An impossible hypothetical: On quality and value
I was having a chat with @abh12345 who spends a lot of time neck deep in crap the numbers of the blockchain and a few things came up and came to mind. This is one that he said,
"give me some metrics and i'll have a look"
But, it is impossible to give those metrics. However, impossibility should never stop people from thinking through various subjects because often, what seems impossible today gets a solution tomorrow or, half a solution or, something else interesting or of use that is completely unrelated but got sparked by the investigation into useless.
Speaking of useless investigations, let's get into the issue of quality on Steem. There s a lot of talk about what is quality and what is not and of course the go to argument is that it is all subjective which means, no matter the metrics applied, they are never going to be a perfect fit and match reward to quality. We don't need to let that stop us talking about if it was possible though.
Let's work under the assumption that there is some super quantum computer with an AI tasked with distributing the reward pool based on the quality of content provided. It is perfect and has worked out every metric imaginable that creates a perfect line of best fit. It is not gameable.
Remembering that this is hypothetical, it has complete control over the reward pool and each day trawls every post submitted to the blockchain over the last 24 hours, evaluating each on the quality and at midnight, votes accordingly. We can be guaranteed that the decision it makes is completely fair and the reward from the pool is perfectly aligned to the quality of the post in comparison to others submitted on that day.
What do you think the result would be?
Let's then look at if there is another metric stream added to the mix on top of quality, value to the platform. This would perfectly align the value to readers, onboarding potential, retention, code and economic capabilities. This is then factored against the quality metric to create either a positive or negative multiplier. Something that is high quality and useful gets a high rating multiplier and something that is missing either will suffer a penalty multiplier.
What do you think the result would be?
Imagine that in this system, no matter what is posted, it will get exactly what it deserves. Again, this is hypothetical and we can ignore the dystopian nature of such a system but, no matter whether you are submitting a new app, a fiction story or a picture of a painting, the system can evaluate it and reward it correctly, no matter the language or location from which it originates.
What do you think the result would be?
Let's then go a bit further and include the engagement aspects of it. Remember, this is hypothetical so we can pretend that we would still engage and interact as if things are rewarded by the community itself. Comments, replies and all aspects associated would not only reward the comments but, apply an additional multiplier to the post and the quality/value matrix so it is: quality + value + engagement
Would you want such a system?
If we go on a little further and look at some of the factors that come under fire such as the Trending section. Trending is then organised in a way that is guaranteed to best satisfy the needs of the community across a wide range of data points that include a host of things that a human is unable to predict.
What is there at the top and what is not?
The AI could then of course create a multilier for consistency where even if an author dosn't have the highest quality content bu, they are able to produce a range of high quality content consistently, they could get an additional volume bonus for producing weight for the content mass.
Would authors be happy with being rewarded on consistency
What if we then go a little bit further and still incentivize manual curation and engagement by adding a positive multiplier to those who are able to accurately support the same content that the AI has supported and a negative multiplier to those who support what the AI has perfectly predicted as harmful to the system. For example, the rewarding of Spam comments.
Would that improve quality and level of support?
One logical step further would of course be for the AI to evaluate the content before it is submitted to the blockchain and stop it from being connected at all if found harmful or low quality. It could then send a report to the poster with a prediction based on yesterday's content results or a warning such as, "do you still want to post this?"
That last one is pretty dystopian already, isn't it?
But, for the thought experiment across the idea of quality and by extension, value, this is probably enough to think about. Yes, this is completely impossible in this current world but, it is interesting to explore considering the amount of discussion that surrounds it here.
What I wonder is, if there was a perfect system that had these capabilities and was perfectly fair on content regardless of the personal relationships or stake held, would people then be happy with the distribution? This impossible system would be the perfect meritocracy where content is evaluated piece by piece on its quality and on its influence on the community and platform and everyone would get precisely what they are entitled to from the pool.
Yes, dystopian but, where do you think your content would appear on this perfectly meritocratic platform?
A lot of authors and posters argue about the distribution of rewards and how it is about who you know and not what you know but, how many would be willing to have the control taken away and a perfect system applied so that they earn what they deserve in relation to the entire platform?
Would they be happy with what they get if their reward is perfectly tied to their independently and perfectly evaluated content, in conjunction with their engagement, curation and value provided to the community and platform?
Of course, this system is impossible and just a hypothetical for a thought experiment but I would predict that if the system did exist, there would be a massive increase in the level of quality across the platform but, most people here will earn zero or, close enough to it as those with real skills dominate the pool.
The pool would still be controlled by stake except stake would be skill, adaptability and the willingness to do ones best; not Steem Power
Taraz
[ a Steem original ]
For sure, lots of people will be unhappy. But maybe less unhappy people than at the moment!
The problem is how the AI will value a post. Because quality could have lots of grey areas. Just like it is now.
If I would write a post about sports betting, you probably would't be interested, even if it was hypothetical the best article of the day! One involved into sports betting would find it a great article!
The same goes for a sewing article, vegan food and so on!
Even an AI robot would not be able to make these distinctions.
Really a bad idea!!! But really bad. This would wipe out to complete genuinity from the blockchain. People will not post things while they maybe would like it because it would have no value and they will not get paid!
Even worse, they will changing their article until the bot says that it has value. This bot maybe just want to make sure, that all the point he thinks are valuable are within the post, destroying the complete idea behind the article.
Hence, would it even allow me to post an article why AI is bad for human kind?
Cheers,
Peter
Yeah, that is why it is an impossible hypothetical.
Absolutely a terrible idea but I had to throw it in also as to not, would be to deny what would likely happen. censorship.
A system which rewards people based on current skill, rather than on willingness and ability to learn, will eventually have no skill at all.
That is true but (for impossible hypothetical argument sake), this system would essentially drive the highest level of competition and as said, the AI isn't gameable which means that it would adapt to the needs of the platform. Again, this is dytopic imo but people would be forced to grow or not earn a living. The eventual no skill would be mass starvation until there is only one.
Highlander of Attrition.
This is not a good idea
I had a feeling a post on this topic would arrive soon enough :)
Just this one for me:
My Sunday post, the Engagement League, might sneak in there under #engagement or #contests for the prizes and number of comments on average that it gets each week. The rest of my content, same place as usual :)
I am pretty sure I will be low-mid level trash that people with too much time on their hands will finally land on after 6 hours solid reading of what was quality above :D
Nice thought but it won't work. You are really thinking out the box and trying to find a solution for this place. The skill being used on here is by how to beat the system. The community has adapted to bend the rules to suit themselves and not the needs of the community. Whatever changes happen on here will be good but also interesting to see how they adapt to bend it in their favor again.
At some point, the AIs will out-perform he greatest of scammers and then there will be no way out of the informational prison they build :)
This hypothetical system would definitely improve the overall health of yhe community and I think that some of my posts would fare much better than they do but some of my posts would definitely see a decrease in rewards.
Even with a perfect system in pkace there woukd be contingents of people who woukd complain and be dissatisfied because that is our nature as a society. We are quick to lay the fault of failure at everyone else's feet instead of owning our own outcomes based on merit. True unfiltered self awareness is a rare commodity and I strive to be as open an homest with myself as humanly possible because I see it as a pure oath to growth and improvement.
I think it would be fascinating to see the results of such a system and to see the community's, as well as the individual's, reactions to the results. This was an extremely thought provoking post!
I would assume that many of the highly rewarded would become a lot less so but, many of the lowest rewarded would remain so. People here who complain about the rewards (myself included), can't be objective enough as they have already invested the time and effort into a post so they assume there is value held in it. I can go out for a jog and it takes energy and time but, no one wants to watch people like me run at the Olympics.
I think later on there will be hybrid versions of this implemented for discovery purposes just like every other site and as AI improves, value will get rewarded but, valueless will struggle massively.
lol This, so this! I am definitely guilty of my efforts that I place on generating a post translating to an emotional attachment to the work rather that being able to look at it objectively and sometimes being honest with myself and saying "Hey dude, that's crap! You are better than this"
I am very interested to see how everything shakes out when and if a version of this is every implemented. It is a good mental tool to keep in the back of my mind as a guide to create better posts, a kind of "what would the AI program think of this post?"
This isn't addressing your question specifically, but we already have spam classifiers, and if we applied them to post comments I bet we could get a pretty good picture of posts that are engaging and valuable to the platform if we simply counted them.
And things that are gameable? Google search results are gameable but you don't hear people giving up on using that.
Machine learning methods to tweak out undesirable behavior is surely possible. Then again that kind of thing seemingly needs to be rather centralized. Steemit is rather centralized already anyway, so maybe it's a rather good fit. Ha....
I use a set of Encyclopedia Britannica from the 70s. Nothing like the musty smell of books ;)
I would assume that various front ends are going to do away with the decentralized idea and make discovery algorithmic on their particular platforms to attract and incentivise or, detract and disincentivise certain activities.
(Re: encyclopedias) I did like those things growing up. The only writing assignment I ever liked involved research reports of which the encyclopedia was one component. And then Google came along hahaha.
(Re: front ends) Yeah, that does seem to be the direction that I hear in terms of talking, but so far nobody seems to have done anything other than show hot, new, and trending feeds as is (promoted, lol).
To be real I think the real flaw here on steemit is yes, the posts quality, but when I dig in the"new" section I often find interesting stuff.
But, my main concern is originality and engagement ( mostly comments). I hate upvoting a post not knowing if that drawing or that text is actually original. I also hate making posts and receive 2 comments, one of which is clearly a bot and the other one a guy that didn't read at all.
It's hard even as a human to curate original content. I really enjoy digging in the crap to find nice people but I'm thinking of giving up.
Because what I often find is that the most interesting people are engaging more that just on steemit. They are on discord and Steem.chat. When I see a minnow with an avatar, I'm 70% sure this guy at least gets it.
/Rant
this was a nice comment.... thanks for your comment... I like comments.
It is amazing that people show up to a social media and assume that they don't need to actually socialize, they don't need to speak or be human. This is the largest issue with all of the bots as they 'play human' (poorly) on a system that is powered by engagement.
Great article.
Where the future is going is that every piece of content will get the exact value it deserves based on the value it brought. That will say a totally perfect content reward system. This is actually not that hard to create. Since most stuff being created we humans can very fast see if it's low value or real value.
Machine learning which is on the rise will be able to detect stuff even faster. We are moving closer to that reality every day. Already right now if you create some genuine real value the chance of you being rewarded from it is super high. In the future this will be close to 99.9% accuracy. This will create a positive feedback loop that will improve everything with compound growth speed.
as accurate as it would be, I think it would be hard to be fair with it from a human perspective across culture/economic classes but of course, the AIs will grow powerful enough to overcome these things also in time. Still sounds dystopian.
Why dystopian?
This system is what I prefer, AI doing the rewarding and bonuses allocation, imagine a group of people, posting daily quality contents and comments and all but still gets to be decided on for reward by those with greater steempower and reputation, Sir Taraz, I really like your effort towards minnows like me, I usually notice that when I comment on your post, you usually do vote me and I'm trying my best to be consistent, people like you should be able to change the current system, I see quality contents on some young creative mind's blog and after reading, you see the rewards given to that person so low and that won't encourage the person, I have a friend who loves drawing on his phone and he is so good, I introduced him to steemit so he could be helping communities with their logos and all and he joined and registered on the platform but today he's no more on steemit, Why? Because people just check his cool paintings without leaving an encouraging comment and reward but I've being trying my best to encourage him by inviting him but my steempower doesn't even show in his rewards at all because I'm also a minnow, so, Sir Taraz I think this system would bring even more people to steemit and encourage them too. Thank you sir @Tarazkp
To reward a lot of people consistently it take more people to powerup consistently yet, most do not. If a wider group of users grew their accounts a little, as price increases they will be able to support others together but those same people are currently expecting a small amount of users who have invested to support a very large number who have not. To retain people from all over the world it is going to take people from all over the world to invest.
Yes I see what you are saying, I'm also part of those investing, if you can check my profile, I powered up with the Steel I've earned so far but still not enough, not up to 1sp, I also want to be part of those who encourage others, it was my friend@jesse12 who introduced me to steemit who also encouraged me by voting my contents so I don't quit,
And what did he do to make it happen?
You have to believe in your stuff enough to make it happen on Steem or any other social platform.
Did he really try or just posted and cry he didn't get any upvote?
OK I agree with that, he didn't really try but at least I tried and didn't get anything until my friend told me to cheer up and encouraged me
Absolutely.. The things you mentioned would make the system fair.. Yes it is impossible in the current situation but as u said we should always explore... The pool would definitely be transparent and properly controlled on the basis of skill...
Do you think everyone would be rewarded?
Absolutely...according to the quality of the stuff they post and not merely by bid bot system..