[netcruiser]

Sounds like its based on a flawed assumption:

"Now, immediately after a block is found, there will be no more transactions in the network to be claimed by a miner making the next block"

The fact is there is usually a large backlog of transactions.

[Paul-E]

This paper was published when there were big fights going on about whether the block size should be increased. Om camp argued for the status quo, which was causing high transaction fees and long backlogs. The other camp argued for a larger block size, which would reduce or eliminate the transaction backlog. Eliminating the backlog was generally considered a good thing, and the arguments were mostly around the block size, but this paper essentially came in and said "wait a minute, we might actually need the backlog once the block reward is 0."

[cyphertruck]

Not to disagree, but to emphasize that the idea that the only way to increase bitcoin capacity is increasing the block size is in error. You didn't push that idea but others do.

Nearly every release of bitcoin over past 8 years has increased capacity or efficiency, often both.

[tromp]

The point remains that eliminating the backlog of high fee paying transactions would jeopardize mining stability and security in the long term.

[uncletammy]

I think this is depends on how large the network has scaled at the time the block reward dies. If BTC is doing 100x visa levels, each transaction fee can have a tiny fee and mining would still be viable.

That being said, I don't BTC has any intention of scaling to that degree (or at all really). The BTC devs seem much more concerned with building second layer products like lightning network. All these products actually reduce the number of on-chain transactions and by extension, miner revenue.

[imtringued]

The problem is that lightning is basically federated Paypal. You have to trust your specific provider to some degree. If he stiffs you, you can get arbitration on-chain. There are almost zero barriers to entry so Paypal can be replaced by Paypal 2.0.

I don't believe for one second that when you increase the size of the audience that anyone will care which provider they use. In practice that means they will just use Visa, Mastercard and Paypal, maybe even by skipping lightning entirely.

[AaronFriel]

How many kWh does one transaction consume?

Is that higher or lower than 8 years ago?

[thebean11]

Number of transactions is the wrong denominator for energy usage of the network. The same number of transactions could be confirmed by a single miner.

The miners contribute to network security (double spend protection), not transaction throughput. Common misconception.

[ximeng]

https://www.statista.com/statistics/881541/bitcoin-energy-co...

"The average energy consumption for one single Bitcoin transaction in 2020 was 741 kilowatt-hours. This was significantly more compared to the cumulative 100,000 VISA transactions with only an energy consumption of 149 kilowatt-hours."

[drdeca]

I don’t think individual transactions are meaningfully tied to an energy cost.

[sjy]

Why can’t you estimate the power consumption of the whole network and divide it by the number of transactions per second?

[xaxiomatic]

You can. But it is not a good metric. Assuming the difficulty adjustments ramp down you could still mine blocks with a couple of raspberry pis.

The energy consumption and capital investment in hardware is Bitcoins security model. As it would require you to put in the same amount of HW and energy to subvert the mining process. Probably we are at a point that this is almost impossible other than a state actor or a global conspiracy.

The block size increase/decrease is also tied to security. There are latency implications as well as the fact that some nodes might drop of if you increased the block size.

In both cases it comes back to security rather than the transactional throughput. I am a huge proponent but still kind of struggle with the idea of how much resources this thing sucks up. But then again if it truly becomes the worlds ledger for wealth preservation... idk ... might be worth it.

[imtringued]

Because that would imply linear scaling. There is a fixed energy cost that doesn't depend on the number of transactions. It goes up and down all the time.

[drdeca]

The choice of the word “meaningfully” was deliberate . Of course you can do that. But I don’t think it is meaningful. The energy use isn’t a result of the transaction.

[troysk]

Feels like natural progression as underlying technology (internet speeds and HDD space) improves. So more network traffic can now get synced across more nodes faster.

[jtsiskin]

I think the analysis still hold even with a transaction backlog, as long as the fees in these transactions are exponential distributed (a few very high value txs, then it trails off quickly)

However I have know idea if this is close to what it really looks like.

[cyphertruck]

The large backlog these days is mostly spam transactions with almost no fees. Bitcoin has increased capacity dramatically in the past decade.

[tromp]

Wrong; there is a large backlog of transactions paying over 50 sat per byte [1], which is over $10 in fees for most transactions.

[1] https://jochen-hoenicke.de/queue/#BTC,24h

[iso8859-1]

Why is it $10? According to https://www.buybitcoinworldwide.com/fee-calculator/

Fee rate: 51 satoshis/vbyte

SegWit transaction with 1 input, 2 outputs: ~172 vbytes

In total: ~8772 satoshis == ~4.48 USD

[tromp]

Because the average transaction is over 400 bytes in size [1].

[1] https://bitcoinvisuals.com/chain-tx-siz

[uncletammy]

https://bitinfocharts.com/comparison/size-btc.html

Make note of the inflection point around 2016. Yeah, it's right around the time your capacity chart starts looking like a wet noodle.

Yeah, that's when a few your scumbag BTC devs removed Gavin Andresen's commit access and killed BTC's last chance of becomming a real currency.