I think the issue is that people have a vision of what Bitcoin is, that doesn't align with the reality of a proof-of-work system with a computation bound hash. The block size is basically orthogonal to the proof-of-work, and the proof-of-work is distributed in the sense that anyone can do it. There is nothing stopping any entity from doing the proof-of-work if they have the resources and capabilities. So this idea that there will be one company that is the only one doing Bitcoin doesn't really make sense if it's cost competitive for other companies to be involved, regardless of the computation or bandwidth requirements.

End users can still validate transactions via SPV, which is completely valid and provable without storing the entire block chain. Consensus relay nodes can do the same thing, although they will start dropping off the network when the bandwidth costs become prohibitive. So in a small block system we have an extremely limited amount of space for transactions, which means the cost will eventually be excessively high and result in settlement between big players. That's the choke point that allows for KYC and AML. You can submit a transaction to the network, but you can't afford to because there isn't enough bandwidth to service your request so you are forced into regulated channels, onto exchanges, or through third party choke points that can be audited. To a certain extent this is already happening.

People who want small blocks for Bitcoin really need to be thinking about using an Alt coin, it was never in the plans to have an extremely low bandwidth pipe to keep things decentralized when the proof-of-work is by default a centralizing force.

fraud proofs don't work because if you want to prevent someone from creating a fraud proof, you never send them the data that would be necessary to prove a fraud ("the block is too big! nobody (but some list of companies) have the bandwidth to support the enormous bitcoin network!"). SPV users don't require (nor check) the block data-- exactly the kind of scenario where you need in place so as to hide fraudulent double spends or other rule violations.

> the proof-of-work is by default a centralizing force

sounds like a good argument for regular pow change hard-forks. also to be watchful for economies of scale-- but economies of scale can hide pretty well i think, even after significant capital accumulation, it's an interesting problem.

here's a recent one: https://github.com/bitcoin/bips/blob/master/bip-0180.mediawi...

But why would a miner give an SPV user the data that proves a fraudulent double spend in the same block?

2. There is no non-blockchain alternative to online censorship resistant cash. So if the price to use it is high, you will either pay the fee for censorship resistance or use your credit card (which is faster, cheaper, gives you rewards or cash back, provides consumer protection, etc., but is not censorship resistant). Altcoins basically act as market forces which allow us to discover what the world is willing to pay for censorship resistant digital cash, and you see this when markets which actually need censorship resistance (eg. Silk Road, AlphaBay, etc.) talk about considering Monero or something else when Bitcoin's fees get too high.

2. There are already alternatives which have better censorship resistance than Bitcoin, limiting the block size does little to fix those issues. I also don't think the market actually values censorship resistance in their digital cash as much as you think.

Alphabay processed $600,000 to $800,000 per day [1]. This is economic activity for tangible goods. In the list of other top uses of the blockchain might be ICOs, which are a way to leverage censorship resistance to raise money without permission from the SEC. On the high end, these accumulated hundreds of millions of dollars within minutes.

Please show me a use case for a blockchain-based cryptocurrency that even comes close to these numbers and that does not need censorship resistance.

[1] https://www.wired.com/story/alphabay-takedown-dark-web-chaos...

The estimate of daily volume provided by blockchain.info obscures our discussion, especially if you consider that the daily volume for BTC/USD pair is around $1B. There could be $1B-worth of outputs in a day all dedicated to providing liquidity to exchanges, but entering/exiting the network is not an application of the network itself.

As for anonymity and fungibility: I believe tumbling services are good enough today, which is why buying on a dark market works. However, it is not perfect and this is why projects like Monero, ZCash, MimbleWimble, TumbleBit and Xim are important.

This isn't the most accurate thing, but looks to me like several millions USD worth of volume just in Bitpay. That doesn't account for all of the other payment processors, or the peer to peer payments, or the exchange trading, etc. I'm not saying dark markets aren't an early driver of Bitcoin, but it's not where the majority of the action is now at all.

Exchange trading is not really using bitcoin, it is buying/selling bitcoin. This is like saying buying a car is an application of cars...

Having said that, Bitpay is processing more than I thought. The data in that article would mean it's about $5M per day... I'd love to see a breakdown of this to figure out why people are spending bitcoin for everyday things if it'd be better for them to just use a credit card (to get rewards, to not have to hold a volatile currency, for consumer protection, for faster payments, etc.)

That and you can buy cash-like goods like gold bars which would be prohibitively expensive or impossible to purchase via credit card.

Bitcoin is almost 10 years old. We have enough data now to be able to build strong cases for what it's actually useful for. It is no longer sufficient to talk about what Bitcoin will do without talking about why it has not done so yet.

Bitcoin as a high throughput transaction network for buying things like coffee is not a good goal. It should be for much larger transactions, but optimized for decentralization and sound money.

Derivative blockchains can be spun off of Bitcoin, and settled periodically to the "reference" blockchain, which is Bitcoin, through lightning networks or other off chain developments (perhaps some sort of forked derivative pegged chains).

Bitcoin only has value as either a distributed non-centrally controlled international settlement network for countries and large banks [0], and/or as a digital Gold.

Optimizing this value out of bitcoin to achieve transaction throughput destroys the necessity of Bitcoin for existing.

Bitcoin proper should remain resistant to change, distributed, unable to be coopted or controlled by a few organizations, and above all, to have the properties of sound money (so resistant to changes to expand the money supply). If it doesn't have these qualities, it has no reason to be valuable. It cannot be all things to all people, but it can to "one" thing very well.

[0] https://thesaifhouse.wordpress.com/2017/05/19/economics-of-b...

Maybe Ethereum will pick up the slack on the transactional side and Bitcoin will be only for "investment".

I am not saying we should never increase the block size, but I am saying we should try all alternatives that make the network more efficient (more tx throughput for same total resource cost) without making it harder for users to validate the blockchain.