Steemit how can this thing be workable long term? - page 24.

smooth

legendary

Activity: 2968

Merit: 1198

Quote from: iamnotback on July 16, 2016, 09:42:27 PM

What is the "recommendation engine"? Please define new terms.

Recommendation engine is a generic term: https://en.wikipedia.org/wiki/Recommender_system

Dan made a post about the specific algorithm he used, but I wouldn't guarantee that the post matches the code. The code is the only real specification.

Quote

So why did you disagree above.

Because of the tension between the monetary incentives and the utility incentives of the recommendations for one thing. For another, the recommendation system in Steem was't addressed in my post. I think it probably didn't exist at the time, but I don't precisely remember (the post includes some comments about the concept as a potentially viable direction to explore which suggests that it didn't).

iamnotback

sr. member

Activity: 336

Merit: 265

Quote from: smooth on July 16, 2016, 09:37:20 PM

Quote from: iamnotback on July 16, 2016, 09:13:30 PM

You can see that smooth's tension between popularity and quality is due to the lack of degrees-of-freedom in ranking groupings:

https://steemit.com/steem/@smooth/voting-is-a-popularity-contest

Btw, I wrote this in one of my early posts in this thread, but I doubt anyone picked up on it. Because I always write FUD. Roll Eyes

There is no tension there. My conclusion is clear that a voting system can't measure subjective quality. You apparently agree.

Disagree. If voting is grouped by shared interests, then voting represents both quality and popularity within your like-minded group.

Users may even have different groupings for different subject matter (hashtags).

Quote from: smooth on July 16, 2016, 09:37:20 PM

There is a recommended posts feature that uses collaborative filtering on your own votes to find posts of (likely) particular interest to you, but the UI of the site doesn't make it obvious. The existence of this feature (if people actually saw it and used it) undermines my point to an extent.

Correct. So why did you disagree above. But please also describe the way this feature computes the like-minded groupings?

But note a filter alone won't correct the monetary rewards, which will leave the incentives game theory in tension of a one-size-fits-all groupthink.

Quote from: smooth on July 16, 2016, 09:37:20 PM

If you are a regular user of the recommendation engine, your incentive is to vote for what you actually like, to make the engine work better.

What is the "recommendation engine"? Please define new terms.

smooth

legendary

Activity: 2968

Merit: 1198

Quote from: iamnotback on July 16, 2016, 09:13:30 PM

You can see that smooth's tension between popularity and quality is due to the lack of degrees-of-freedom in ranking groupings:

https://steemit.com/steem/@smooth/voting-is-a-popularity-contest

Btw, I wrote this in one of my early posts in this thread, but I doubt anyone picked up on it. Because I always write FUD. Roll Eyes

There is no tension there. My conclusion is clear that a voting system can't measure subjective quality. You apparently agree.

There is a recommended posts feature that uses collaborative filtering on your own votes to find posts of (likely) particular interest to you, but the UI of the site doesn't make it obvious. The existence of this feature (if people actually saw it and used it) undermines my point to an extent. If you are a regular user of the recommendation engine, your incentive is to vote for what you actually like, to make the engine work better.

iamnotback

sr. member

Activity: 336

Merit: 265

I always make a 5000 kg blubber splash when I drive in the pool, because I paradigm shift the way to think about an issue:

Quote from: iamnotback on July 16, 2016, 08:37:46 PM

My first comment is a reply to smooth's comment:

Quote from: https://steemit.com/money/@biophil/is-steem-paying-for-groupthink-the-game-theory-of-steem-part-3#@anonymint/re-smooth-re-biophil-is-steem-paying-for-groupthink-the-game-theory-of-steem-part-3-20160717t013639963z

Quote from: smooth

It's getting harder and harder to find all the hidden gems on Steemit these days.

That is unavoidable in the current system design, due to the one-size-fits-all reputation system, i.e. each user will have different priorities but the design of the system doesn't accommodate such degrees-of-freedom.

If some of the voting power shares your preferences, that content will rank higher than content that interests none of the voting power, but assuming that interests are reasonably diverse then ranking will be more or less uniform and uncorrelated to individual preferences. So thus more or less in order for any content to rise up, it must be a groupthink effect.

An improvement would be some algorithm which allows each grouping of like-minded interests to have their own separate ranking computation. The monetary reward algorithm would also need to change, so as to reward content that ranks highly in any grouping.

I elaborated:

Quote from: https://steemit.com/money/@biophil/is-steem-paying-for-groupthink-the-game-theory-of-steem-part-3#@anonymint/re-trogdor-re-biophil-is-steem-paying-for-groupthink-the-game-theory-of-steem-part-3-20160717t021129871z

Quote from: trogdor

I suppose that one possible response is that whales have an incentive to preserve the value of their investments, and the best way to do that is to promote a system of fair voting and promote the integrity of the system

The one-size-fits-all ranking system (c.f. my other reply to smooth below) makes it impossible for whales to act rationally, because they can't compute a set of votes which would reflect their individual preferences for quality which might be shared with other like-minded users.

Thus as far as I can see, the system disincentivizes the whales from participating in voting, for they will come to see that either they become one dysfunctional groupthink monolith or they more or less effectively nullify each others votes in terms of anything other than a uniform ranking which is functionally equivalent to no ranking.

You can see that smooth's tension between popularity and quality is due to the lack of degrees-of-freedom in ranking groupings:

https://steemit.com/steem/@smooth/voting-is-a-popularity-contest

Btw, I wrote this in one of my early posts in this thread, but I doubt anyone picked up on it. Because I always write FUD. Roll Eyes

AlexGR

legendary

Activity: 1708

Merit: 1049

Quote from: iamnotback on July 16, 2016, 08:57:43 PM

Quote from: AlexGR on July 16, 2016, 08:53:25 PM

See, you are already making money Cheesy

Lol. Thanks for the vote. Now I can afford to buy a stick of bubble gum.

Try your luck with an introducing myself post, with proper tags etc... you might be surprised Cool

smooth

legendary

Activity: 2968

Merit: 1198

Quote from: iamnotback on July 16, 2016, 08:59:09 PM

Quote from: smooth on July 16, 2016, 08:55:24 PM

Quote from: iamnotback on July 16, 2016, 07:47:10 PM

I signed up so I can write some frank blog posts and see what sort of reception they get:

https://steemit.com/@anonymint

Apparently login is disabled at the moment, so I can't post a blog.

It seems they fixed the bug I reported in this thread yesterday and now I see the total balance for 10 SP estimated to be $28.

It isn't a bug entirely, though maybe undesirable behavior. The blockchain uses a 7 day median price feed. Given the enormous price increase it was lagging. That usually won't be the case.

The original "bug" (source of confusion for n00bs) I reported here was that it was displaying the estimated price per SP, not the estimated total balance in dollars. That seems to be fixed.

It wasn't! The feed was off by a factor of 10 at one point, after the price rose by factor of 10 over a few days.

Quote

Next they need to display the estimated balance in the local currency. Some people can't convert dollars to local currency in their head. Display currency should also be selectable.

Good idea.

smooth

legendary

Activity: 2968

Merit: 1198

Quote from: magicalacademy on July 16, 2016, 05:57:59 PM

Quote from: tobysc on July 16, 2016, 05:13:42 PM

-
Basically what I mean by this is that posts of higher quality generally gets more votes than those of lesser quality.

The problem is that dan, ned and smooth decide what a quality post is. And so far they have decided that boobs and steem praise is a quality post.

That's not exactly true. I don't vote for Steem praise more or less ever (I think circlejerking is absolutely bad for the platform) and sometimes downvote it. Dan downvoted the girlsgonesteem-nsfw posts.

After the July 4th payout there are many more users with stakes that matter. Not as much as Dan or me in a particular instance, but given that I don't even vote much they may matter more overall.

There is a sort of evolving consensus about what is 'good' content for the site. Both whales and less-stacked users matter in that consensus.

iamnotback

sr. member

Activity: 336

Merit: 265

Quote from: smooth on July 16, 2016, 08:55:24 PM

Quote from: iamnotback on July 16, 2016, 07:47:10 PM

I signed up so I can write some frank blog posts and see what sort of reception they get:

https://steemit.com/@anonymint

Apparently login is disabled at the moment, so I can't post a blog.

It seems they fixed the bug I reported in this thread yesterday and now I see the total balance for 10 SP estimated to be $28.

It isn't a bug entirely, though maybe undesirable behavior. The blockchain uses a 7 day median price feed. Given the enormous price increase it was lagging. That usually won't be the case.

The original "bug" (source of confusion for n00bs) I reported here was that it was displaying the estimated price per SP, not the estimated total balance in dollars. That seems to be fixed.

Next they need to display the estimated balance in the local currency. Some people can't convert dollars to local currency in their head. Display currency should also be selectable.

iamnotback

sr. member

Activity: 336

Merit: 265

Quote from: AlexGR on July 16, 2016, 08:53:25 PM

See, you are already making money Cheesy

Lol. Thanks for the vote. Now I can afford to buy a stick of bubble gum.

smooth

legendary

Activity: 2968

Merit: 1198

Quote from: iamnotback on July 16, 2016, 07:47:10 PM

I signed up so I can write some frank blog posts and see what sort of reception they get:

https://steemit.com/@anonymint

Apparently login is disabled at the moment, so I can't post a blog.

It seems they fixed the bug I reported in this thread yesterday and now I see the total balance for 10 SP estimated to be $28.

It isn't a bug entirely, though maybe undesirable behavior. The blockchain uses a 7 day median price feed. Given the enormous price increase it was lagging. That usually won't be the case.

Quote

Why did I receive 10 SP tokens and my gf's sister received only 5 SP yesterday? Did they increase the signup bonus? If they did, maybe they are trying to improve the attrition rate by giving newbies more incentive to not abandon their 10 SP.

I have no idea. There is a minimum account balance as a consensus value but it is actually lower now, not higher.

AlexGR

legendary

Activity: 1708

Merit: 1049

See, you are already making money Cheesy

iamnotback

sr. member

Activity: 336

Merit: 265

My first comment is a reply to smooth's comment:

Quote from: https://steemit.com/money/@biophil/is-steem-paying-for-groupthink-the-game-theory-of-steem-part-3#@anonymint/re-smooth-re-biophil-is-steem-paying-for-groupthink-the-game-theory-of-steem-part-3-20160717t013639963z

Quote from: smooth

It's getting harder and harder to find all the hidden gems on Steemit these days.

That is unavoidable in the current system design, due to the one-size-fits-all reputation system, i.e. each user will have different priorities but the design of the system doesn't accommodate such degrees-of-freedom.

If some of the voting power shares your preferences, that content will rank higher than content that interests none of the voting power, but assuming that interests are reasonably diverse then ranking will be more or less uniform and uncorrelated to individual preferences. So thus more or less in order for any content to rise up, it must be a groupthink effect.

An improvement would be some algorithm which allows each grouping of like-minded interests to have their own separate ranking computation. The monetary reward algorithm would also need to change, so as to reward content that ranks highly in any grouping.

iamnotback

sr. member

Activity: 336

Merit: 265

I signed up so I can write some frank blog posts and see what sort of reception they get:

https://steemit.com/@anonymint

Apparently login is disabled at the moment, so I can't post a blog.

It seems they fixed the bug I reported in this thread yesterday and now I see the total balance for 10 SP estimated to be $28.

Why did I receive 10 SP tokens and my gf's sister received only 5 SP yesterday? Did they increase the signup bonus? If they did, maybe they are trying to improve the attrition rate by giving newbies more incentive to not abandon their 10 SP.

I figure that one of their internal calculations is that it probably costs > $100 to acquire an ecommerce user normally, so they probably figure if they spend that much per user then they receive fair value, given Dan's plans to monetize the site with a marketplace.

P.S. I have (currently back of the napkin idea) suggestion to make which I think may help them drastically improve the attrition rate. I need to write it down and think it out a bit to clarify if it is correct.

Zer0Sum

legendary

Activity: 1588

Merit: 1000

Quote from: CoinHoarder on July 16, 2016, 05:50:36 PM

STEEM PROPOSAL: Abolishing Liquidity Incentives And Reverting The Funds To Provide Actual Autonomous Liquidity

https://steemit.com/sip/@coinhoarder/steem-proposal-abolishing-liquidity-incentives-and-reverting-the-funds-to-provide-actual-autonomous-liquidity

So the person who designed the mind-blowingly over-complicated Bitshares pegging system...
Designs something a 12 yo might draw up on the back of a napkin in 5 minutes.

>>> Just give ALL rewards to ONE person that provides the most "liquidity" every hour in a system with no tx fees.

Of course, your ideas will make $3.00 because none of the 30 insider whales will touch it.

Also, the whole "vote posts in a tiny 15 minute window within the first 30 minutes"...
Can only be designed for the whales to game and control the system (zero Steemit posts are time sensitive).

Quality music (Appetite for Destruction) or films (Shawshank Redemption) often take years to become hits by word of mouth...
So there is no logical reason for a blogging platform to disproportionately reward snap decisions on content.

I'm really starting to think Steemit is full of existential holes... and could crash and burn.

AlexGR

legendary

Activity: 1708

Merit: 1049

Quote from: iamnotback on July 16, 2016, 06:35:28 PM

My understanding is everyone is moving away from ads towards gamification.

You basically have to spam ads in order to generate income. And the income is net loss overall to your ability to compete for usership.

My own idea was about opt-in ads. If you don't want them, you don't see them, you get paid less. But at least it's an option for usd acquiring steem which are then turned to ads. It breaks the closed economy loop.

iamnotback

sr. member

Activity: 336

Merit: 265

My understanding is everyone is moving away from ads towards gamification.

You basically have to spam ads in order to generate income. And the income is net loss overall to your ability to compete for usership.

AlexGR

legendary

Activity: 1708

Merit: 1049

Quote from: iamnotback on July 16, 2016, 06:29:49 PM

Quote from: AlexGR on July 16, 2016, 06:26:11 PM

Quote from: iamnotback on July 16, 2016, 06:06:20 PM

Any way, we already documented that advertising can't generate more than roughly $15 per user per year.

Facebook is a bad benchmark due to less targeted content.

A discussion platform can provide very targeted content and -by extension- very targeted ads.

The difference in efficiency between targeted and non-target ads is night and day and can justify ad revenues for targeted ads exceeding 100x compared to non-targeted.

Sorry it is much worse for Reddit:

https://www.quora.com/Why-are-reddits-advertising-revenues-only-8-million-in-2014-compared-to-Instagram-with-700-million

https://www.quora.com/Why-does-Reddit-make-only-10-million-in-revenue

I can't even remember seeing ads on Reddit.

iamnotback

sr. member

Activity: 336

Merit: 265

Quote from: AlexGR on July 16, 2016, 06:26:11 PM

Quote from: iamnotback on July 16, 2016, 06:06:20 PM

Any way, we already documented that advertising can't generate more than roughly $15 per user per year.

Facebook is a bad benchmark due to less targeted content.

A discussion platform can provide very targeted content and -by extension- very targeted ads.

The difference in efficiency between targeted and non-target ads is night and day and can justify ad revenues for targeted ads exceeding 100x compared to non-targeted.

Sorry it is much worse for Reddit:

https://www.quora.com/Why-are-reddits-advertising-revenues-only-8-million-in-2014-compared-to-Instagram-with-700-million

https://www.quora.com/Why-does-Reddit-make-only-10-million-in-revenue

AlexGR

legendary

Activity: 1708

Merit: 1049

Quote from: iamnotback on July 16, 2016, 06:06:20 PM

Any way, we already documented that advertising can't generate more than roughly $15 per user per year.

Facebook is a bad benchmark due to less targeted content.

A discussion platform can provide very targeted content and -by extension- very targeted ads.

The difference in efficiency between targeted and non-target ads is night and day and can justify ad revenues for targeted ads exceeding 100x compared to non-targeted.

iamnotback

sr. member

Activity: 336

Merit: 265

Topic: Steemit how can this thing be workable long term? - page 24. (Read 32368 times)