices approximately 200k insurance. With enough resources and luck, someone could +EV roll back your transaction, but the cost scales exponentially.

The current viable solution for small inperson transactions is pretty much replicating the banking system. Current versions of abstracting the banking system into something decentralized are clunky/buggy to say the least.

Cash handling is expensive (cash pickup logistics, compliance fees, banking fees, insurance fees, you name it). But it IS from one bearer instrument (cash) to another bearer instrument (Bitcoin) - and that’s hugely valuable. Wire transfer fee at Suntrust for instance is currently at $50, for example: https://transferwise.com/us/blog/international-transfer-sunt... https://www.suntrust.com/content/dam/suntrust/us/en/personal...

I'm not sure why you group BTC with "all other cryptocoins" when it's clear that your arguments are specific to BTC and its clones. There are many other cryptocurrencies with vastly different approaches to scaling that your post does not cover.

Ethereum is already a lot faster than that, and will probably get to Eth2.0 with sharding and PoS in the not too distant future. Then your ETH transactions will be blazing fast, secure, and cheap as chips.

Of course, they've been saying all of that is not so far away for a couple years now, but we will see. I'm tentatively optimistic.

Ethereum is still on proof of work, so it has the same problem as bitcoin. If they made confirmations take 1 minute instead of 10 minutes, that just means their confirmations are 1/10th as reliable as Bitcoin's confirmations. Nothing was solved there.

And PoS has been "not so far away for a couple years now" because they keep finding problems with it.

Ethereum uses a variant of GHOST, which is why they get away with 15-second block times.

https://github.com/ethereum/wiki/wiki/White-Paper#modified-g...

Ethereum's PoS is currently implemented in seven independent and fully interoperating clients. Multi-client testnet should roll out soon, and production is expected first quarter 2020.

Even if the block time can be lower, ethereum supports a comically low amount of transactions per second.

https://ethstats.net/

9943589 (gas limit) / 21000 (minimum price of a ethereum transaction) / 14.57 (average block time) = 32.5 transactions per second.

What's your point, exactly? My initial comment was that ETH is faster (15tx/s vs 3-5tx/s) than BTC and that layer1 scaling solutions for ETH are probably coming soon.

Everyone that has replied to you is just providing more info than I did.

And your response is "ethereum's transactions per second are comically low right now"?

What is my point exactly? I feel like this should be obvious to anyone that doesn't have a bias towards crypto because they are gambling a lot of money on it.

Making confirmations faster usually just tends to make them weaker. The security of each confirmation is based on the work down (or cpu hours). By making confirmations take half the amount of time, this only makes confirmations half as secure, because the amount of CPU hours spent is the same. You can also make confirmations faster by shrinking the amount of work done per block. This means that while confirmations are faster, it will also take longer for your transaction to get into a block in the first place.

Take a look at this and notice how coinbase requires over 10x more confirmations for ethereum than bitcoin.

https://support.coinbase.com/customer/en/portal/articles/593...

The number of confirmations needed to feel like a tx is secure and the throughput of the network are not related. Regardless of that, you're still wrong about security and confirmations, using your own source.

ETH ave. block time: 10s. Time to get to Coinbase confirmations (35): ~350s

BTC ave. block time: 10m. Time to get to Coinbase confirmations (3): ~1800s

So I can make 5-6 transactions that coinbase thinks is "secure enough" on Ethereum in the time it would take me to make one transaction with Bitcoin.

Add to that the fact that the throughput of the network is higher, and ETH is actually about 17x faster than BTC, and that's totally ignoring the on-chain scaling that ETH is much closer to than BTC.

Also I don't currently hold any crypto, so you can stop pretending that everyone who is more informed than you is actually just biased. I cashed (of my mostly BTC holdings) in 2017.

Um no.

I show you this to prove to you that 1 eth confirmation isn't equal to 1 bitcoin confirmation. If it was anywhere equal in security, there would not be a >10X difference. And also multiplying confirmations by average block time and seeing that it is lower than bitcoin doesn't mean coinbase thinks it is more secure. It can also mean that the value of the average eth transaction on Coinbase is lower.

A 51% attacker will be able to churn out 100x the number of confirmations if confirmations were made 100x smaller. It doesn't matter how fast you make them or how many tricks like GHOST you add.

These are some helpful websites/articles on understanding confirmation security.

https://medium.com/@nic__carter/its-the-settlement-assurance...

https://howmanyconfs.com/

Here's why GHOST supports shorter block times for the same security.

In Bitcoin, the chosen block is the one with the most hashpower in the linear chain behind it. Abandoned forks contribute nothing to that.

With GHOST, the chosen block is the one with the most hashpower descended from it, in all forks. This way all the forks contribute to a block's security.

With a linear chain, propagation time is an issue. If blocks are too short it's more likely that miners will find more than one candidate block, causing the work on abandoned blocks to be wasted. With GHOST, no work is wasted.

The original GHOST paper, which proposed the algorithm as a modification to Bitcoin, has some math to calculate, for a given propagation time, the block time that would produce equivalent security to a linear chain. It was significantly shorter.

As far as I can tell, GHOST is only used to mitigate high hash power loss when confirmation times are lowered to 15 seconds.

Orphaned blocks are very rare in Bitcoin.

https://www.blockchain.com/btc/orphaned-blocks?offset=0

Even if you somehow saved all this wasted hash power, you would only maybe increase hash power by less than 1%!!!

Read my comment here for more details.

https://news.ycombinator.com/item?id=21550997

Sounds like we vehemently agree. Orphaned blocks are indeed rare in Bitcoin, because they have ten-minute block times. GHOST only helps when block times are so short that you get lots of orphaned blocks.

And that's how Ethereum gets away with 15-second blocks.

But you haven't proven that 1 ETH 15 second block is equal or even in the same ballpark in security to a 10 minute BTC block.

>that just means their confirmations are 1/10th as reliable as Bitcoin's confirmations.

This argument is tired. There is no rule that dictates that the same amount of work or stake is equal to the same amount of security across different coins. I trust ETH transactions just as much as I trust BTC, and there are others that I trust equally also.

And now that BTC costs more in fees, their solution is to move to less secure off-chain layers. I'll take on chain ETH, monero, BCH, and a few others any day over a low security off-chain solution like what BTC offers. I may hold BTC, but I'm not buying it, and I'm not using it.

I would like to point out that the lighting network isn't "less secure", it's cryptographicly sound. It is less secure in the sense that you need to keep your lighting node online all the time, but this can be solved with other relegated solutions.

This is plain wrong. Lightening is not on chain except at entry and exit points. By nature this means that anything that happens between is less secure because it has not be settled publicly. Until something has been printed into the blockchain with enough proof of work behind it to be considered immutable, it is simply not in the same domain of security. Comparing these two states is a joke because one of these states is ignoring the very reason why bitcoin was invented (rapid public immutable settlement). Until that settlement has occurred, you must do the very thing that bitcoin was created to remove... trust third parties.

That does not sound correct on the lightning network. I cannot find any reason it would allow someone to run off with your funds. Can you explain the attack?

In order to use the lightning network, you are making a smart contract on the normal slow bitcoin ledger with another party. You need to insert a pre-determined amount of money into this system to be able to spend it.

Most consumers will not want to do this because they don't know who they will be buying from and don't know how much they are going to spend. If they decide they want to buy from random website Z one day, making a lightning network connection between themselves and that website will be useless because establishing a lightning network connection needs to go on the regular slow block chain, so it'll take as long as a regular payment.

The only way it works is if a major exchange like Coinbase sets up lightning network agreements between other major exchanges. Then the user puts bitcoin into Coinbase and uses the Coinbase website to make lightning network payments on your behalf. This means you will probably put all your money into an exchange like Coinbase who can then run off with your money.

Okay, but the lightning network is not just "here's all my money, please don't run off". There's crypto around it, a multisig protocol designed to prevent exactly that, no?

If it was just some crap exchange, why would anyone be bothering, and especially why would it be taking so long?

Um no, you deposit all your bitcoin into Coinbase's bitcoin wallet so there's no crypto enforcement that you own the bitcoin anymore. That's the only way Coinbase can safely use your bitcoin on their lightning network connections.

They need to commit regular bitcoin into a lightning network channel on the main blockchain before they can spend it.

That's not lightning. That's just how exchanges work now.

That is how exchanges have always worked, but the lightning network encourages people to use centralized exchanges that already have lightning network channels between all other exchanges and sellers so they don't have to waste time committing unknown amounts of bitcoins between 100+ vendors.

> If they decide they want to buy from random website Z one day, making a lightning network connection between themselves and that website will be useless because establishing a lightning network connection needs to go on the regular slow block chain, so it'll take as long as a regular payment.

The "network" part of lightning network allows you to make a payment to anyone else on the network (trustlessly) through multiple hops. For example, if you opened a channel with a friend, and your friend had a channel open with Starbucks, you would be able to pay Starbucks through your friend without having to trust your friend.

And this works today. I have one channel open on the lightning network, and regularly make payments to random nodes on the network. I pay << $0.01 in fees for these transactions, and they settle in ~a second.

So what happens if you aren't linked even indirectly?

What if the trust chain is 20 users long?

Who's going to spend the resources to find this chain? And has it been tested to scale, or is it just like Bitcoin in the early days where people said it would scale and never tested it?

People are going to gravitate to putting all their money into a large exchange like Coinbase to avoid these problems.

There is no trust chain. You have no idea how lightning works.

Every hop only knows the previous and next one, nothing else, not even the length of the chain.

Your computer/phone finds the route and if you’re ok with privacy trade off you can outsource this task to your hub for a fee in single satoshis.

People are going to gravitate towards convenient solutions making all sorts of security, privacy and financial risks, that’s why it’s important to have multiple alternatives with different trade offs.

There is a trust chain. If there was 0 trust involved then you could pick random nodes instead of needing to find a specific path. Sure there's cryptographic trickery involved so it is harder to steal the money, but it can still happen.

Large centralized lightning nodes will form and they will act like banks and be able to stop payments to arbitrary entities like Paypal, and even make the payment get stuck for days or weeks depending on how long the timeout is.

From the paper itself.

"the network will look a lot like the correspondent banking network"

No, you’re flat out wrong. There is no trust and it’s impossible to steal money.

You need to find a path because path needs to have channel capacity required for your transaction, that’s why you can’t pick random nodes, not because you want to trust them.

Stealing somebody’s funds from channel is just as hard as stealing bitcoin - you need to crack encryption or you need a private key.

You have no idea how lightning works, please do some reading before making incorrect claims.

Nope, you are just echoing the basic understanding of the tech that is on every bitcoin propaganda outlet.

Each node can still steal money from the node next to them by turning hostile such as not passing the secret and closing out their channel and getting time locks to expire, which is many magnitudes easier than cracking encryption or stealing a private key. The network has already been DDoSed. Now imagine the incentive to do it when it handles millions of dollars.

Not passing the secret means not spending the output in HTLC, which will then be claimed back by sender due to timeout.

You can unlock your HTLC without telling the other party the secret. You just close the channel by yourself and tell everyone else the secret and then ddosing so the other party can't redeem his HTLC before his HTLC expires.

> You can unlock your HTLC without telling the other party the secret. You just close the channel by yourself and tell everyone else the secret

Do you not see the contradiction in your statement? You can’t spend your htlc without revealing the secret to the other party. If your stealing procedure involves ddosing random unknown party from being able to access any internet whatsoever to lock them out of minuscule amounts lightning payments are used for - good luck with that.

Nope there is no contradiction. Again you haven't really thought about how this works. You are guaranteed to know who sent you your HTLC because the payment is settled through the channel. You cannot cloak this in any way.

You only need to ddos the lightning network so it won't record their HTLC when it expires. There is no need to ddos them, though I would argue it would be equally easy to do. And so while the other party will eventually see the secret on the block chain, they can't get their HTLC unlocked on the network before it expires.

And the lightning network has already been easily ddosed! We will see it happen more often if the lightning network ever gets bigger. This isn't some hypothetical situation.

> ddos the lightning network so it won't record their HTLC when it expires

this is not even a thing. lightning network is collection of independent nodes, there is no registry of payments and channels, every node keeps their own record.

> And so while the other party will eventually see the secret on the block chain, they can't get their HTLC unlocked on the network before it expires.

why? if a party is stuck in a payment and especially if they notice they are getting attacked - they will take steps to mitigate. the moment commitment with secret is broadcast and in the worst case - the moment it gets confirmed on chain the other party will know the secret and can unlock their htlc. the risk is essentially the same as with deep reorgs - very low because those are very rare events. and the hassle is most definitely not worth the value of a stolen micropayment, which is the primary purpose of LN.

It doesn't work that way. If Coinbase operates a Lightning node, you simply open a channel to their node, from your Lightning wallet. You don't give Coinbase your money.

Thanks for writing this. It's the first time I've understood a "lightning won't solve scaling problems" argument. If you need a payment processor to process your payments, then you might as well use a payment processor in an existing fiat currency... It's a fair point.

You've been misled...

Edit: Here is an example of a lightning network enabled wallet that works as a lightning node itself. No payment processor needed: https://lightning-wallet.com

I'm not sure how this helps. Clearly I can always run my own lightning node, but the person I'm paying has to accept the off network payment from me. That's fine for some kinds of transactions (like my employer paying my salary). But how am I supposed to convince a retailer to accept my off network payment? Similarly, if I want some random person to send me money, how do I convince them to run a lightning node and to let me connect to it? I mean, we can't even set up VOIP connections without a middleman these days because of firewall issues.

I think there is a fair amount of logistics to work out here. I may be wrong, and that would be great, but at the moment I'm not quite seeing how this is going to work in practice.

It won't work, it's already done. There used to be support from major companies like Dell, and "core" blew all that support on a convoluted, over-engineered nightmare while choking the main network. The momentum is gone and it's never coming back.

I'll be deeply surprised if lightning is ever used at the retail level by more than a couple shady internet casinos and John Galt themed vape stores.

> But how am I supposed to convince a retailer to accept my off network payment?

Because you're both connected to a common lightning node, at least indirectly. There are several public lightning nodes that accept connections and they are joined together. Connect to one and you have access to their network.

> Similarly, if I want some random person to send me money, how do I convince them to run a lightning node and to let me connect to it?

Same as above. The lightning network is, of course, only going to work with other nodes who are also connected, but there is an incentive to join, since when you do, you can transact with all of its peers.

It's been quite a long time since I read up on how lightning works, so maybe I'm misunderstanding something. At least in the version of the protocol I looked at the node that connects to the Bitcoin network has to promise to pay. It's literally a payment processor because they need the capital to guarantee the payment. Is this not the case any more? If it is, then I'm still in the same boat where the Bitcoin network facing node can decide not to service me.

> If it is, then I'm still in the same boat where the Bitcoin network facing node can decide not to service me.

Jesus, this is precisely what the lightning network is for! Please, do a little research before dismissing it wholesale. The node you are connected to cannot refuse to pay, because you have a signed, but unpublished transaction of what he owes you (if he owes you). If he suddenly stops contact and says fuck you, you can just publish that transaction.

Seriously, this has always been the idea from the get go, so don't go blaming that you haven't read the latest update on how lightning works. If you don't understand something, don't go around spouting that it must not work.

No offense, but could you please calm down? I have actually done some research -- to the point of knowing how to implement lightning. However, it is a long time ago and things may have changed. I'm asking you for more information because you seem to know more than me. I'm not trying to shout you down. I'm trying to understand what I'm missing. It's totally fair if you don't feel you have the time or patience to explain it to me, but for both your sake and mine, it's probably best if you just say that and I won't bother you any more.

"Bitcoin (and all other cryptocoins) are just not usable for mass transactions. Buying anything with bitcoin takes over 10 minutes"

Anybody here know the currency Nano? The ORV (Open Representative Voting) consensus mechanism makes it possible to find decentralized consensus in ms without the power-intensive PoW.

For a first impression, you can perform a real on-chain transaction here: https://nanospeed.live/

Reply on news.ycombinator.com