You are viewing a single comment's thread from:

RE: How Does Steemcurator01 Support Affect Authors' Behavior?

in WORLD OF XPILAR3 years ago

Just to revisit some points, since it's been a long conversation. The original implementation of /promoted was done for people who transferred STEEM or SBD to the @null wallet. For performance, I guess it would be much faster to follow that model and check that wallet's incoming transfers instead of looping through every post to look for posts with beneficiary settings.

OTOH, the beneficiary never changes, so that's the kind of thing that you could capture once, the first time you look at the post, and never check again.

Finally, on the point about using a tag to signal promotion, the nice thing about this signaling method is that people could start immediately after curation support is available and the UIs could follow along when ready. In this case, the beneficiary method would probably be more convenient than the "transfer to wallet" method because curators can validate the setting with a simple mouseover (after the payout goes above 0.00).

A quick-start idea that just occurred to me is to set up an account to do a recurring automated post (burning a large percentage of rewards) to publish a list of posts that have been promoted via either/both of the above methods. If I can find some time during the next couple weekends, maybe I'll give that a try and just see what happens.

cc: @o1eh

Sort:  
 3 years ago 

Looking at the wallet makes a lot of sense. Paid-for-promotion where the "buyer" is sacrificing their stake in the knowledge that an auto-upvote won't follow. It's what bidbots should have been.

A quick-start idea that just occurred to me is to set up an account to do a recurring automated post (burning a large percentage of rewards) to publish a list of posts that have been promoted via either/both of the above methods. If I can find some time during the next couple weekends, maybe I'll give that a try and just see what happens.

I'll be interested to hear how you get on - I'm struggling to make progress with most of my projects at the moment... motivation is lacking.

I said to somebody recently about the well known coding term "Garbage In, Garbage Out". I didn't realise it at the time, but sub-consciously I was referring to my work in writing Steemit a new front-end. What's the point in making something look pretty, when 90% of what it's being filled with is garbage. I started "blacklisting" certain things - e.g. #krsuccess but then I realised I'd be blacklisting so much. Too much. Too many users with too much content created purely for bidbots. Which led me to a selfish route - creating a VIP version of Steemit where users are whitelisted, not blacklisted. Your content is included by invitation only. Any abuse of that invitation removes those privileges. Getting users accepted could be a pain but... I could gamify it like the plagiarism game. A user applied, everybody has their own reputation to approve or reject an applicant.

I don't know. I'm exhausted thinking about it and it's destroyed all the motivation I had to work on it.

I saw a couple posts by you after this comment, so it looks like maybe you got a little bit of a second wind. My only suggestions on the whitelist would be to consider (eventually) letting individuals decide whether or not to activate it and providing the ability add or delete people according to their own tastes.

The abuse is frustrating, but I think your identification of the swarms of 15SP delegation voters has helped a lot. I notice that the votes per block count has leveled off around 6 now.

Also, I came across this article a couple days ago, and I find it strangely reassuring when I see that the same sorts of problems happen in other places too. I haven't read the related preprint, yet, but I might take some time and do that. It might contain some Steem-relevant information.

Also, I thought their differentiation between "diagnostic" vs. "screening" was interesting. Seems like your script to identify potential 15SP delegation abuse is a good screening technique within that same framework.

 3 years ago 

My only suggestions on the whitelist would be to consider (eventually) letting individuals decide whether or not to activate it and providing the ability add or delete people according to their own tastes.

I think I've got a 2nd wind - my mind's been wrestling about how to do the black/white list and it was only fairly recently that an appropriate approach felt right. I'd also tried refactoring the code a lot to make the multi-tag queries more efficient... which didn't work so I had to roll back and "lost" quite a lot of effort.

My current thoughts are that the idea of users having their own whitelists that they can add and remove users from is the existing friends list - so I'd piggy back off that. When you mention the ability to activate my whitelist - are you suggesting that the "star user" tag is not on anybody until it's activated? (I've already implemented the ability only to see "star users").

The abuse is frustrating

I'm planning to create other tools to help fight against abuse which should be fairly easy to implement. My 15SP abuse only highlights abuse in the comments and not posts themselves which will be a bit more difficult to identify but I have an idea (posts that have received a lot of upvotes that are only worth a few cents). This should hopefully return even more SP to the Steem team.

The other one is analysing the posts that steemcurators04 - 08 are upvoting so identify upvoting of crap and see if there are linked accounts within the wallet addresses and transfers. With a view to highlighting behaviour like this before it's too late.

Sorry, got distracted by this and need to log off now

Coin Marketplace

STEEM 0.18
TRX 0.16
JST 0.029
BTC 62497.97
ETH 2428.72
USDT 1.00
SBD 2.65