Hacker News Reader

Your Bitcoin-mining heater will only approach 100% efficiency. Reverse-cycle air conditioners, while not feasible everywhere (my Australian one guarantees operation down to −10°C or −15°C, not completely clear which), are typically more like 300–400% efficient, possible because they’re moving, rather than generating, heat. So if we assume you need 35kW of heating for a given space, a reverse-cycle air conditioner may be able to consume only 10kW, while your Bitcoin heater will need to consume 35kW to generate as much heat. So it’s not free electricity even then.

You have to be careful thinking about what efficiency you are talking about.

At least while the hardware is new, you are getting paid to have your heat on with a mining rig.

Regardless of how efficient your heat pump is (that is Heat output divided by electrical input) you are always paying for the electricity to run it. Even if it has perfect thermodynamic efficiency, even if it is magical and has impossible perpetual-motion efficiency, you are still _paying_ to run it. In the magical scenario the cost goes to zero, but that is still not better than getting paid to run.

You can't compare thermo efficiency between something which generates money and heat and something which consumes money and generates heat until a crossover point when the hardware gets old and starts consuming more electricity than it can pay for. If you have a 400% efficient heat pump you would have to wait until your mining equipment generated 75% of the money it took to power, after that the heat pump becomes preferable.

Or you have to do money efficiency and talk about the capital cost and operating cost/profit of running each system

> the mining difficulty will keep on rising even when your hardware is turned off

This is not accurate. I mean, this is generally the case, yes. But technically speaking, difficulty only increases if there needs to be an adjustment in the block-generation time. If in the last two weeks, the average block has been generated faster than 10 minutes, then it means that miners have become faster (either because of better technology or because of more of them entering the market), so difficulty increases. But it could happen also that blocks get slower and difficulty goes down.

What I'm getting at is that constant raising of difficulty levels is not an artificially decided, it's a consequence of the market.

Indeed. As value of BTC goes up, number of miners goes up, hashrate goes up, difficulty goes up.

As value of BTC goes down, the opposite occurs. This happened following the last crash.

All of these affect one another but there isn’t clear and strict causal relationship or ordering.

Value follows difficulty but also difficulty follows value.

Value doesn't follow difficulty. That's the "labor theory of value" fallacy.

On the average, higher difficulty doesn't lower new Bitcoin supply, because difficulty adjusts to maintain steady supply.

Difficulty is affected by price of Bitcoin and cost of production. Difficulty can still rise if Bitcoin price development is steady or negative, due to newer mining hardware that is more efficient, or due to new sources of energy that are cheaper.

Value follows difficulty because difficulty is valuable in itself - it is a measure of security of bitcoin ledger.

Okay, I can see that argument, however I don't think it has much of an impact on price, certainly not to the point of "value follows difficulty".

You can arbitrarily decide what a "secured" transaction is by requiring a certain amount of prior blocks. There is no extra value in security beyond "secure for all intents and purposes".

I guess you could say higher difficulty equals fewer minimum prior blocks and therefore better performance for the system, but in practice the common minimum block requirement isn't going to fall below two and that's what many sites have been using even at far lower difficulty levels than today.

You’re right that number of blocks until transaction is considered safe is arbitrary but you can’t just arbitrarily raise this number to compensate for lower difficulty because interval between blocks is 10 minutes and so in effect you’re reducing usefulness (and value) of the whole transaction system.

High difficulty is certainly non trivial value proposition of bitcoin, among many others.

> You’re right that number of blocks until transaction is considered safe is arbitrary but you can’t just arbitrarily raise this number to compensate for lower difficulty because interval between blocks is 10 minutes and so in effect you’re reducing usefulness (and value) of the whole transaction system.

Yes, I admitted as much, but it's not a linear value relationship. There's a point where difficulty that is "too low" makes the system unacceptably vulnerable, but at that point the coin must be almost worthless and the incentive to be a malicious actor is equally low. There's another point where difficulty is high enough that for all practical purposes, a malicious actor has no chance of turning a profit. No further increase in difficulty beyond that point adds any value. Bitcoin is well past that point.

I wouldn't rule out that the belief that "more difficulty = more security" may cause certain people to go bullish and buy, therefore driving the price, but it's not such a direct relationship as price vs. cost of mining.

Given that, presumably, bitcoin mining rigs don't magically downgrade, how can the difficulty ever go down?

If his cost in electricity and connectivity exceed his revenue from mining, the miner will stop mining even if he has invested a lot of money in mining rigs.

But they won't likely sell the rig immediately. Which means that as miners halt their mining, if the difficulty would start falling again, it'll eventually make those miners resume their operations and halt the fall of difficulty.

If difficulty drops to the point where it's profitable again, they turn the rig on.

If, as a result, difficulty rises to the point of unprofitability again, they turn them off again.

This is basically equilibrium. It's not theoretical, it's what actually happened lately. Difficulty dropped along with price, now it's rising again along with price - but it's still below historical highs.

In fact they may never sell the rig, used mining rigs aren't worth much

ASICs might be worthless, but crypto miners using GPUs do rotate them, pawning them off as "lightly used" on auctions.

Presumably they fail at some point, and if the cost of electricity is greater than the profit they are turned off.

Yes! I keep seeing these projects (solar -> BTC, excess energy -> BTC) and I keep running the numbers and running into the same issue.

difficulty inflation

And I never understood why others don't see it too. Glad you found the same issue because I thought I was missing something.

Note: Solar has the same issue you pointed here which is that it only runs part of the time but the difficulty clock is still running.

Also: add in the halving of the reward! That happens in 1 year or so.

edit: updated reward halving estimate

>> buying it and only running it a fraction of the time is even moreso.

If you have an intermittent power supply, slap a battery on the system and turn it into a continuous power supply. Over a period of time the next amount of mining done will be the same. My point is whether you mine intermittently or continuously you will average out to the same amount of mining at the end of the day/month/year. The only issue is how you tune your battery/mining rig ratios to optimize the hardware investment.

There's nothing to optimize; the battery costs far more than the value of Bitcoin mined.

But what if you steal the batteries? Hell, steal the electricity, too! We're talking Bitcoin here!

The real money making opportunity is running the security firm that guards all those remote expensive rigs and batteries from being stolen.

At some point Bitcoin mining difficulty either:

1. must hit a ceiling and the rate of increase will significantly slow,

2. or Bitcoin mining will break SHA-256 and render it insecure.

I find 1 to be a the most likely outcome however the most plausible path for 2 would be a novel cryptanalytic attack on SHA-256 which could be exploited to increase mining power. It seems very unlikely that 2 would occur because mining power approaches the brute force limits of SHA-256.

No. There is no way for Bitcoin mining to "break SHA-256 and render it insecure". There isn't enough energy in the whole SUN to do that.

Well Bitcoin mining could "break SHA-256 and render it insecure" by:

1. exploiting cryptanalytic weaknesses in SHA-256,

2. successes in new computing paradigms such as a large scale quantum computer using Grover's algorithm.

3. or turning the local galactic cluster into computronium for SHA-256 miners. Although at that point we have bigger problems. =)

If you think SHA-256 will not be broken anytime soon then that suggests that you believe that mining difficulty increases will slow down.

> It seems very unlikely that 2 would occur because mining power approaches the brute force limits of SHA-256.

Can you elaborate on this a bit more? I'm not following what you mean by the "mining power approaches the brute force limits of SHA-256."

If Bitcoin mining power continues to increase exponentially, it will eventually get to the point that it will brute force the output space of the SHA-256. This won't actually happen because 2^256 is a very^256 big number.

someone tried a basic version of this https://www.engadget.com/2018/03/09/qarnot-qc1-cryptocurrenc...

If you look at ASIC-resistant currencies (like ETH), their difficulty does not change that much over time.