It took me understanding every detail of Bitcoin, including the UTXO, the OP codes, how it enables layer-twos, how and why it is pseudonymous, and the BIP process, to realize all other cryptos are scams and trash. People believe the stablecoin or CBDC is a threat to Bitcoin. They are wrong because Bitcoin exists precisely to cure the world of inflation, which is only enabled by central banking itself.

Replies (67)

Don't get me started on Monero. If you want privacy, use lightning with unadvertised channels over tor, and run an internal dedicated bitcoin node also only over tor. Connect to your lightning node over a personal VPN to your internal network or over tor. If you want to sanitize/de-link your bitcoin, run it through a mixer, trade it over robosats, or just put it onchain through Boltz or liquid. I've gone through the privacy-coin phase and realize, bitcoin is the only way.
I watched @Super Testnet give a speech about it. He thinks the lightning network is more private than Monero. I see all alt coins as inflation attempts on the bitcoin network. I don't want any inflation aside from a little paper pocket cash because I think that is also more anonymous than monero.
DNMs won't use lightning and exchanges don't ban lightning. DNMs only use monero and exchanges ban monero. repeating whay super testnet says will not magically give you perfect understanding of what's going on and why. he just wants to tell you what you already want to hear.
A professional user can (with FCMP) spend Monero more privately than cash. Professional means. Runs his own OS. Uses anonymity networks. Cash has serial numbers and gets tracked at many places already. It's fungibility is by decree. Untraceanility is breaking down for years, while Monero improves.
Lightning only has similar sender privacy guarantees IF AND ONLY IF you're running your own LND node with a channel opened to an onion peer with a UTXO that's untraceable to you. Using a custodial wallet like Wallet of Satoshi? No privacy. Phoenix or Zeus with Olympus? Little privacy from the LSP, the metadata makes timing attacks trivial. Only exception being ecash but it's custodial and anonymous nature makes it extremely susceptible to counterparty risk at the mint. "Lightning has awesome privacy" is true only in the perfect case. For most they have worse privacy than even the noob Monero user that uses a remote node with no VPN with coins purchased from a CEX. Let that sink in. The worst case scenario privacy wise for Bitcoin, is superior to all but the most schizo levels of Lightning Bitcoin usage when using Monero. That is why Monero is used on darknet markets and lightning isn't and on-chain Bitcoin is all but ceased. Monero is private even for the noob, Bitcoin is public for all but sysadmin Linux expert types.
I don't buy your assessment that an untraceable UTXO is *required* on your ln channels for complete anonymity, though it certainly does improve it. Omce you open a channel, the only public information available is who you opened a channel with, not who you transacted with. One does not need to be a sysadmin to run a private ln node, there are packaged solutions available. Agreed the tools can be improved to increase privacy in a straightforward way.
I'm talking about getting to the level of privacy that Monero provides by default. Yes having a nonKYC UTXO is vital for having no link of your channel activity to your identity from you channel peer. You need to look into timing analysis and BGP vulnerabilities in the Tor protocol. Lightning hops are vulnerable to the same attack vectors and the more metadata you give, the more trivial the deanonymization.
Yes, I understand these vulnerabilities of tor and the lightning network. The meaning of "privacy" is a matter of degree, not a binary quantity.
I mean so you did the meme right? run your own LN infrastructure with CJed UTXOs, rather than "just use monero" for strong default privacy? there isn't any foundational reason Bitcoiners should have a problem with Monero. except theyre attached to their ideas about what Bitcoins success looks like.
Super Testnet's avatar
Super Testnet 2 months ago
STN makes good points about LNs privacy and good points about Moneros weaknesses. but he completely REFUSES to ever acknowledge the UX problems in achieving *comparable privacy guarantees* with LN. this is just the latest version of "LN will fix bitcoin privacy" that we've been hearing for 7 years now. it's not that it's not possible, its that its hard and if your security guarentees were broken you would never know because you dont have L1 consensus-level security.
These are NOT dnms and he knows it. These onion services have completely different threat assessments than a DNM. and that they use LN cannot be considered a stress test of the network under adversarial conditions.
also in addition to the UX hurdles on Bitcoin or LN
Hanshan's avatar Hanshan
The thing is though Until something has been battle tested and has been around for five to ten years *in an adversarial environment,* we don't really know if it's a good privacy tool or not. Bitcoin was the same way, it was "private" for a while. But it was private because nobody was looking. Lightning is in the same condition right now. It looks good on paper and I'm hopeful. But It's difficult to manage and we honestly don't know at this point if it's going to work out long-term or not. I certainly don't consider it a drop-in replacement for known and tested solutions for transactional privacy. So I coinjoin, move through LN and use Monero as much as possible.
View quoted note →
Super Testnet's avatar
Super Testnet 2 months ago
> he complete REFUSES to ever acknowledge the UX problems I acknowledge that UX problems exist in LN. I deny that similar problems are absent from monero. Per the monero website, here are some UX problems with monero: - Users shouldn't reuse a monero *address* - Users shouldn't even reuse a monero *wallet* (i.e. you should delete your wallet after each use and make an entirely new *wallet* for the next use, according to the monero website) - Users should run a monero node, not a phone wallet The website doesn't say the following thing, but users should also use tor/i2p/vpns if they use a monero *node,* but it's needlessly difficult to do so because many important people in the project basically disagree with that. The monero github even says in its readme "Monero isn't made to integrate with Tor" and discusses the hoops users need to jump through to get it to work with tor. It's got serious UX problems and most monero influencers just don't want to admit that because the phone wallets are easy -- and the monero website recommends not to use those!
So much bullshit, so little time... lets start here, mind showing me where on the monero website it says "users should delete and make an entirely new wallet for each use"?
Super Testnet's avatar
Super Testnet 2 months ago
Evidence against reusing a monero wallet: “If you need perfect unlinkability of your receivables, the only solution remains to use a separate seed (separate Monero wallet).” source: Evidence against reusing a monero address: “To prevent the payer from linking your payouts together [you must] generate a new subaddress for each payout.” source: Evidence for running a monero node, not a phone wallet: “It's always advisable, especially for privacy-conscious users, to use a personal node when transacting on the network to achieve the highest rate of privacy.” source: https://www.getmonero.org/get-started/faq/ Evidence of difficulty of using monero on tor: image
Super Testnet's avatar
Super Testnet 2 months ago
> These are NOT dnms and he knows it They are per the definition used by law enforcement and by wikipedia. If you've got a strange stricter definition of "it only counts if drugs are widely sold there," that's a you-problem > These onion services have completely different threat assessments than a DNM Nope. They have very similar threat assessments, for the following reason:
Super Testnet's avatar
Super Testnet 2 months ago
> You need to look into timing analysis and BGP vulnerabilities in the Tor protocol While looking at that, also look at how those vulnerabilities apply to monero's Dandelion protocol: Key quote: "our analysis of Dandelion and Dandelion++ indicates that they do not offer high anonymity either…an adversary that controls 20% of the nodes…[can] intercept [enough] transactions [to where] the median entropy is about five bits…equivalent to 32 possible originators per transaction."
I don't understand why it's worth the effort to make a big deal out of the address reuse mistake. does lightning not have bolt12 and lnurl? if you reuse these, the sender will infer that it's you both times.
Super Testnet's avatar
Super Testnet 2 months ago
> I don't understand why it's worth the effort to make a big deal out of the address reuse mistake Because address reuse is bad for privacy (the monero website recommends against it) and yet most monero wallets are designed to encourage it > does lightning not have bolt12 and lnurl? if you reuse these, the sender will infer that it's you both times Yes, that is equally bad for user privacy
I agree, address reuse is bad for privacy. but I have used multiple monero wallets and they tend to show a fresh, unused subaddress by default, just like most bitcoin wallets. you can also tell which specific subaddresses got paid, and by how much. if someone guesses that you own different subaddresses and decides to pay you on the "wrong" one, you can see them doing that. it's not very exciting to talk about opsec mistakes that are possible on both monero and lightning, or which can be easily mitigated. thry both have reusable addresses and neither of them forces you to reuse it. additionally, tor is already baked into monero's electrum fork and it's a preview feature in cake wallet. some wallet developers have started looking at i2p as well. I am unhappy that it's necessary but I wouldn't exaggerate how difficult it is to integrate. image
So this is the point in the conversation where STN starts talking about problems with blockchain privacy, but somehow makes it all about Monero. i guess because the monero website talks about these problems? and yea, transacting on an L2 solves some of those problems. LN remains an beta-tested and largely ignored payment system and you don't know how it's going to play out. for example. in actually practice, since most users either have one big channel with an LSP or are using phoenix or muun or something, you don't have unlinkabiity of txs on LN EITHER. so GTFO with that bullshit
And since I don't want to do a real long one of these and go through and count all the disingenuous bullshit, just one more thank goodness everybody now knows that you can avoid the trials and tribulations of passing those flags to your Monero node, by just running a full stack lightning node.
You can stack dandelion++ with Tor or i2p. Plus Monero has both on chain AND network level privacy. If dandelion++ is broken all that is proven is that your IP did A transaction. Not what the transaction consisted of since reciever and amounts are still blinded. With lightning, the privacy is entirely reliant on the obfuscation of the chain of hops. If origin and desination are linked the whole transaction is revealed and linked to you.
Super Testnet's avatar
Super Testnet 2 months ago
I linked to the Tor domain Thanks to clearnet proxies, any dnm can be used on clearnet They are DNMs if they are a marketplace accessible via Tor or similar, there is no requirement that they somehow eliminate all clearnet versions too
Super Testnet's avatar
Super Testnet 2 months ago
Do those monwro wallets also show a contact list? Such features encourage address reuse, thus creating additional UX hurdles -- if you recommend a "user-friendly" monero wallet, you're probably recommending a privacy-harmful monero wallet
STN is also the guy who argued for a week that "monero is 'traceable' because the sender knows where the payment went"
Super Testnet's avatar
Super Testnet 2 months ago
To get decent privacy on monero, you have to install and configure a monero node Might as well get superior privacy by installing and configuring an LN node instead
Super Testnet's avatar
Super Testnet 2 months ago
> STN starts talking about problems with bbllpckchain privacy, but somehow makes it all about Monero These problems apply to other blockchain-based privacy systems too The fact that other terrible systems also have these problems is no excuse for monero
nobody who has even done both of these thinks they are equivalent , and you look dumb suggesting they are similar.
leave it. transaction unlinkabiity is better using monero subaddresses with the monero gui wallet or whatever than with Phoenix or muun or any sovereign node with just a channel or two and that's #facts folks
monerod --proxy 127.0.0.1:9050 or add proxy=127.0.0.1:9050 to the monerod.conf file then its running behind Tor. Is that difficult? What they mean is that Monero nodes are not made to accept INCOMING connections via Tor. You can proxy transactions via a hidden service but that's it. Making a new wallet per transaction is schizo IMO but your criticisms about having to make new addresses and the like are all REQUIREMENTS for Bitcoin on-chain even to the same vendor. You can reuse the same address with the same person with Monero's and there will be no on-chain link still since each address is a stealth address similar to BIP 352 addresses. Running a Monero node is best practice but even using a remote node offers superior privacy than most on-chain and lightning Bitcoin setups. Its evident to me that you've no experience with Monero usage as you'd know everything I'm describing here by using the official Monero GUI wallet or the like. Please do proper journalism by using the protocol you criticize. I've extensively used and researched lightning at this point thus giving me true insight to the benefits and limitations.
Super Testnet's avatar
Super Testnet 2 months ago
Then how come I can trace my xmr transactions to you but you can't trace your xmr transactions to me? I'll even use Phoenix if it helps you
yeah, and these super dangerous wallets also have a way to export your seed phrase. this can be privacy harmful if you accidentally write it on your forehead they have a way to view your transaction history and some of them let you export it to CSV. this is incredibly dangerous for privacy in case you make a mistake and fax it to the police station
are you drunk? that isn't what we are talking about. YOU'RE saying because there's a possible *active attack* on a monero user that could be executed by a sophisticated adversary that could result in them linking two subaddresses together (OMG), users should always send from a new wallet by default. (which is bullshit) and as if most LN users transactions aren't trivially linked together by their custodian or the one or two LSPs they have a channel with... now you want to talk about knowing where the money goes? well ok I guess but you haven't said anything about your first FUD yet.
I found a critical vulnerability that affects all monero wallets. they display the transaction amount on the user's computer screen. this will leak information about how much cocaine the user buys if the chupacabra sneaks up behind them and looks at the screen. I disclosed the vulnerability to the featherwallet developers and they have been ignoring it for 8 months! none of the lightning wallets are affected by this vulnerability because you cannot use lightning to buy cocaine anywhere.
it would be a lot easier to have a discussion about it if certain parties didnt disingenuously insist that obviously dissimilar things were comparable it requires A LOT more explanation to make it clear to the non-technical person why for example, its harder to link spends in monero than for most LN users BuT iT TAlkS abOuT uSIng a dIFFeRenT WAlleT oN gETmOnEro dot ORg
Super Testnet's avatar
Super Testnet 2 months ago
> Is that difficult? Yes, for most people. Manipulating configuration files requires a lot of background knowledge that most people don't have. Using command line arguments too. You also neglected the part about installing linux in the first place, and then installing tor. For most people, I suspect they would assume "installing tor" means "installing the tor browser," not running "pkg install tor" or "sudo apt install tor" or whatever it is since it depends on which linux operating system you're using. And if someone gets past all that, then you have to somehow find out that (1) you have to keep it running (2) the port switches from 9050 to 9150 if you're using tor browser -- and if you're on Windows, everything is different there too. So yeah, it's difficult. Not for me or you, but we've got a lot of background knowledge that you can't assume for most people. > Making a new wallet per transaction is schizo IMO Yet -- according to the monero website -- that's the only way to get transaction unlinkability in monero. You think *lightning* has a UX problem? Meet monero. > your criticisms about having to make new addresses and the like are all REQUIREMENTS for bitcoin on-chain even to the same vendor. You can reuse the same address with the same person with Monero's and there will be no on-chain link There will be no "on-chain" link, but that person will have an off-chain link that is just as provable. An identical thing is true for monero's addresses and bitcoin's xpubs: if you trust the sender with your privacy, you can give him a monero address or a bitcoin xpub (or a bitcoin SP address, more recently) and if he reuses it, then "on-chain" there will be no link between different transactions sent by that person. Unless the sender makes mistakes too. But even if he does everything *else* perfectly, he made this mistake: he kept your address or xpub and reused it -- and that means he kept a record or proof that links multiple transactions together. If the address or xpub leaks to untrustworthy people, e.g. if he is subpoena'd for that information, your adversaries learn those links and can prove them in court. I think that's a terrible for privacy. Do you? Or is it only bad when it applies to bitcoin, and perfectly fine in the case of monero? > Please do proper journalism by using the protocol you describe I've used monero plenty and I think it sucks.
Super Testnet's avatar
Super Testnet 2 months ago
> which is bullshit If "create an entirely new wallet for every tx" is bad advice then maybe submit a PR to withdraw that advice from the monero project's website > most LN users transactions [are] trivially linked together by their custodian or the one or two LSPs they have a channel with I acknowledge that most LN users are not using it in a privacy preserving way. Do you acknowledge that most monero users aren't?
Alright I got major counterarguments to your points here but can't outline them in detail 'til after work so I'll be back. In summary: adding Tor to your Monero node is optional because of dandelion++ but is absolutely required with Bitcoind or the like because it doesn't have a dandelion implementation to fluff transactions. To add the Tor proxy its just as trivial for both daemons thus isn't any harder for Monero. The official Monero GUI has a check for socks proxy too so you can't get much simpler than that. Bitcoin core has the same but once again, it's required for Bitcoin due to no dandelion stem. For your big point regarding address reuse even the most ideal cryptographic solution for privacy will always be compromised if an attacker, like a state agency who subpoenas, gets the private key. Not a gotcha, every single cryptographic system has this vulnerability: PGP, Tor, SimpleX, Signal you name it. If its decrypted on the device, then the adversary has the plaintext if they have access to the device or private keys, simple as. Monero is way simpler than setting up LND. I can tell you've not used Monero because you refuse to concede this point. I've set up a public monero node via systemd, set up Retoswap, set up p2pool recently with xmrig, the latter being the most difficult. Each of these way simpler than learning how lightning channel liquidity works and using the embedded LND node on Zeus. This difficulty went through the roof when setting up LND via CLI. Tor is often preinstalled on most Linux distros and often is enabled via systemd. Port doesn't switch from 9050 to 9150 that's why they're separate. The Tor browser creates a whole separate Tor process than the daemon version that uses 9050 by default. Almost all distros use systemd so once For is enabled it will restart itself if it crashes or when OS is restarted. Its set and forget. This error of yours betrays your inexperience in Tor routing. Everything's always more difficult on Windows. Not a Monero gotcha, Bitcoin has the very same UX hurdles (worse since you absolutely need Tor to protect your IP when broadcasting transactions from bitcoind due to lack of dandelion++) Cake wallet mobile + random remote node gets you similar or easier UX and way better privacy than Phoenix, Zeus LSP, Spark WoS or other custodial solutions. Cake Wallet plus your own node, or MoneroGUI with a local node is way easier than running your own LND behind Tor while offering superior network level privacy AND better on-chain privacy with greater security due to being on chain. You don't need to worry about watchtowers, force closes, etc.
Yena's avatar
Yena 2 months ago
Lots of more privacy problems listed on the OFFICIAL getmonero website. Lightning does not have these issues. Admits monero users will be POOR: "If you use Monero but give your name and address to another party, the other party will not magically forget your name and address." "If you give out your secret keys, others will know what you've done. If you get compromised, others will be able to keylog you." "If you backup your seed in the cloud, you'll be poorer soon" read it for yourself! source: https://www.getmonero.org/get-started/faq/#anchor-magic
Yena's avatar
Yena 2 months ago
So when Facebook, reddit, Pornhub, etc had onion domains were those DNMs?
he apparently has a lot invested in making it sound like LN is used for purchases in adversarial conditions redefining what we all call DNMs to "anything that has a .onion" is convenient for him.
Yena's avatar
Yena 2 months ago
I'm sure there's subreddits where buyers and sellers meet, also there is porn subscriptions so I don't see why all three of these wouldn't be DNMs. Just different rules for each platform. Actually anyone could proxy a market to an onion domain and it wouldn't be possible to know otherwise. Therefore, you couldn't actually prove any online shopping isn't a DNM. right?
mister_monster's avatar
mister_monster 2 months ago
If bitcoin exists only to cure monetary inflation and any guy can spin up an attempt at inflating it, it's not very good at it's only purpose, is it? Of course, those premises are incorrect. Bitcoin exists to create a peer to peer electronic cash system. Curing monetary inflation is only a part of that. And these other things exist primarily to provide alternative peer to peer electronic cash systems based on peoples idea of what is an improvement. They're not fungible with bitcoin and therefore cannot inflate bitcoin, the only way they devalue bitcoin is by attracting users that like their feature set. The free market at work, for money, competition based on merit, it is natural selection. If bitcoin is the best and the only thing worth using, they'll have little impact I'm the short term and virtually none long term.
you have to <ominous music> pass some flags to your node
Super Testnet's avatar Super Testnet
> Is that difficult? Yes, for most people. Manipulating configuration files requires a lot of background knowledge that most people don't have. Using command line arguments too. You also neglected the part about installing linux in the first place, and then installing tor. For most people, I suspect they would assume "installing tor" means "installing the tor browser," not running "pkg install tor" or "sudo apt install tor" or whatever it is since it depends on which linux operating system you're using. And if someone gets past all that, then you have to somehow find out that (1) you have to keep it running (2) the port switches from 9050 to 9150 if you're using tor browser -- and if you're on Windows, everything is different there too. So yeah, it's difficult. Not for me or you, but we've got a lot of background knowledge that you can't assume for most people. > Making a new wallet per transaction is schizo IMO Yet -- according to the monero website -- that's the only way to get transaction unlinkability in monero. You think *lightning* has a UX problem? Meet monero. > your criticisms about having to make new addresses and the like are all REQUIREMENTS for bitcoin on-chain even to the same vendor. You can reuse the same address with the same person with Monero's and there will be no on-chain link There will be no "on-chain" link, but that person will have an off-chain link that is just as provable. An identical thing is true for monero's addresses and bitcoin's xpubs: if you trust the sender with your privacy, you can give him a monero address or a bitcoin xpub (or a bitcoin SP address, more recently) and if he reuses it, then "on-chain" there will be no link between different transactions sent by that person. Unless the sender makes mistakes too. But even if he does everything *else* perfectly, he made this mistake: he kept your address or xpub and reused it -- and that means he kept a record or proof that links multiple transactions together. If the address or xpub leaks to untrustworthy people, e.g. if he is subpoena'd for that information, your adversaries learn those links and can prove them in court. I think that's a terrible for privacy. Do you? Or is it only bad when it applies to bitcoin, and perfectly fine in the case of monero? > Please do proper journalism by using the protocol you describe I've used monero plenty and I think it sucks.
View quoted note →
Depends on how recent the wallet was last synched but it takes a bit, just as syncing from a Bitcoin core node takes a bit. There are light wallet solutions akin to Electrum for bitcoind like Edge wallet but similar to Electrum there are major tradeoffs privacy wise. You could hypothetically run a node on a phone if it has enough storage (around 260gb unpruned last time I checked) but similar to Bitcoin core (about 570gb last time I checked) storage is the limiting factor. Most node in the phone LND implementations rely on light wallet syncing like Electrum or Neutrino nodes.