Hacker News Reader

Comment by BIackSwan | Hacker News Reader

Genuine question - What about a "cashless" currency based on blockchain? (a la bitcoin). Wouldn't it be and not be under any single entity's control at the same time?

View on HN

One lesser point, bitcoin is pseudo-anonymous: you can always follow the money even if you can't identify all the people who handle it. It might be argued then that any entity who knows what you're spending and where has plenty of control over you. There are anonymous coins that don't have this property.

The biggest danger with blockchain is that if it ever fails, it would fail for technological reasons and its failure would be instant and complete. Like encryption being broken. Instead of government money, where the failure is telegraphed years in advance, e.g. by reckless abuse of the printing press.

The equivalent of breaking encryption on bitcoin is perfectly forging paper money, or breaking the security on any financial institution that enables you to steal all their money.

Also, if someone can break sha256, not only is bitcoin screwed, but so is the Internet as we know it and every secure institution period. The overnight "instant and complete" breakage of fundamental assumptions about security would destroy everything. So interestingly, if you could break sha256, you could probably break into almost any financial institution and steal their money as well.

The security of a blockchain is in its popularity. As long as the code is open (and I'll admit, I am not a fan of how bitcoin mixes the default interface (BitcoinQt) and the default implementation (BitcoinCore) together so much) and there is investment in the market, there will be strong financial pressure to consistently and diligently analyze and track the source code for weaknesses.

Bitcoin at 3B has at least a dozen security firms that have looked into its crypto. If you had another crypto that broke a trillion you would have thousands of companies constantly safeguarding and testing the crypto of the currency in the same way Google offers ludicrously high bounties for anyone who can break Android/Chrome/ium security.

Not necessarily. What we're seeing with Ethereum, for instance, which seems to be the main blockchain technology that banks want to use right now, is that it was "customized" to fit the banks' needs to the point where the blockchain used by the banks doesn't even need to be public, the way Bitcoin is. It also doesn't need to work based on the "consensus" mechanism. And it could be further modified so that each transaction can also be wiretapped/blocked directly by the NSA, and so on.

I don't necessarily blame Ethereum for this, although they did have a hand in following the banks' rules and essentially writing software for them, but this would've probably happened even without Ethereum agreeing to do it. Maybe the banks would've forked Ethereum's technology and then they could've done whatever to it, or they could've written their own tech from scratch.

Cash and current forms of electronic payment have the nice feature that the seller gets their money almost immediately. As I understand it, bitcoin transactions take a while to confirm. I wouldn't want to stick around in the shop until some miner got around to my transaction.

Small payments using BitPay require one confirmation, and that typically only takes a minute or so. Their fee (1%, I think) covers reversal risk.

Now imagine being lined up at the equivalent of a cash register with every single person in front of you taking "a minute or so" beyond what a transaction currently requires. Nobody's hiring additional cashiers (or whatever they'd call them) or self-checkout kiosk supervisors (to go along with the extra kiosks) if there isn't any extra money to be had.

You fire one cashier and install 100 self checkout machines. Problem solved!

And replace those cashiers with people who have other job titles to manage inventory shrinkage. There's no getting around the fact that you need eyes near the doors to make sure that everything gets paid for, and you need enough eyes to monitor the traffic: more parallel traffic, more eyes needed. (Anecdotally, I see more staff at self-serve checkouts than at cash registers in stores that have both, yet the cashier lines move more quickly than the "more convenient" kiosks.) I'm sure there's an ideal world living in some people's minds, but it's not part of the external reality and probably never will be.

I avoid self checkout - slow and inconvenient. But it's just a matter of improving the technology: we'll get there eventually.

I avoid self-checkout because the entire machine is usually set up around the assumption that I will try to walk out of the store without paying for one or more items, or by futzing the bar codes so that I pay $1 for a $30 beef roast.

The self-checkout technology will never improve as long as that assumption remains. If you can't trust the customer to not steal a pack of gum, you can't trust the customer to not steal an entire beef tenderloin.

The professional cashier does not need to keep every damned item on a giant scale after scanning it, because they are an agent of the store, and presumably trustworthy enough to notice and stop shoplifters and scammers. And they usually don't have to halt the entire checkout process and wait for a manager to come by and restart it whenever anything even mildly out of the ordinary happens, such as using a coupon.

You can't improve the technology when the premises are wrong. And the premise is that you can turn a grocery store into a glorified vending machine.

If we put an RFID tag on every item and use an RFID gate then we would just have to walk past the gate which is capable of detecting every item without you having to take it out of your cart or bag. A turnstyle could then force you to pay and disable the RFID tags. Cost and transmission distance of passive RFID tags are the major obstacles, I believe.

How do you detect RFID tampering or the use of RFID-blocking shopping bags or pouches?

Remember, the store is more interested in protecting its operating margins than in providing more convenience for the customer. If the technology does not allow a store exec to believe that it can stop a $30 steak from walking out the door unpaid, it will never be installed.

Where are you going to put the food if you fill up your entire grocery store with 100 checkout machines? Also spending a minute or so waiting for the checkout machine to clear my transaction still sounds like huge pain in the ass.

One confirmation on average is approximately 10 minutes. Most small value payments are acceptable with zero confirmation once the payment has been broadcast to the majority of nodes which only takes a few seconds.

Yes, you're right. No confirmation, just broadcasting.

The "takes a while to confirm" is a shortcoming of politics in bitcoin today. Blocks are limited to 1MB, there are more than 1MB of transactions almost all the time now, so some transactions (the ones paying the lowest fees) get delayed multiple blocks and take forever to be confirmed.

That will eventually be fixed, and once we are back to normalcy where transactions aren't limited by block size, you can regularly get - even on feeless transactions - multiple confirms in seconds. Your transaction won't be finalized until it enters the block, but once you have a half dozen or more confirms the probability of failure of that transaction entering is extremely low - well below 1% at that point, insofar as I have never seen it happen myself and don't know anybody who has had a multi-confirm transaction miss acceptance.

The blockchain is mostly a solution in search of a problem. Bitcoin is cool, but I doubt it will ever have large-scale applications beyond DNMs and hacker/hobbyists.

- The current biggest problem that bitcoin faces is one of scale - and there's no clear solution in sight. Why on earth would you want every little transaction on the blockchain? If I buy a stick of gum, that doesn't need to be on public record for eternity. It is impractical and unwieldy.

- Centralized systems are always more efficient than decentralized ones, and systems naturally tend toward centralization. Bitcoin falls down under centralization, which we are also currently witnessing with the hash power tied up by a few largely unknown Chinese operators.

There are proposed solutions like pseudo-centralization with LN where bitcoin just becomes a settlement layer for off-chain contracts - and the details behind LN are still so primitive it is not clear it will actually solve any real problems.

What you are talking about in practice for something at scale is a centralized blockchain of some sort, which is really just a distributed, immutable database, with some sort of trusted signing authority. What exact advantage does that give aside from potentially non-reversible transactions, which are a clear non-starter for most consumer purchases? There are already services that offer true non-reversibility for special cases where this is desired.

There has so far been no viable application for blockchain technology as a replacement for cash except black/grey market purchases.

In theory, yes. But currently, there exist enough structural problems, as much I know, that Bitcoin or similar schemes can not replace cash.

Further, even with a total secure system, there will always be the pressure of the governments to control it.

Reply on news.ycombinator.com