Hacker News Reader

Comment by scyclow | Hacker News Reader

I wouldn't think so. BTC uses an unspent transaction output model, which is like breaking up a plot of land into smaller pieces, and treating each square inch as fungible. ERC20 tokens are based on an accounting model, so you're just incrementing and decrementing balances -- nothing to trace there.

View on HN

The "accounting model" is even easier to trace, I'm not sure what you mean by that. For example, here are all USDT transactions: https://etherscan.io/token/0xdac17f958d2ee523a2206206994597c...

Here's what I took away from the suggestion. Let's say I send you $100 USDT and Osama Bin Laden sends you $100 USDT. Then you send Justin Beiber $100 USDT and keep the remaining $100 USDT. Where are the tainted tokens?

For Bitcoin these are UTXOs which are easy to trace, but not so for Ethereum and its tokens which just rely on accounting not UTXOs. Right?

If Bin Laden sends $100 USDT to you, your address will be tainted because Osama sent you coin and proof of that will be in Transfer Log of ERC20 token. So if you send any amount of that token anywhere, recipient address will be also tainted, doesn't matter if it was Bieber part or Osama part.

Got it, so now Justin Beiber's wallet is tainted too.

> Right?

No.

> Let's say I send you $100 USDT and Osama Bin Laden sends you $100 USDT. Then you send Justin Beiber $100 USDT and keep the remaining $100 USDT.

That's a misconception, you can do that exact same thing with Bitcoin and you are left with the same question. Bitcoin outputs can be combined into a single output (and they routinely are). See this diagram: http://syskall.com/bitcoin-notes/svg/transaction.svg

Even Ethereum's rationale acknowledges that Bitcoin's UTXO model provides a higher degree of privacy: https://eth.wiki/fundamentals/design-rationale

I never thought of it that way. I wonder what the article's author thinks about Ethereum's privacy.

Reply on news.ycombinator.com