Hacker News Reader

Comment by JonathanBeuys | Hacker News Reader

What the article is missing is the incentive structure.

For example: "The number of entities sufficient to disrupt a blockchain is relatively low: four for Bitcoin".

This probably refers to the 4 largest mining pools.

Yes, they could temporarily annoy some people by not processing their transactions. But it would come at a giant cost to them. They would be out of business quickly as miners can switch to a different pool in minutes.

That is like saying the US economy is brittle because if Walmart, Apple, Amazon and ExxonMobil would just send everybody home and set their premises on fire, there would be a problem.

View on HN

Completely agree. It's impossible to inspect any aspect of Bitcoin's design without considering its incentive structures. They are the innovation in my mind. Every blockchain/crypto/Dao/<insert buzz word> forgoes one of Bitcoin's perfectly balanced incentive structures in favor of some "innovation" (speed, programability, founder/VC enrichment) which always inevitably leads to centralization and exploitable risk somewhere in the system.

Every time it happens it hardens my opinion that a L1 should be like the foundation of a building - it should have one job (be hard money), it should be simply and elegantly designed, it shouldn't have any cracks (a.k.a misaligned incentive structures), it should evolve very slowly if at all, and it should be designed to be built on top of. L1 exists to anchor L1+N to the real world and that's it.

> simply and elegantly designed

I don't think Bitcoin qualifies as such. Sure, projects like Ethereum are far more complex, but there are projects that significantly improve on Bitcoin in simplicity and elegance. Bitcoin script in particular is not that simple and full of warts.

I think by definition it does, which is to say, its popularity proves it. I just think there's a hindsight thing going on.

I liken Bitcoin to e.g. the Model T or the first grey brick cellphones. It changes everything forever because it works very well, but also is quickly improved upon and thus eventually won't be that popular.

No other system/project anchors a digital currency to the physical world as effectively as Bitcoin. But I do agree, eventually L2+ will be much more popular, in so far as much more people will interact with them. That's kind of the idea. L2s will still always rely on L1 for its anchor to the physical world (and thus physics).

I'm unfamiliar with L1 and L2 as presented here. Do these both refer to the Bitcoin blockchain?

If so, no. I strongly predict Bitcoin, both the blockchain and the currency will nearly, if not fully, "die," as the Model T and grey phone have. Ethereum is already much better at it's core, and it's not hard to think even Ethereum will be improved upon.

In the long run, literally all bitcoin has is history and name recognition. Like the Model T, you will see it revered like something in a museum, and also just about nobody will actually use it.

> Ethereum is already much better at it's core

What you and many others seem to overlook is first mover advantage, network effects and most important: trust. Bitcoin is not Yahoo or Facebook, it is not a company product which has to improve and break things all the time.

In the long run Bitcoin is a store of value. That requires trust, which takes time. To gain trust over time it has to move slow and steady because there can be no mistakes.

When it comes to money and trust it is almost impossible to get a second chance (see Terra UST). Just one mistake could result in everyone distrusting Bitcoin.

What I think others overlook is that in practice (and despite what the retconners are trying to say about bitcoin is gold or some nonsense) -- is that Bitcoin was NOT designed to be a blockchain platform. It was designed to be fungible currency. And it's an excellent proof-of-concept...but with severe flaws that will kill it, especially in light of the fact that we're talking about *open source software, which begs to be forked, and effectively has been.

Your argument strikes me as very odd -- you seem to suggest that nothing can beat Bitcoin in theory, but Ethereum is already beating it in practice.

> Your argument strikes me as very odd -- you seem to suggest that nothing can beat Bitcoin in theory, but Ethereum is already beating it in practice.

In theory everything is possible. I do not follow the beating really. They can both exist with different use cases.

Bitcoin wants to be digital gold (and money) and Ethereum wants to be a smart contract world computer running dapps. They solve different problems.

But Ethereum is a technically superior option for both cases. It does what Bitcoin can't do, but it also does what Bitcoin does better, and has a roadmap to improvement beyond that. Bitcoin is still wildly slow and destroys the environment, and Ethereum is on track to mostly not do that, to say nothing of "post-Eth" projects like Pulsechain.

Now, I'm sure Bitcoin will enjoy name recognition for some time, but in my experience, always bet on the tech.

Bitcoin slow? I make virtually free, instant, micro-payments on Bitcoin hundreds of times every day (Podcasting 2.0). I also accept Bitcoin payments on my e-commerce site that settle instantly for fractions of a percent in fees. I also route hundreds of Bitcoin payments per day using a raspberry pi which draws an extremely low amount of energy. Not sure which "reality" you're referring to...

Man, your use case is one I hope I'm wrong about. I love the idea of microtransactions and tipping and such. But, I never much paid attention to Lightning given that the BTC price didn't really move when it came out, which strongly suggested to me that not enough people really cared.

It's impossible to consider Bitcoin without considering lightning. Two sides of the same coin. It's still early in its adoption curve for sure, but it has been gaining traction.

- Twitter uses lightning for its tipping mechanism

- The Podcasting 2.0 standard is all lightning based

- NCR is integrating lightning

- Blackhawk is integrating lightning

- Kraken has enabled lightning

- Robinhood is integrating lightning

- Square/block is investing heavily into lightning development and enablement

I'm probably missing some big ones...

I'd suggest doing more research before commenting about things you don't understand. Layered Money by Nik Bhatia would be a good place to start.

The concept of layered system design, especially in money and technology, is not new.

Ethereum is an inherently flawed system.

I'd suggest providing actual ideas, or better yet practice, instead of reading assignments?

The reason I know more about Ethereum than Layered Bitcoin stuff is that that is what more people (myself included) are actually using today to do cutting-edge stuff on.

Moreover, it's pretty easy to first-principle your way into why it's probably better; layered Bitcoin saddles you with a particular kind of technical debt -- you're literally trying to hack Bitcoin into something it wasn't intended for, and the Bitcoin code is of course mostly immutable. Ethereum, on the other hand,is built for it.

Sure, all software has problems, but a car built from the ground up is likely going to be better than a Model-T with addons.

(edit)

Bahaha, I checked out the guys book. He's a Bitcoin Maxi who apparently quotes -- Nassim Taleb aka the guy who thinks that BTC will go to zero. Not looking like a worthwhile thing.

Calling Lightning a "hack" is like calling TCP a hack of IP. First principles of protocol design literally dictate layered modular stacks.

Actual ideas:

Bitcoin is layer 1 of the internet of value. It only needs to be secure, decentralized, and stable. It has one job; enforce digital property rights while maintaining the above characteristics (Think the internet layer: IP)

Layer 2 (lightning, liquid, etc) build on top of the secure, decentralized, stable foundation of L1 Bitcoin and create standards/protocols for application/user comms (think the transportation layer: TCP, UDP)

Layer 3 (Strike, TBD, Sphinx, CashApp, etc) applications to fulfill specific use cases built on top of layer 2 (think the application layer: all the things you do on the internet)

Which projects are simply designed to be hard money?

That depends on what do you mean by hard money. A capped supply?

well, there are multiple factors, supply cap is one. Unchangeable issuance schedule (i.e. unchangeable monetary policy) is another gigantic aspect. No other project can guarantee the same immutability because no other network is defended by the bitcoin mining network.

Bitcoin is the only commodity on the planet who's supply is completely unaffected by demand.

I already see discussion on the bitcoin dev mailing list about how to deal "with block rewards being too low to maintain acceptable security" in the distant future [1]. We can expect more of such discussion as the block subsidy keeps halving while still dominating the tx fees [1].

An emission like 1 per second forever is much more immutable due to guaranteed miner incentives and lack of arbitrariness. And thus a better basis for hard money [2].

[1] https://lists.linuxfoundation.org/pipermail/bitcoin-dev/2022...

[2] https://john-tromp.medium.com/a-case-for-using-soft-total-su...

lol, new here? That discussion has been going on for YEARS. Discussion doesn't negate anything I said.

Bitcoin is a terrible foundation to build transaction processing on, since it is extraordinarily slow - it would fail to handle the needs of even a small village.

The "solution" of slapping something like LN on top of it is "let's completely forgo all of Bitcoin's guarantees and build something else; while still being hurt by Bitcoin's cripplingly slow speed when creating our network".

Also, the incentive structures being discussed are only meaningful as long as the miners believe that Bitcoin will keep increasing in value. If it looks like BTC is dying, they have every reason to try to squeeze the last few drops of value from it by running an attack on it and running away with what real money they can extract.

Fedwire is a terrible foundation to build transaction processing on, since it is extraordinarily slow - it would fail to handle the needs of even a small village.

The "solution" of slapping something like banks on top of it is "let's completely forgo all of the Fed's guarantees and build something else; while still being hurt by Fedwire's cripplingly slow speed when creating our network".

Also, the incentive structures being discussed are only meaningful as long as the Banks believe that the Fed will keep increasing the money supply. If it looks like USD is dying, they have every reason to try to squeeze the last few drops of value from it by running an attack on it and running away with what real money they can extract.

> completely forgo all of Bitcoin's guarantees

ah, he doesn't understand at all how lightning works. He's just repeating sound bites.

It's the other way around: I took a look how Lightning works, and understood that, unless you open and close channels constantly, you don't interact with Bitcoin at all. Repeating soundbites you've just heard without understanding the system would be saying "LN fixes Bitcoin's speed problem".

While a channel is still open, transactions happening on the channel aren't in any way protected by Bitcoin's guarantees (they get LN's guarantees, which are different). And those are the only transactions that have any speed boost.

It would be easier even than that, observe what happened in 2017. Bitfinex said they would allocate the BTC ticker to the present BTC chain no matter what the hash power indicated was the actual canonical chain and that was the end of it.

Contrary to popular belief, miners aren't choosing anything when it comes to BTC. They're rubber stamping the decisions of the BTC core developers, who are blessed by the exchanges such as Bitfinex, and they are collecting fees on that rubber stamp.

That sounds like the right thing to do?

Why would the hash power decide what is called BTC? If Dogecoin some day attracts more hashing power than Bitcoin, then Bitfinex is supposed to assign the BTC ticker to Dogecoin?

We and a group of partners run a restaurant together. One day you decide to turn it into an apothecary for the dispensation of poisons, and keep the name and pretend that it's still a restaurant.

Any mechanism which rubber stamps the above process as legitimate and collects a fee for doing so is corrupt, letalone over the objections of an enormous fragment of both customers and partners in the restaurant in question. Lack of objection or approval from those whom the change is in the interests of, or who are simply too ignorant to understand what is going on, don't change that.

That is effectively what happened with BTC.

The way that it was supposed to work would be that the hashpower would choose the canonical tip of the chain if there were two conflicting ideas about what the chain should actually be. BCH stuck to that ideal and through two contentious and many more peaceable forks proceeded according to it. BTC did not.

Hash power in the context of potential forks in the BTC chain, not a different coin.

You're both already explaining the reasoning through your wording, "present BTC chain" and "potential forks". If you fork a project, even if yours becomes more popular, that doesn't mean you get to take over the original name.

> If you fork a project, even if yours becomes more popular, that doesn't mean you get to take over the original name.

That's exactly what happened with BTC. And yet almost everybody calls it Bitcoin.

BCH changed the blocksize and thus became incompatible with the original network. If you're claiming that the project that kept the blocksize the same and remained backwards compatible is the fork, then I'm done here and I wish you the best of luck.

It was the original plan to change the blocksize all along.

Satoshi said in 2009: "The existing Visa credit card network processes about 15 million Internet purchases per day worldwide. Bitcoin can already scale much larger than that with existing hardware for a fraction of the cost. It never really hits a scale ceiling. If you're interested, I can go over the ways it would cope with extreme size."

Some latecoming BTC core devs cancelled that plan, with zero justification, changed the consensus mechanism for the chain from hashpower to fiat declaration by fraudulent centralised exchange running a pump and dump scam with fake money, and sabotaged the entire project in the process into complete uselessness, and to this day we're stuck with people who believe a fundamental financial revolution is going to take place on a chain that was forced to 4tx/second and get very angry, confused or surprised when you point out that such a change is clear and obvious sabotage.

If you're claiming otherwise, then I'm done here and I wish you the best of luck.

I think you're missing some context about (what is now called) Bitcoin Cash.

Look up the debate between big blockers and small blockers if you're interested in learning more; a quick summary, many people wanted larger block size so bitcoin could process more transactions.

I believe several newer core developers staged a coup to prevent this from happening, despite several of the longer-term contributors being in favour of it. The original contributors (sans satoshi, who was mostly inactive at this point), had their access to the git repositories revoked.

This video goes into a bit more detail (https://www.youtube.com/watch?v=XfcvX0P1b5g) but basically the people opposing a larger blocksize had financial incentives.. a lot were in bed with a company called Blockstream which benefits from their development on sidechains (now the lightning network). Had the block-size been increased, Bitcoin would be much more useful as a means of exchange (more transactions would be processed, and they would be cheaper).

This is where Bitcoin Cash is now, but it doesn't get the BTC ticker symbol because Bitfinex and other CEXes have a lot of influence as well.

> Had the block-size been increased, Bitcoin would be much more useful as a means of exchange (more transactions would be processed, and they would be cheaper).

Of course the flipside of this is that it would become very expensive to run a node. Bitcoin's blockchain is ~400 GB after 13 years. BCH has a 32x larger blocksize cap, so we can extrapolate that if anyone actually used that fork, it would grow a 12 TB blockchain over that same period. As a result running a node would become expensive and the network would become much less decentralized (the very thing people are complaining about in this thread!)

This would be true if the number being 32x'd was actually high, but that is not true. It was 1 megabyte. 32 megabytes over ten minutes is not even 512kbps. Other chains that process much more throughput, as well as BCH itself, has been benchmarked on very modest hardware pushing that boundary and higher.

Simple fact of the matter is that the whole justification for the permanent 1mb limit was an outright indefensible lie, pushed on idiots, and supported by censorship. Nothing more, and nothing less.

You're completely right.

Hashpower between chains is only relevant if they are compatible (same consensus rules), as the accumulated proof of work is used as an indicator of which chain should be followed. Once a chain hard forks to alter the consensus rules like Dogecoin, Bitcoin Cash, etc, there is no amount of additional hashpower that would make up for the incompatibility -- a node will simply ignore the invalid blocks.

That's fair enough, but I believe the paper stayed in its bounds; it is correct that given the means or the willpower control could be assumed. What you should be more worried about than blatant double-spends are all the hidden ways having 51% of network control allows you to cartel, centralize, and siphon funds away from smaller, newer players in the system.

https://saito.tech/the-double-spend-attack-is-not-the-same-a...

But those companies only exist because the US regulatory framework allows them to exist as long as they follow the rules. If the CEO of Walmart sent out an email saying that everyone was fired and that all stores should be demolished, he would be fired by the board of directors within hours. Also other executives would respond to the email telling everyone to ignore the previous email while they see if the CEO was hacked.

The entire point of crypto is to bypass the regulatory framework. In some cases that may make for a more robust system in some cases it may not.

Yes they’re incentivized not to, but when push comes to shove, they can basically choose their own form of “truth” and ignore/disrupt/revert transactions and others would need to fork the chain. It’s not a given which would succeed.

This has happened before

No need to fork the chain. Miners can just switch pools.

Valid transactions are enforced by nodes, not miners (there are WAY more nodes than mining pools). All of a miner's incentives persuade them to mine on the longest chain, lest they lose out on potential profits. A group of nodes and miners can certainly fork if they please, but the network simply won't care. e.g. BCH/BTC

Reply on news.ycombinator.com