I remember reading that it's hypothetically possible for Bitcoin to get caught in a trap if the price falls extremely precipitously. The difficulty only gets adjusted after a certain number of blocks, and if the profitability of mining drops fast enough, you could find a situation where it is impractical to mine enough blocks to trigger the difficulty adjustment.
My impression was that this was mostly just a problem in theory, and would not be triggered by the sort of decline in prices we're seeing in recent days. I assume the same is true for Ethereum?
Bitcoin difficulty adjusts every 2016 blocks (10 min blocks * 2016 = 2 weeks), but ethereum difficulty adjusts every block, so presumably ethereum wouldn't be affected by this.
Even if this did occur to Bitcoin, the community and miners would likely work together to put out a hardfork that adjusts difficulty downward to start the chain.
That is not how it works. The author of the transaction decides the transaction fee. If nobody has put a high transactionnfee in their transaction, you can't include it into the block you are mining.
Which means you wouldn't bother trying to mine a block. Seeing that no one is bothering to try to mine a block--and so their transactions are taking forever to complete--people would increase the fees they are putting on their transactions (which doesn't even require them to know what is going on, as that's what they would presumably try to do anyway) and eventually enough high-value transactions would exist to make it worth trying to mine a block with them included.
I'll just wait for you to take that short term loss and restart the mining when it's profitable again. Since that's what everyone thinks there's an incentive trap there. And now someone that actually does decide to take the loss has an incentive to make a 50%+ attack on the network to double spend their way into profitability.
Not a crypto fan but regarding the 51%+ attack that is sort of covered by game theory. By performing the double spend they doom the network itself so it will either be forked so it may continue or entirely die out.
Either way the party carrying out the attack is unlikely to be able to benefit from it, especially if a prerequisite of the viability of the attack is the network already being incredibly week.
All other miners drop out of the market because it's unprofitable. That leaves you as the only remaining Bitcoin miner.
Wouldn't this leave the network extremely vulnerable to 51% attacks though? If the recovery mode is that miners keep mining in the hopes of being the last man standing so that they can control the chain afterwards, then...
Nevermind, just answered my own question. Multiple miners are likely to keep mining in the hopes of being the last one standing and being able to 51% Bitcoin, which will keep mining competitive and get you to the block difficulty adjustment even if it's momentarily unprofitable.
Maybe I'm missing a key detail here. Let's say Alice, Bob and Carol (and no one else) are all mining bitcoin. For simplicity, one block is mined per day, the block reward is worth $300, and so the system is in equilibrium when they all spend $100 a day on electricity. Difficulty adjusts every 10 blocks.
Now the value of bitcoin drops precipitously - let's say they wake up one day and the block reward is worth $3. Let's say they all shut off their machines. Dave wakes up and sees no one is mining bitcoin, and thinks "I can be the last one standing". He now has to spend $300 * 10 = $3000 (in the worst case where the drop is right after the last adjustment) to mine ten blocks, for which he receives $30 worth of rewards.
But how does this help him? If the difficulty adjusts downward, Alice, Bob and Carol can all just start mining again. It's not like they're permanently locked out of the system because they turned off their mining rigs. So Dave has spent net $2970 for...what exactly? The ability to run a 51% attack and double spend to recoup his cost? To run the double-spend, he has to re-mine a second side chain at the way-too-high difficulty! I'm skeptical that this is an attractive proposition for Dave.
If you have a use for the heat generated (like heating a home or business) then running a miner means getting free heat, even if the cryptocoin generated does not pay for the electricity. Current cryptominers are generally too noisy for home heating though (although GPU's might be quieter).
100% efficient at the home, but not necessarily overall. Power generation itself and transmission to the home can be much less efficient (and much more expensive!) than, say, home natural gas heating.
Depends on the area and many other factors, e.g. how the electricity is generated. In much of the northern US/south eastern Canada, electric heating is much more expensive and less efficient overall.
To put it in perspective, heat pumps typically have an "efficiency" (coefficient of performance) of 250-350%, meaning they pump 2.5-3.5 times as much energy as they consume. So it's roughly 1/2 to 1/3 the cost to heat the same amount of space.
the problem is that as the outside gets colder, heat pumps get less efficient, so places that need the most heating benefit the least from heat pumps, and those are the places that use the most amount of energy for heat.
So yes, I agree, but heat pumps have cold weather limitations that mean they only benefit relatively warm climates.
True, but if the heat pump is designed to operate in very cold weather the minimum efficiency is 100% (backup resistance heating), so you will always at least break even versus pure resistance heating. Discounting equipment costs of course :)
Also, I've never seen a building (in the U.S.) that relies on electric heating if temps in the area will get much below 10°F (-12°C). At that point, it's always natgas/propane/fuel oil/wood.
> if the heat pump is designed to operate in very cold weather the minimum efficiency is 100% (backup resistance heating), so you will always at least break even versus pure resistance heating.
But the heat pump has a "outside" part, where there is a fan and other electrify using stuff. Can you not risk that more heat is generated outside than is moved inside?
Sadly I've seen many buildings (apartments and homes) that use electric baseboard heaters, and you probably have as well. They just aren't very modern.
That’s older heat pumps. Modern heat pumps are over 200% efficient at 0℉ and over 100% efficient down to around -15℉ Which makes them suitable for many places that don’t have relatively warm climates.
Only if you disregard how it‘s produced. Burning gas to generate electricity to heat a home is much less efficient than just burning the gas within that home, for example.
I like to keep my mining rig on a handtruck, and then I cart it from room to room to keep warm. The children so love gathering around it, warming their hands by the ASICs, as they listen to their father tell spooky stories about fiat currency and central banking and the time we had a real house with a working furnace and lost it because daddy leveraged it with an adjustable rate mortgage to buy a dumb ape avatar on Twitter.
This is exactly what I do in my residential home — winter heating — the only time I'm mining (Monero, typ) is when my heat pump can't keep up. I have my main miner eject heat onto my feet/workstation in lieu of an electric space heater.
From my own energy monitoring applications, I recon about 40% of my winter heating needs are met in this manner. I DO use a heat pump for heating between 40°F-60°F, but the efficiency of such a device rapidly diminishes as the temperature falls below freezing. Below 20°F, an average heat pump begins to be LESS efficient than simple resistive heating.
Current heat pumps can do quite a bit better than that.
Mitsubishi's residential air-source ductless mini-split heat pumps that have their "Hyper-heating inverter" technology, and similar systems from Fujitsu, are still about twice as efficient as electrical resistance heating down to around 0℉. They don’t start getting beat by resistance heating until around -15℉.
Their capacity starts dropping at around 23℉ and would be around 75% at -15℉ so if you need to go below 23℉ you do have to oversize your system.
I'm hoping the crypto crash lasts for about six months or so and a lot of these miners who were taking my precious gaming cards start to struggle and have to sell all their old 3080s / 3090s / 6800 XTs / 6900 XTs for around half of what they paid. If I can pick up a 6900 XT for $499 or so, especially used from a miner who's had it undervolted and likely running at 80% or less all day, I'll bite.
I want the pain to really hurt. They deserve it. I would blame NVIDIA and AMD for not ramping up production, but the sad fact is, there's only so many foundries capable of producing GPU chips and those are TSMC and Samsung, so it's really a matter of not having enough capacity.
You probably don’t want to buy a used mining card. There were reports about drastic performance decreases after a certain time of mining with them. I don’t find the link any more, but it was on HN as well.
>I'm hoping the crypto crash lasts for about six months or so and a lot of these miners who were taking my precious gaming cards start to struggle ...I want the pain to really hurt. They deserve it.
I quit gaming to focus on disinformation because when I went off to college, folks from that community went so far as to walk uninivted into a house party thrown by an ex-girlfriend to repeat dead memes then act surprised when the response was nearly kinetic.
(I was starting to get involved with the local green movement instead of Pokemon Green, they were forever angry sitting alone at home for hours never led to a girlfriend. Always whining about finding a girlfriend, never a partner, the rude little "incel"s)
Anyways, I agree that the Bitcoin boys deserve to be pain, and they should pray to whatever god they wish that it's only psychic after being so harmful to the planet to try to route around consent (and AML laws - 10k is 10k dumbasses).
The hard truth is that Lyft, Airbnb, Bitcoin, Grubhub, a lot of it is reinventing the cyberpunk lifestyle you could have lived in the 90s or could TODAY with a little cash and respect for the fact that people are not programs and do not have the memory of goldfish.
> If I can pick up a 6900 XT for $499 or so, especially used from a miner who's had it undervolted and likely running at 80% or less all day, I'll bite.
It's sad that places like Ebay are useless due to so many scams, and Craigslist suffers the same issues if you don't "keep it local" -- at least when I used to do deals on CL, I had the option of just taking a Louisville slugger to someone if they tried to walk away with both the currency and the merchandise.
(Or getting in my car and driving off if they tried to negotiate too hard over ten bux.)
Anyways thanks for raising this idea of undervolting -- I had a Game Gear as a kid so I'm forever on the lookout for ways to avoid my battery going dead.
(It took six AAs and would die just as you got close to Lake Erie, the thing fucking sucked power worse than a Dogecoin farm)
>I would blame NVIDIA and AMD for not ramping up production, but the sad fact is, there's only so many foundries capable of producing GPU chips and those are TSMC and Samsung, so it's really a matter of not having enough capacity.
My worry is they'll build foundries across SE Asia, and then folks like Apple will coerce the smaller countries in parallel more easily, now that China has built up a middle class that, much like the times leading up to the Tienanmen protests, is once again ready to rise up against their totalitarian oppressors.
Take a look at the "big board":
Taiwan is staring down the barrel of the PRC, and South Korea is forever under the threat of annihilation since Kim Jong Dumbass could launch a massive kinetic strike at any moment, and major powers like China have stated they have a no first strike nuclear policy.
(On my end, I'm just one guy, but I also am feeling hamstrung by a decade plus long promise to only relatiate, never intrude, in the "cyber" realm... but I'm just one guy, taking a break from making purposefully inflamatory GIF laden conference slides since apparently diplomacy gets you nowhere :-) )
Even if you only intend to do ML stuff with it, the price-to-performance ratio of Nvidia's next-gen GPUs are going to be insane. Just the jump from 8nm to 5nm alone will be pretty crazy.
nobody knows what the prices will be, so until then I'm not sure we can declare price to performance anything.
As it stands, the previous generation shows that prices can become arbitrarily high if supply is low. Two years and a crypto crash later the cards are getting close to their MSRP from two years ago.
GPUs seem like they're the new .22LR in terms of the MSRP being meaningless if it's never on the self.
It always confused me how many people stocked up on the latter, since if you want to actually have a resistance like the founding fathers intended, not just have a series of manchild meltdowns at the idea of a "well regulated" militia, ending with getting shot by the senate sergeant at arms, what I was told you do is run up on them like they're Reinhard Heydrich and use at least a .22 magnum, with hollowpoints, so they enter the skull, split open wide, then bounce around, rather than .223 rifle round or some double plus P bullshit that will go in one side of the head, out the other, and through a wall into the head of someone who's not a totalitarian scumbag.
But I was very drunk and had my phone off when I had THAT conversation. I've quit drinking alcohol with strangers since. Zdraví! :-)
Anyways, to swing it back to gaming or whatever:
>As it stands, the previous generation shows that prices can become arbitrarily high if supply is low
The issue is the supply will always be low if the MSRP is too low and/or folks don't enforce the law when people buy up consumer goods for illegal enterprises to the detriment of the hobbyists and other peaceful civilians the MSRP is meant to be set for.
>The issue is the supply will always be low if the MSRP is too low
The issue is can people make the supply low and drive the cards price up where such activity is profitable, or not. You can declare the MSRP 'too low' but if there are plenty of cards available at MSRP no one will pay the inflated price.
The supply was very low, which allowed people to drive the price up for opportunities to engage in some profitable arbitrage with those few who were interested in overpaying for the cards. After a short time of ubiquitous scalping, the manufacturers themselves raised the price to eat the amount being arbitrated because why wouldn't they if there were still people buying at that price? At that point why would they make more cards if they can make money selling fewer cards with less cost?
The 'market finds the price' approach doesn't work if everyone involved including the manufacturers are all looking to drive the price up for their own profit, which is what happened. Lots of people made money as middlemen for running bots etc to snap up whatever supply of cards there were, without adding any value or improving anything but their own finances.
Now translate that to everything else you buy and you can see that if the same could be done to baby formula, or cars, or houses, or eggs, and you can clearly see it would be a negative on society in general. You just need someone with a big enough fund - like say blackrock buying up homes, etc.
>The 'market finds the price' approach doesn't work if everyone involved including the manufacturers are all looking to drive the price up for their own profit, which is what happened.
Agreed. Folks want to focus on one variable (the MSRP) -- with the idea "the market" will adjust that variable, but "the market" is not a healthy one -- if people are buying a bunch of GPUs to create an electronic currency that is less private than cash, then using it for a host of illegal activity you can do... with cash... and in the process distorting the market for scientific equipment (as the one parent hinted at with their comment about "just" doing machine learning) then only changing one variable (the MSRP) isn't the appropriate solution.
But you get people who would scoff at the idea of an invisible man in the sky having near religious meltdowns if you tell them there's also no invisible hand in the market :-)
How does the recent decision by the ETH devs to delay a significant change to the difficulty algorithm factory in? This 'difficulty bomb' would have made mining significantly less profitable.
>The Difficulty Bomb was originally put in place for a variety of reasons:
>- To act as a deterrant for miners who wish to continue mining the Ethereum 1.0 (Proof of Work) chain once the network moved fully to the Ethereum 2.0 (Proof of Stake) chain
>- To ensure that the core developers of Ethereum 1.0 are "forced" to upgrade the network via a hard fork (if only to delay the effects of the bomb) so that the network does not stagnate innovation-wise
