You are viewing a single comment's thread from:
RE: Today's Truffle Picks: Quality Steemit Posts that deserve more Rewards and Attention! (03.03.2018)
I think you need some adjustments.... Number 6 is Just a copy of the steem bluepaper.
Definetly not worth that much SBD.
Is this a blunt copy and paste? TrufflePig relies on @cheetah to detect plagiarism, too bad he didn't caught that :-/
Are you thinking in implementing some way of manual Review? The Idea behind the bot is great, but maybe when It have some big sp on It, some manual Review might be needed.
@cheetah is pretty good, but not perfect.
Also, i Will start to campaing soon for people to add some sp to this bot, because finding 24 good posts among the Sea of shit that is steemit is a pretty amazing feat, and these posts really deserve more rewards!
Truth is, I'm not planning for manual review, at least not on a regular basis. I rather try to improve the bot when he taps into edge cases or suffers from structural problems.
Latest example, yesterday he endorsed a post by
haejin. Of course, from algorithmic point of view, this makes perfect sense.haejingets constantly enormous rewards (about ~200SBD per post). Hence, if one of his posts just gets 20-30$ by chance (or massive flagging), this is by definition a truffle. But is it, really? No, of course not, becausehaejinbasically exploits Steemit. So I had to improve the bot by filteringhaejin's contributions in general.Coming back to your point, indeed, there could be manual review services operating on top of the bot. In fact, there is already one: @sneakyninja said that she or he regularly consults @trufflepig for inspiration for the daily sneaks (the bot even made into one of the sneaks, yeah!).
But doing a regular manual review of the bot's output myself is definitely out of reach. So if you delegate SP to the bot, you have to live with the risk that sometimes it can make errors and feature dubious content. Still, this might be worth it. As you said, if you are able to reward 24 good posts on a daily basis and just 1 shitty one by accident, but thereby drive Steemit in general more towards a focus on quality posts, the one false positive might not be that bad after all :-)
Still, I'll try to mitigate shitty recommendations as good as I can from a technical perspective by improving the algorithm.
Anyway, it's great to hear that you already trust the bot even so far to convince others to delegate SP to @trufflepig. Big thank you!
Good points you made. I like the idea of automation of our day to day lives, but i always think that robots shouldn't be left running around without supervision. They may take over the world.
The only sad part is that i don't have too much SP to Delegate yet, and a bit of it is already delegated to another amazing project that is @utopian-io.
But i will definetly start a campaing to raise fundos to the bot. I only have 400 followers, wich even though is pretty good, i dont think most of them have too much SP either.
But i will try to reach other whales that seems to have a good faith on the platform. Heck, i will even try to reach @ned himself. Don't know how though...
All other curation services that i follow are pretty amazing, but they may still be kind of biased, because of humans.
I will let you know when i finish the article praising the bot and asking for delegation for it.
BTW, are you planning on setting investiment returns for the bot posts earnings? i think this seems to attract a lot of SP delegations around here.
Yes, @utopian-io is a cool project, good choice :-)
I haven't thought about it, yet. I am also not so sure how the details could work out.
ROI of delegating SP to Bidbots is rather straightforward. The earned SBD from each bidding round can be paid out immediately. But how do you payout SP if this requires a 13 week powerdown? This would involve a lot of book-keeping.
One way would be to distribute the sbd earnings from the bot posts (like the daily picks and comments) proportional to the delegation.
The self vote from the bot would go closer to payout as possible, to stimulate upvotes for curation rewards.
The SP earned should be kept on the bot to increase its Power.
Maybe some automated SQL script to gather the data from the delegation and assing the SBD rewards proportion to sponsors.
I think there is some Scripts or similar programs around that do ano automatic payout for sponsors. Maybe you should contact @elear. I think utopian sponsors pay is automated.
Probably Will be less than the bid bots roi, but It is some incentive.
And you , as creator, take like 5 or 10% cut of the rewards, to keep the bot running.
Hmm interesting thought. Still, I wouldn't want people to invest in the bot for a ROI, because than bidbots are much better anyway. But if getting a little bit back makes people give more, than why not?
And you are right, distributing daily SBD income according to vested shares seems like a good idea.
Why SQL? I don't want to maintain a SQL server. The beauty of the bot is that it is (except from some very basic caching) completely stateless. All statefulness is externalized by the means of the blockchain. I definetly want to keep it this way at all costs. Having to manage state makes stuff from an engineering perspective extremely complicated and annoying.
Maybe in the long run. However, truth is, I wouldn't be so generous in the beginning. Currently the bot does not pay for itself. So maybe I would start with more something like 50/50 to make sure server costs are covered.
Again, I don't want people to invest in the bot to make money, but rather to support content creation and put something automated up against the inflation of bidbots who do not care about content at all.
Just found this: https://steemprojects.com
Add trufflepig there!
Yeah, people are more willing to contribute If they get something in return.
I fact i have a Very Basic understanding of coding/programming, so It was Just a Dumb suggestion based on what i read about SQL. Just ignore It.
Makes total Sense, and i dont see It would harm the investor Desire to delegate a bit.
Just need keep It clear How the return Will be calculated when the funding campaing starts.
Also, if you think about it: This is actually pretty cool, it singled out the BluePaper as a pretty valuable publication. The bot cannot know about plagiarism unless it is flagged by @cheetah! So to him this post is the original BluePaper :-)
Indeed... Too bad It isnt the author.
I browsed a bit among these hardcore plagiarizing accounts. There are all not very valuable with quite a low reputation. Maybe it makes sense to include reputation in the final ranking step.
For example, if your rep is below 40 your ranking score is halved. However, that increases the false negative rate. Fresh new users and minnows that produce fine and original content, but have a low reputation because of their account age, would be punished by this as well. Hard to tell if this is worth the risk. I would suggest to observe the bot for a while and see if more plagiarized content pops up by these accounts. If so, filtering will be needed.
Maybe a blacklist of confirmed scammers/plagiarists?
Add a warning on his posts after the blacklists, and leave it open for a future review if they improve.
Yes, I would love to get my hands on such a blacklist. I tried to obtain the one used by @steemcleaners but haven't reached anyone on Discord, yet. But this is definitely on the agenda.