Hacker News Reader

Comment by daneel_w | Hacker News Reader

"... the BonanzaMines System is shown achieving 55 J/TH whereas the Bitfury Clarke achieves 40 THash/s at 56 J/TH while the Canaan Avalon A9 does 30 TH/s at 58 J/TH."

55 J/TH versus 56 and 58 J/TH from competing vendors. Small margins of course matter at scale, but this (preliminary) improvement in energy efficiency, less than 2%, isn't really impressive. Nor does it seem like an amount that can make a dent in the energy consumption problem of Bitcoin mining. Are they aiming at selling their product cheaper than the competitors' offerings? Doesn't sound like Intel. Am I missing something?

View on HN

> Nor does it seem like an amount that can make a dent in the energy consumption problem of Bitcoin mining.

If you make a dent in the energy consumption problem, doesn't the network automatically adjust to pop the dents out (increase the difficulty)?

It does if miners persistently aim at their maximum power allotment and keep investing in an equal amount of new hardware and extra hashrate. The pivot point where operational costs outweigh the revenue seems to sit just beyond the power allotment for most operations, at least until something radical happens with the power efficiency and costs of mining equipment.

Yes it does.

Indeed. The price of a bitcoin and the price of energy are the only two variables in the function that determines the energy consumption.

The depreciation of mining equipment is a third variable, and it is rather significant at this point in time.

As an example, the Antminer S19 Pro retails for $8-10k and is rated at 3250 W. Running it 24/7/365 would consume 113.88 MWh of electricity over a four-year lifespan. At average marginal hydro power rates ($0.05) this would cost $5,694 while at typical residential rates it would roughly double to $11,388.

The hashrate is an input into the difficulty calibration but not the energy consumed. Energy consumption is not an input that could be considered by the network because it's not observed globally (or even locally, really).

EDIT: the equilibrium determining whether and how much capital to invest in hashing equipment is not related to 'difficulty' -- it's a human economic factor that considers difficulty as an input.

> The hashrate is an input into the difficulty calibration but not the energy consumed

my understanding is that this is correct - others are introducing "cost of electricity" in an economics statement, not the math of difficulty itself

Any increase in power efficiency will lead to lower operating costs, which will lead to a higher hashrate, which will lead to an adjustment of the mining difficulty.

Energy consumption determines the cost, whereas the benefit is determined by the value of a bitcoin, and the demand for on chain transactions. Cost will always approach benefit.

They are also comparing against some older generation ASICs...the S19 from Bitmain[1] which 'mature' tech today gets 34J/TH, 95TH/s which is like... 40% better?? It's a super weird comparison point that they chose. Or maybe I am reading something wrong here...

[1] https://support.bitmain.com/hc/en-us/articles/900000253583-S...

The S19 Pro is down to 29 J/TH system efficiency, not far off beating the Intel design by 50%.

Indeed, it has a ring of "Hello, fellow youngsters" to it that Intel is even in this space. Maybe AMD and ARM has them thinking they're behind in other areas

I'm surprised that Logitech isn't already shilling Bitcoin Mining Heated Gaming Chairs and Ergonomic Crypto Mice and Mechanical Hashing Keyboards and Bluetooth NFT Whoopee Cushions (every time you sit on it, it mints a fresh NFT for each farting sound it emits, and automatically posts it to your twitter feed).

> Bluetooth NFT Whoopee Cushions (every time you sit on it, it mints a fresh NFT for each farting sound it emits, and automatically posts it to your twitter feed).

Do I still have to pay gas fees to mint a fake fart? How about I rip a real one and we call it even?

The Beats Whoopie Pill functions both as an bluetooth microphone input and speaker output device, with an ultra-low-frequency-high-power woofer. But it's a real pain in the ass to install and uninstall. Fortunately, Kyle Machulis has written an Emacs mode that supports it.

https://www.youtube.com/watch?v=D1sXuHnf_lo

https://imgur.com/gallery/4yEX49T

This product does not even need to be sold in order to boost interest in Intel stock. They can bring it up every quarter and keep the hype going.

The idea that more efficient mining chips will reduce Bitcoin's energy consumption is just wrong. Miners will just buy and use that new chip for as long as the value of the block reward exceeds the energy cost for mining it.

The mining network should always converge to an equilibrium where X$ worth of electricity is consumed for producing (a little more than) X$ worth of BTC, no matter how efficient mining chips are.

Personally I feel that Bitcoin miners are being unfairly targeted whereas the real issue is unclean energy production. Bitcoin is an easy target because a lot of people believe it is overvalued or worthless which gives rise to arguments such as "Bitcoin mining is unnecessarily wasteful".

> Nor does it seem like an amount that can make a dent in the energy consumption problem of Bitcoin mining.

Jevon's paradox indicates that they will make it much much much much worse.

Why sell it at all? They can just mine directly.

I believe this is the modus operandi of most manufacturers; selling new models in refurbished condition after a few months of earning revenue on the hardware. Former Swedish Bitcoin miner brand KnC accidentally admitted this, then backtracked and changed their stance to "extensive quality control".

Reply on news.ycombinator.com