> You realize that the US government has many supercomputer clusters it owns directly to support its own data collection and analysis needs right? If the US decided BTC was a meaningful security threat, they have a lot of power they can bring to bear and anyone thinking otherwise is fooling themselves.
While the US government does have supercomputer clusters, mining-wise they're orders of magnitudes less performant per compute instance than a dedicated ASIC mining rig. Since the target's Bitcoin:
The supercomputer clusters are assumed to run top of the line gear, let's say 10 Tesla A100s. A recent Hashcat benchmark puts a Tesla A100's SHA256 performance at about 35 GH/s.
https://gist.github.com/Chick3nman/d65bcd5c137626c0fcb05078b...
The single compute sled can therefore be said to have 350 GH/s of potential Bitcoin-toppling performance.
On the other hand, we have the Bitmain Antminer S19 XP Hyd (255Th), where a single instance can brute force SHA256 at 255 TH/s, or 255,000 GH/s.
https://www.asicminervalue.com/miners/bitmain/antminer-s19-x...
It would take 700 10xTesla A100 servers to beat one S19. In order to beat a single instance of an S19, the government can do any of the following:
1) Directly beat them with more servers: Not feasible. To get that many compute servers would require putting a lot of orders through the whole computing industry, which would take years to fully roll out.
2) Directly beat them with ASICs: Feasible, but hard to keep a secret. Miners are sensitive to big entrants buying up everything, and it'll be made apparent in mining forums that something's up.
3) Make it illegal by banning PoW: That's currently happening right now with some states. What we're seeing now is a move to friendlier states, or a move out of the country, primarily the former in the public eye.
> This is the country that deployed a very sophisticated worldwide Trojan to make an attack on a nuclear reactor in a hostile country. The key here is if they can take over a huge swath of available worldwide compute no problem, they wouldn’t even need to touch their sufficient in house and commercial compute resources available to them.
It should be noted that the event in question specifically targeted Windows systems & the PLC systems attached to them, and it was only successful because of lax security procedures with regards to USB flash drives that allowed them to gain access into the air-gapped systems & run their payloads.
https://en.wikipedia.org/wiki/Stuxnet#:~:text=It%20is%20init...
The type of sophisticated attack being outlined here would at the very least involve the unpatched vulnerabilities of:
- Windows
- Linux (& all of its distros)
- Hypervisors
- IoT device firmwares
- Anything that AWS/Azure/Google Cloud uses
- Router firmwares
- Android
- iOS
- ALL of the OSes for the various XBoxes & PSXes
Each of which will require their own customized payloads, exploitable vulnerabilities, & custom compilation of the necessary software to perform the attack. These vulnerabilities also need to be persistent & non-traceable, which is increasingly becoming rarer in the long term (even though there may be bumps along the way sometimes). If no C&C systems are in place, then this widespread outbreak would require all of the systems mentioned above to run at a predefined timestamp. Otherwise, the traffic to/from the C&C systems would be noticeable by the public.
Even if all of that were to align perfectly, they'd still need to beat the ASIC miners mentioned above, and as shown by the RTX 4090's listed performance, it'd take 7000 Tesla A100s to beat ONE Bitmain Antminer S19 XP Hyd (255Th).
(It should be noted that A LOT of devices listed above are MUCH less performant than 1 Tesla A100 in terms of SHA256 hash computes.)
Simply put: It's not worth the time to create such a series of exploits when the resulting compute has a fraction of the compute power of an ASIC.
Consequentially, that means that the target of focus should be on ASIC miners themselves. However, there're several obstacles there as well:
1) Most miners (especially large miners) are mostly running the same Linux servers to control the ASIC miners. That means that the miners share the same exploit surface as AWS/Azure/Google Cloud/anyone using Linux, and benefit from the same security patches they're getting.
2) Limited surface area: ASIC miners are designed to have the bare minimum functionality to perform their sole job: compute SHA256 hashes. The consequential area of attack is limited & can be made impossible to attack with such restrictions on what the ASIC miner accepts.
3) The large miners have proper firewalls & networks in place to isolate each ASIC miner from each other via vLANs, making proliferation across the system harder to perform & easier to detect.
> Bitcoin proponents love to pretend it’s undefeatable and confuse IT in government with offensive capabilities.
See outlined offensive strategies above, & the realities of achieving such a target.
> They also liked to think it was anonymous until researchers put in a modicum of effort to show how it’s not.
The privacy issue was noted in Satoshi's whitepaper [0]:
[0] https://bitcoin.org/bitcoin.pdf
[0]> The necessity to announce all transactions publicly precludes this method, but privacy can still be maintained by breaking the flow of information in another place: by keeping public keys anonymous. The public can see that someone is sending an amount to someone else, but without information linking the transaction to anyone. This is similar to the level of information released by stock exchanges, where the time and size of individual trades, the "tape", is made public, but without telling who the parties were.
[0]> Some linking is still unavoidable with multi-input transactions, which necessarily reveal that their inputs were owned by the same owner. The risk is that if the owner of a key is revealed, linking could reveal other transactions that belonged to the same owner.
> They get the strategy of the attack fundamentally wrong. They think it’s to take over Bitcoin. It’s not. Driving value to zero is sufficient. And that is a much different case (making transactions take longer and longer, disrupting on-ramps and off-ramps, doing hostile take overs of data centers running mining operations, perhaps in conjunction with other governments, having spies infiltrate these groups, etc etc etc).
That case is definitely different, which is why it's separate from the attacks mentioned above. What's mentioned here is a fiat attack, wherein the cutoff of exchanges in/out of BTC is targeted, with varied results. The solution for this is the proliferation of BTC in place of USD for international settlements, which involves outreach into receptive governments that want an alternative system in case of problems with the existing western fiat regime.
> As the xkcd joke goes, your super duper cryptography is no match for a guy with a $5 wrench.
The axiom for this is that the person can be affected by a $5 wrench in the first place, which is not assured. It also assumes that the person being targeted has sole control over a given system, which is also not 100% assured.