Once they successfully do the attack, they can take over the blockchains and allocate all cryptocurrencies to themselves, cash out, and it'd probably register as a good profit on their income statements.
Well, assuming any kind of rationality in the crypto market is always hazardous: meme coins (doge coin, shiba inu) are worth billions in market cap, so do hard forks of bitcoin, and Ethereum didn't not collapse after they decided to rollback a bunch of things they didn't like during the DAO “hack” while they are advertising Ethereum as immutable and tampering resistant.
You can never be sure how this market will react in any event (well, except Elon tweets I guess)
I used to assume that, but there are coins which have been attacked successfully before and people still use them anyway. I don't attempt to explain the logic there :P
1: https://medium.datadriveninvestor.com/itcoin-unhackable-it-h...
Usually it's just about the currency becoming worthless (usually only a problem in government controlled currencies that have loans in foreign currencies), or people just not accepting it anymore.
There's no inherent value in any currency, it's just a useful fiction because it's really hard to go get five goats to trade somebody for a laptop.
Given the number of countries using the dollar as their reserve currency, value could collapse quite a ways without the US doing anything.
It's not likely, but neither is a 50 + 1 attack on bitcoin given the amount of computing power currently tied up there.
Also, Bitcoin has significantly more vulnerabilities than just a 51% attack or a crash as nodes need to be connected to the internet and therefore can at least in theory be hacked. What exactly happens after that point is anyones guess, but it could easily be worse than a traditional 51% attack.
Sure, but the value of the dollar (like everything else) is in how easy it is to get (supply). The US GDP is not a measure of the supply of dollars on the market, it is a measure of how often those dollars are exchanged.
> Bitcoin has significantly more vulnerabilities than just a 51% attack or a crash as nodes need to be connected to the internet and therefore can at least in theory be hacked.
That's fair, if we're including implementation vulnerabilities, it does expand the potential problems quite a bit. The post I was responding to was specifically about a 50 + 1 attack. Though if we include these, we should probably also expand physical currencies' problems to include counterfeiting in that case.
The US GDP directly translates into the amount of taxes paid. Suppose a 50% drop in the value of USD happens. If nothing else happens taxes increase by 100%, and people need to acquire more USD which drives the value of the currency up.
You might wonder about economic hardship, but historically that causes deflation. +/- 2% per year without printing money is easy, but there is a reason you don’t see huge swings without the government printing money.
> counterfeiting
An issue for any physical currency but electronic currency actually operates independently. Trying to counterfeit large quantities of cash is therefore surprisingly difficult even for nation stale level actors. A great fake bill can get passed around, 10 billion in fake bills which is still a trivial amount would trip up various monitoring systems.
? GDP is produced and denominated in USD, domestic taxes are paid in USD. How does a drop in the relative value of USD compared to a foreign currency influence domestically paid taxes in any way?
Oh definitely. I did not mean to imply that taxation in USD does not have a modulating effect (and if I came across that way, I apologize). I'm just trying to say that you can have some pretty wide swings even with that.
> but there is a reason you don’t see huge swings without the government printing money.
You don't generally see it, no. Mostly because other countries use the dollar as a store of wealth, and they don't want to tank the value that they've squirreled away. I'm bringing it up as a hypothetical with the same likelihood as a 50 + 1 attack on bitcoin.
China, for example, seems to hold ~2 trillion in USD (~60% of 3.5% trillion total reserves).[0] I assume most of this is held as digital assets, as there's only ~2 trillion physical dollars in circulation.[1] Dumping that much on the market would cause some pretty wacky inflation (as well as destroy China's stability). In a similar vein a 50 + 1 attack on bitcoin would allow an actor to give themselves all of the bitcoins, but would also destroy the value of those bitcoins because nobody is going to trust the network after that.
Both of these attacks are really only useful if you value hurting other actors more than you value your own resources/your own sense of self-preservation.
[0]: https://en.wikipedia.org/wiki/Foreign-exchange_reserves_of_C... [1]: https://www.uscurrency.gov/life-cycle/data/circulation
"Clue" and "holding coin" have not so much overlap.
And there's plenty of reasons to want to do so: Proof-of-work currencies are what make ransomware payments possible, and there's a large daily CO2 cost for proof of work currencies. The question is simply when is the cost of the externalities greater than the cost of acting?
[0]: https://www.crowdstrike.com/cybersecurity-101/ransomware/his...
So short term: having the high prices paid by crypto people is nice. But it destroys their relationship to their old customer base. Depending on where the crypto market will be in 5, 10, 15 years, this can be the 5x of their market cap or the 0.2x
ETH mining at 500TH (a solid chunk over 51% just for ease) would net you 300 ETH an hour...that's 1.1 million USD, which means you would have to be able to purchase millions of dollars worth of real world items that would be released to you within the time-frame of your attack (in this case an hour) to make this even approaching "worth it" from a risk/reward viewpoint and even that calculation is ignoring just how profitable they'd be just mining normally for the foreseeable future.
No one is waiting for just an hour worth of confirmations for millions of dollars of real world goods exchanged for ETH.
No, not for Bitcoin. GPUs are no good at mining Bitcoin; you need an ASIC for that.
> Once they successfully do the attack, they can take over the blockchains and allocate all cryptocurrencies to themselves […]
No. Having a majority of mining power does not enable anyone to assign currency to themselves, except for the mining reward. That would require faking digital signatures.
Since Bitmain asics are currently being made at around 16nm nodes, using 3nm process that Nvidia has access to - think about that for a bit.
If I remember correctly, the difference in transistor density improves as 1/alpha whereas area of transistors, if everything gets shrunk with alpha decrease, should decrease as square of 1/alpha. I.e., if alpha differs two times, there should be 4 times more shrunk transistors whereas practice shows only 2 (and often less) more shrunk transistors.
This is because you cannot make transistors too small in every dimension, they will leak.
So the difference between 3nm and 16nm will be less than five times. Not 25 times as difference between these parameters would suggest.
The difference in getting masks and first prototype should be quite substantial in terms of upfront payments and time-to-production.
I remember that 25 square millimeters proof of concept in 180nm process needed $50K of money and half an year of delay. The 35nm (5 times less) prototype of the same area would cost about $500K and more than half an year delay. The difference between 16nm and 3nm is about the same, I can expect several millions of dollars of upfront cost of small chip and year long delay before prototype arrives for 3nm process.
These Bitmain guys should be assumed to be not stupid and they most probably have access to latest processes after second prototype. Yet they choose 16nm - we must ask ourselves "why?" and Nvidia should too.
As a matter of fact, Nvidia had its share of problems with tiny processes in past.
"A GPU attack on BTC" Jésus help us
See, if they just stopped selling them, people would ask questions. If, however, you just reduced the supply by like 80%, said it "was due to all those crypto miners hoarding them", and then proceeded to use that hoarded 80% to do a 50%+1 attack, you'd be able to -quietly- take over, and turn around and sell those coins to the still preserved market.
applies tinfoil hat In fact, who is to say that isn't exactly what they're doing, and hence the existing shortage?
So they're going to manage to not sell any cards (without disclosing that as a public company), set up the infrastructure and do a 51% attack on bitcoin in front of everyone, then "cash out" as if there would be people lined up waiting to buy bitcoin when this has just happened in front of their eyes?
Unlikely.
Nvidia wouldn't do this because it might hurt GPU sales.
I imagine that if Nvidia were to openly "steal" someone's bitcoin using a 51% attack then the victim could sue Nvidia in a US court and things could get interesting.
What laws would apply and what would the outcome of such a case be? And could the act also be considered criminal?
What if you are a drug dealer or dictator?
Why doesn't Nvidia use the cards themselves to mine? Not for the attack but for the profit?
Each card is obviously cheaper to produce than they are selling it for. Why not maximize all the profit?
Assuming this was even possible, your plan has a fatal flaw.
Who the heck is going to buy the BTC back?
I go back in time with a super computer and mine all the coins the first year and lock them up in my wallet. I then return to the present day.
Are they worth anything?
I have mined and sold $2,800 of ETH in the 9 months I’ve owned it. The extra power has cost me a couple hundred dollars.
I’ve made anywhere from $3 to $75 per day during that time (thank you Shiba Inu launch date for the highest numbers).
I’m continuing to make about $3 per day. You can look up current profitability for any graphics card at whattomine.com.
Bottom line is I could have added infinite graphics cards to this setup and the profit would’ve gone up linearly.
Anyone that doesn’t think GPUs are high due to ETH mining has no idea what they’re talking about. The analysis from OP seemed ridiculous to me. I didn’t even see LHR mentioned in the article which are GPU manufacturers attempting to make GPUs unprofitable for miners so they can only be used for gaming. (Hint: didn’t work very well).
That’s at least six months off and has already been delayed repeatedly.
The 3080ti’s are easier to find as they’re LHR and price per hashrate ratio isn’t good. What I mean by easier is that bots have a tendency of going after cards with price/ratio.
It doesn’t help either that DDR5 RAM is also highly scalped. PC memory is in direct competition to GPU memory and vice versa.
For real though, cut that out.