Live data from Hacker News

Twitter doesn't care about spam?

twitter.com

71–80 of 192 posts

Re: Twitter doesn't care about spam?

#71
post #9

Option d: nobody outside of the Twitter TnS teams (including Paul Graham) has any idea of what Twitter is detecting as spam and what not, and thus can't actually judge how good a job is being done. First, we don't see the spam that was outright blocked, so we have no idea of what the false negative rate is. Second, I bet that this is not a binary block/allow decision, but there are all kinds of ways of reducing the e…

Just tweet metamask lol - its the magic word. Edit: WARNING this will immediately fill your feed with spam. Don't click anything that you get.

Hilarious. This is actually true.

https://twitter.com/nickstinemates/status/148716175094431334...

Re: Twitter doesn't care about spam?

#72

Twitter is also incredibly bad at identifying offensive Tweets - https://inteoryx.com/htmls/TwitterOffensive.html

What is universally or objectively "offensive"? Is a young attractive woman who is only following me to try to get me to follow her OnlyFans, "offensive"? Is a die-hard materialist atheist, or a strident born-again Christian "protected by the vaccine of God", "offensive"? Is plain nudity "offensive"? Violent photos or movies? Vanilla sex? Hardcore sex? BDSM sex? Bad words? (Bad word filters are easily defeated with c…

But it could at least be self-consistent. Sometimes identical tweets by different users are given different verdicts. Twitter's offensiveness detection seems about as good as Google's search results.

I've just realised a potential reason for this: we're seeing a spam-detection algorithm several years down the line from when Twitter started. That's several years of cat-and-mouse, where the spammers have access to much higher quality data than Twitter does. If their algorithm was simple, predictable or accurate, the spammers could just work around it. It's plausible that Twitter has run the numbers and determined that this is the best they can do, at the moment.

Re: Twitter doesn't care about spam?

#73
post #64

I have a four letter dormant Twitter account. I get about four mentions/replies per day that link to some nondescript crypto airdrop since several weeks. I manually mark each and every such tweet as spam and block the user, which needs 5 clicks or so. Nothing seems to help.

> four letter dormant Twitter account

You know at some point Twitter will just arbitrarily snatch your user name up and assign your account a randomly generated sequence. It's been done before without a given explanation. Just keep that in the back of your mind.

Re: Twitter doesn't care about spam?

#74
post #9

Option d: nobody outside of the Twitter TnS teams (including Paul Graham) has any idea of what Twitter is detecting as spam and what not, and thus can't actually judge how good a job is being done. First, we don't see the spam that was outright blocked, so we have no idea of what the false negative rate is. Second, I bet that this is not a binary block/allow decision, but there are all kinds of ways of reducing the e…

Just tweet metamask lol - its the magic word. Edit: WARNING this will immediately fill your feed with spam. Don't click anything that you get.

"Metamask" is pretty effective, but it only gets better if you combine it with a bunch of other triggers. I've seen some people deliberately craft tweets to get all of the bots to show up at once, e.g.

> metamask trust wallet support seed phrase recovery bitcoin shib ethereum network nft help moon coinbase defi dapp shib dogecoin giveaway retweet elon musk free nfts crypto punks opensea free raffle dao hacked banned from Instagram snapchat facebook i need an artist logo designer

Re: Twitter doesn't care about spam?

#75

Earlier quoted context omitted.

Just tweet metamask lol - its the magic word. Edit: WARNING this will immediately fill your feed with spam. Don't click anything that you get.

Hilarious. This is actually true. https://twitter.com/nickstinemates/status/148716175094431334...

Interestingly the spam is usually under the "possibly offensive replies" fold, so Twitter is detecting that somethings up but falls short of blocking the spam altogether.

Re: Twitter doesn't care about spam?

#76
post #43
post #3

Relevant tweet from Elon Musk yesterday: "Twitter is spending engineering resources on this bs [support for profile photo from NFT] while crypto scammers are throwing a spambot block party in every thread!?" - https://twitter.com/elonmusk/status/1484456594775678976?cxt=...

The best part about this is that the people above telling us they haven't seen spam on Twitter in years just need to click this tweet and scroll a bit.

I did, saw no spam. Clicked "More Replies" three or four times. Still none unless you count https://twitter.com/saylor/status/1484546928704696321

Can you link a few? Maybe we have algorithm sort set differently? Are you showing "Newest First" or whatever it's called?

Re: Twitter doesn't care about spam?

#77
post #59

Earlier quoted context omitted.

Crazy how we can have such different experiences. I follow less than 100 accounts and easily 50% of the responses are always spam/scams. It makes Twitter almost unusable for me. The most ironic part is that Twitter will show the spam/scams, but hide other replies, from real people, that Twitter thinks are mean.

I follow around 1000. I see almost zero spam. This whole thread is pretty confusing to me.

Could it be that you are perhaps unaware of what is a cleverly veiled spam reply or a scam reply? Lots look conversational, even genuine.

You only have to mention NFTs, even in a negative light, and you generally get drive-by likes from verified NFT-related accounts, and if you check their 'Likes', they've liked dozens more tweets since yours. The idea being to draw attention to itself with a certain demographic. Twitter has its own option to report people for this exact behavior.

Likewise, the number of QRTs in replies to things boggles the mind.

Re: Twitter doesn't care about spam?

#78
post #45
post #19

Earlier quoted context omitted.

Option E: Nobody can even agree on what Twitter spam is. Is it bots? I follow a few bot accounts and get value out of them. Is it blatantly promotional content? This actually seems to be one of the intended uses of Twitter. Is it low-quality content? I feel like lots of people earnestly tweet out low-quality stuff, and who gets to judge what's low-quality anyway?

tweet> really hard day today, my great aunt died of covid after a long fight reply> I know we all probably must have heard about Bitcoin but don't know how it works, I tried it in a week ago and i made a return of $10500 after a week of trading, connect with my trader at @SpamBtcAccount1234

Sounds like the variable is whether one engages with Bitcoin discussion. I don't, and have observed no spam.

Re: Twitter doesn't care about spam?

#80
post #61
post #16

Earlier quoted context omitted.

Huh? Can you explain this one, please?

Metamask is a crypto wallet manager, and often people mention it in Tweets because they're having some kind of problem using it, so spambots reply with links to fake "Metamask support" pages that helpfully guide you through the process of handing your private keys over to someone who will immediately steal all your bitcoins and monkey JPEGs. That particular type of spam has been around for a while now so people must…

Thanks for the explanation. As I've stayed far away from cryptocurrencies, I thought it was Covid related at first.
Post reply on HN