Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Do not rug on me: Zero-dimensional Scam Detection

arxiv.org

Post reply on HN
Print view View original

131–140 of 158 posts

Re: Do not rug on me: Zero-dimensional Scam Detection

#131

Post by 40four »

cush wrote: post #117
40four wrote: post #102

This title should be changed. It is heavily editorialized. It’s not what the paper is about. It is just causing the typical crypto haters to get out their pitch forks. This article is about fraud detection. The conversation should really be about the machine learning techniques used, and the actual conclusion, which is using those techniques, ” This implies that new malicious tokens can be detected prior to the malic…

The title is in the conclusion of the paper. Yes, they built a fraud detection system... and with it discovered that 97.7% of tokens were rug pulls "Based on this theoretical foundation, we provided a methodology to find rug pulls that had already been executed. Not surprisingly, we found that more than the 97,7% of the tokens labelled were rug pulls."

I understand, but it isn’t the conclusion. I read it obviously, the sentence I quoted is in the conclusion as well. The title is spurring the same old boring “crypo bad/ crypto good” arguments we’ve been over 1000 times.

We should be talking about the techniques described in the paper, and the fact that they determined it is possible to detect early on with some confidence if a project is solid, or a scam.

This should be a positive for the “crypto bad” folks if you ask me. But this is getting lost.

Re: Do not rug on me: Zero-dimensional Scam Detection

#132

Post by GaylordTuring »

synu wrote: post #70
Sargos wrote: post #60

Earlier quoted context omitted.

If you go to Uniswap right now you won't see any of these tokens. You run no risk of getting scammed by using Uniswap unless you go out of your way and manually add them.

Looking at the paper abstract, it says the identified scams by looking at transactions. So it seems at least some people are manually adding them and getting scammed. Or are you saying anyone who does that is by definition using Uniswap wrong and shouldn’t be counted as being scammed on Uniswap?

Remember that you can be aware of something being a scam but still trade it, since you believe that the thing won’t collapse just yet. A very risky behavior of course, but it doesn’t mean that you have been scammed yourself.

Re: Do not rug on me: Zero-dimensional Scam Detection

#133

Post by nuclearnice1 »

Sargos wrote: post #80
synu wrote: post #70

Earlier quoted context omitted.

Looking at the paper abstract, it says the identified scams by looking at transactions. So it seems at least some people are manually adding them and getting scammed. Or are you saying anyone who does that is by definition using Uniswap wrong and shouldn’t be counted as being scammed on Uniswap?

The paper does not analyze the transactions directly unless I'm missing that piece. There are a number of reasons accounts would interact with the scam tokens such as the scamming group filling up the liquidity pool, creating fake volume to add legitimacy, and other steps to complete the scam. There's also the obvious case of the scammers getting a user to actually buy the token (perhaps through spam email or "pump g…

> The paper does not analyze the transactions directly unless I'm missing that piece.

It looks at transactions and attempts to classify coins as scam or not.

It doesn’t make any claims about the volume in scam coins. That’s more a clickbait headline and HN thread thing.

It’s most trying to contribute an algorithm for identifying scam coins.

See the appendix for features from the transactions that they used for their algorithm.

Re: Do not rug on me: Zero-dimensional Scam Detection

#134

Post by ceejayoz »

mritchie712 wrote: post #97
miohtama wrote: post #55

Some comments: While the report findings might be correct on the number, the terminology and conclusion in related discussion are often not. No real person ever traded these 26k tokens. The creation of liquidity pool should not be considered a launch, as you still somehow need to convince other people to buy it from you. Nothing was "launched". These trading pairs are more akin to spam, to show up in the search resul…

spam email is a good analogy. Email would suck if you needed some central authority to grant you permission to send. spam filters work on email and will get better for crypto.

> Email would suck if you needed some central authority to grant you permission to send.

You do. If Gmail, Yahoo, or Microsoft block your sends, you might as well give up sending email.

Re: Do not rug on me: Zero-dimensional Scam Detection

#135

Post by GaylordTuring »

LightG wrote: post #101
Sargos wrote: post #60

Earlier quoted context omitted.

If you go to Uniswap right now you won't see any of these tokens. You run no risk of getting scammed by using Uniswap unless you go out of your way and manually add them.

Does Uniswap prominently advertise that as a warning to users? I don't know anything about the space, but it sounds like a fact and statistic that should be front and centre in peoples face all the time.

Yes, when you add a token manually, you’re warned that anyone can add a token and that you should exercise caution.

Re: Do not rug on me: Zero-dimensional Scam Detection

#136

Post by cush »

zeroclip wrote: post #28

Bit higher than expected but not that surprising. Deploying a new token costs almost nothing, but may lead to outsized rewards for scammer. It can be easily automated. Reminder that 85%+ of all email sent is spam.

That's a fun stat. We're 6x less likely to be spammed by opening a random email than scammed by investing in a random token on Uni.

Re: Do not rug on me: Zero-dimensional Scam Detection

#137

Post by mritchie712 »

ceejayoz wrote: post #134
mritchie712 wrote: post #97

Earlier quoted context omitted.

spam email is a good analogy. Email would suck if you needed some central authority to grant you permission to send. spam filters work on email and will get better for crypto.

> Email would suck if you needed some central authority to grant you permission to send. You do. If Gmail, Yahoo, or Microsoft block your sends, you might as well give up sending email.

right, but you don't need permission to start and you earn or lose your reputation based on your actions.

it's the same with uniswap. you can list without permission, but you won't have any credibility unless your actions build them.

Re: Do not rug on me: Zero-dimensional Scam Detection

#138

Post by PaulHoule »

dannyw wrote: post #12
onlyrealcuzzo wrote: post #7

What constitutes a scam? I'm about as much of an NFT hater as you can be. But people spend money on lots of "dumb" stuff. So why is an NFT a scam if Pokemon cards aren't? Both of them are artificially scarce. Both of them are mainly about buying cute pictures. And at least for a lot of the buyers, it's about speculating the price will go up in the future.

Not a lot of people spend 10k or 100k on pokemon cards. Quantity matters.

There are the people who spend figures like that on Gatcha games:

https://www.youtube.com/watch?v=z-fxfuWhff0

I was a fan of Fate/Extra and Fate/Extella, so I was curious about Fate/Grand Order and found it very hard to run on any Android emulators, the NVIVIA Shield, AMZN Fire tablets and the other off-brand Android devices I have a huge collection of. Eventually I got it running and found the English translations make no sense at all, but it is moe-driven more than story driven and it wouldn't bother fans at all.

Re: Do not rug on me: Zero-dimensional Scam Detection

#139

Post by chucksmash »

parkingrift wrote: post #121
hailwren wrote: post #114

Earlier quoted context omitted.

Not sure what gemini has to do with this, but most people who actually think crypto is interesting will tell you that _any_ centralized exchange (to include coinbase, gemini, the NYSE, even the ones in the future that haven’t been made yet) is against the ethos of crypto. Uniswap is an entirely different class of thing. The only scams possible in the _logic_ of Uniswap are those which would be considered bugs in the…

Ah the “No True Scotsman” fallacy. No true crypto enthusiasts would use a centralized exchange. If they’re using an exchange, they must not be a true crypto enthusiast. https://en.wikipedia.org/wiki/No_true_Scotsman

"Not your keys, not your coins" has been repeated ad nauseum on crypto forums for years. It is very much part of the ethos and received wisdom, even if often discarded.

To say that centralized exchanges whom you trust to keep your assets in their custody run counter to the spirit of a system conceived with trustlessness as the defining feature is not rhetorical voodoo or fallacious reasoning.

Re: Do not rug on me: Zero-dimensional Scam Detection

#140

Post by PaulHoule »

onlyrealcuzzo wrote: post #25
Kiro wrote: post #14

Earlier quoted context omitted.

> The terms scam and malicious token are not being used identically by all researchers. For example, papers such as [44] use the terms scam and under-performing token indistinguishably, leading to inaccurate results and classifications. In the current paper, we define a malicious token as one released by a developer or a group of developers with no intrinsic value or use case. I guess the test token I put on Uniswap…

Who decides what the use case is? Pokemon cards technically have a use case. So do crypto kitties and smooth love potion - even if it's somehow slightly "dumber".

The real purpose is the expectation you'll sell them to a greater fool.
Post reply on HN