KNW Token Q&A Session with Knowledge.io Head of Product, Steve Englander

Transcript from a Q&A session on the t.me/knowledgeio Telegram group chat

The Knowledge whitepaper can be found here: knowledge.io/whitepaper

Question from Interviewer:

As stated in the whitepaper, Advertisers can incentivize users to contribute reviews based on their Knowledge Score offering KNW tokens.

Let’s say a large conglomerate of companies decides to advertise their products on the platform. They purchase millions of tokens, and pay users to review their products.

Don’t you think smaller advertisers without the financial capabilities would be trampled upon?

Their products wouldn’t gain massive audience because of this.

Is there a way to solve this problem?

Answer from Steve:

There are several common methods for handling this type of situation. Similar to the current model in the programmatic advertising ecosystem, they use a 2nd price auction model to prevent hoarding in the manner you describe. It’s called a Vickrey auction.

We’re going to explore all options available, and since we intend to leverage the existing programmatic infrastructure to some degree further down the road to achieve massive scale, we’ll already be leveraging that model.

Interesting to note, the programmatic landscape has evolved to the point today that they are looking to move to a first price auction model.

Question from Interviewer:

Can you explain how this auction method works?

I still can’t see how it solves the problem I stated.

Answer from Steve:

https://en.wikipedia.org/wiki/Vickrey_auction

Unfortunately, I’m not a mathematician, so I would not be the correct person to explain this at length, but there is that Wikipedia page which explains it in some detail.

Question from Interviewer:

Read about it. I don’t think you understood my question. The users this time around are not buying the product so no need for auctioning.

They are reviewing the product based on their Knowledge Score.

Example. If I have 1000 tokens and I want users to review my product. Let me say I pay each of the reviewers 1 KNW token each. So I would only be able to target 1000 reviewers?

I can’t stand a chance to compete with large conglomerates that have millions of tokens and can reach millions of people.

How would this problem be solved?

Answer from Steve:

Okay. I believe I understand your question. I thought that you were asking how advertisers will be prevented of accessing a lions share of the supply by overbidding intentionally to win over smaller advertisers.

The question you’re asking requires a quite complex explanation, and I’m not sure you have the complete understanding of how the ratings, reviews, and influencer type content works within the platform, although I could be wrong.

Only certain users will emerge as Knowledge Stars for any given topic, and less users for combinations of topics.

These users are involved in correctly answering questions throughout the ecosystem — the questions have topics attached to them.

Those topics also relate to advertiser’s products.

So advertisers — who are also the bulk of sellers in the marketplace — will discover these Knowledge Stars as they emerge and be able to make performance ad model (CPA) style deals with Knowledge Stars for assisting in the referral of the sale. This is the model that we think works the best, that most advertisers are already used to, especially in the ecommerce space — most of them already run affiliate programs.

So the total supply is limited, there is choice on behalf of the Knowledge Star in whether to accept it, and the Knowledge Star could essentially lose some credibility if the products they recommend are not of the level of standard they mean to represent.

There’s much more I could say about it within our development team discussions regarding how the calculations and algorithms are designed to work, but I don’t think it’s a good idea to get into more detail beyond that at this point. Unless you have a follow up.

Question from Interviewer:

Since the advertisers approaches people with high Knowledge Scores in a topic or combination of topics, there might be a shift of paradigm.

Only a specific set of people would garner attention from advertisers while the rest wander within the community.

In a scenario whereby a trusted Knowledge Star expert or group of trusted experts becomes bad, they might give bad reviews about the advertisers products and they would still be paid because the advertisers pay before review.

Don’t you think these would cause a problem in the community? How would these problems be combated?

Answer from Steve:

It would be left to the advertiser/seller’s discretion to terminate the relationship with a Knowledge Star. The Knowledge Star could also terminate at any time.

Response from Interviewer:

The advertisers would have lost large amount of money before the contract is terminated

Counter-Reponse from Steve:

Let’s rewind a moment.

CPA model payments are performance based, so they only end up paying a commission for sales referred.

It happens post-event.

But we would require the payment to be “deposited” and make it so the system could draw upon this deposit to handle those distributions.

Let me look back at the question to see if I’ve addressed all of it.

Knowledge Stars are not trusted really. It’s conditional and the scores vary over time. This gets into more of the calculations and algorithms that assess the weightings of a Knowledge Star against the rest of the community and other “similar” Knowledge Stars, and while interesting, is too detailed to explain here.

Nothing in the system is trusted except math. It doesn’t require trust. The Knowledge Score is fluid.

Question from Interviewer:

I like the idea of using CPA model in curbing the issue of Knowledge stars reviewing poorly about the advertiser’s products.

A drawback here is that a reviewer might spend his time and put his best into reviewing a product and because of unforeseen circumstances, purchases didn’t turn out quite well through his reviews.

Does it mean that all his efforts would be wasted?

Answer from Steve:

I wouldn’t say wasted. There is some value to both the advertiser/seller and the community in general. But the Knowledge Star wouldn’t necessarily be restricted to exclusivity on retail products that appear in multiple retailer inventories. I prefer not to discuss more about those specifics, but let’s just say there will be different “types” of deals that can be made with retailers.

The Knowledge Star can have multiple relationships.

Question from Interviewer:

Also, it is stated in the whitepaper that Data pools would submit shares of knowledge from their users.

I hope this would be encrypted if not the vulnerability of the entire system would be at this point.

Answer from Steve:

That data would most definitely be encrypted in transit.

In this context, the data pools are primarily publishers of apps, websites, and other internet connected platforms.

Response from Interviewer:

That’s great.

Counter-Response from Steve:

They could also be advertisers. And other ad tech platforms.

Question from Interviewer:

I need more explanation on “littlebit”.

Answer from Steve:

It’s the analog to a satoshi.

Question from Interviewer:

It wasn’t also explained how the percentage of KNW tokens awarded to a data pool for operating and maintaining the pool is calculated

Answer from Steve:

The process by which KNW tokens are distributed to participants on the Knowledge.io ecosystem is just like bitcoin pool mining. We call this process meta-mining. Just like a bitcoin pool mining, which has:

1. fixed supply

2. increasing difficulty over time

3. halving of rewards

KNW tokens are distributed in a similar way, fractionally and proportionally based on contribution of Knowledge to the pool, meeting all 3 criteria above.

In the Knowledge.io ecosystem, our rewards work like bitcoin pool mining. Publishers are mining pools, and their users/visitors are individual miners on the pool. Rewards are issued based on the individual’s contribution and the publisher earns tokens for operating the website, app or platform. Anyone who wants to advertise or interact with the knowledge.io ecosystem has to buy or earn KNW tokens.

Our token is about rewarding users for sharing Knowledge. Via our SDK/APIs and ad platform, users are able to earn tokens when answering Q&A style questions, and for a few other things that add value such as humanity verification and identity verification.

As people answer questions, we store their scores of correct, incorrect, interest, and intent based responses against topics associated with the questions they’re answering on the Knowledge Score blockchain. The calculations for correctness and the associated reward are distributed across nodes to achieve consensus, and then the KNW token which lives as an ERC-20 compliant token on the Ethereum blockchain is told to update address balances periodically using smart contracts.

There’s more to it but users are an analog to miners submitting a hashrate to the publisher pools.

Response from Interviewer:

This is the wholistic explanation. Thanks. Am asking for just a specific part from what you just stated.

Counter-Response from Steve:

Please clarify. Unless I already answered it above.

Clarification from Interviewer:

Rewards are issued based on the individual’s contribution and the publisher earns tokens for operating the website, app…

There is no explanation for how the pool is paid for operation and maintenance. I understand how users in the pool are being paid. What about the pool maintener. How is he paid

Continued Answer from Steve:

They’re paid in a similar way as a mining pool, in cyrptographic hashing physical mining.

Question from Interviewer:

OK thanks. Are there plans in the future to create your own blockchain?

Answer from Steve:

I’m leaning towards yes. Possibly several. For various purposes, as subsystems on the ecosystem. An example would be to detect bots. But I can’t really say more than that at this time.

I hope you’ll understand that as stewards of the responsibility of decentralizing the entire system most in accordance with fulfilling it’s purpose, we shouldn’t discuss these things at this time.

Question from Interviewer:

Concerning the data pool ( publishers for example). Let me say I have a website and I integrated KNW API into it.

Through the API, I am able to publish contents of advertisers.

I would be paid in KNW tokens definitely. But how would the users of my website be paid since they don’t even know that I am advertising for an advertiser in the Knowledge ecosystem.

My users would just see an advertisement and keep on reading an article on my site. Or let me say they click on the advertisement, how would they be paid.

I am making an assumption, that my users haven’t heard of KNW before.

Answer from Steve:

1. We leverage an OpenID server to Login with Knowledge. In some cases users will be logged in already when they reach your website, because it’s connected in some way to the tags (code snippets) for the API/SDK, and in this way we would be able to recognize the user and award KNW tokens.

For the users that are not logged in, it is in the best interest of the publisher to get their users to login. Through various means, either by some prior access using a Login with Knowledge auth on the website, or within the ad unit, the user who is familiar or not logged in, or the user who is unfamiliar will be able to login.

2. The ad units will have some layering, and the Q&Ads unit will have something similar to the AdChoices icon layer. In this way, we could show the KNW token icon there for users who are not “Logged in with Knowledge.

Question from another Visitor:

Ok, KNW token will be POS or POW?

Comment from Interviewer:

POW definitely.

Answer from Steve:

KNW tokens are ERC-20 standard tokens on the Ethereum blockchain.

Counter-Comment from Steve:

There might be some POS or something new under the hood as a subsystem.

But primarily yes, POW in what I’m describing above.

Continued Answer from Steve:

3. We can support regular banner ad type ads, but advertisers will likely want to use the Q&Ads units more heavily within our system. We do want to however allow them to leverage the Knowledge Score targeting externally within other ad platforms over time.

Comment by Interviewer:

This means that sites like blogs were users only view and read information would be beneficial to only the publishers and not the users of their site.

Continued Answer from Steve:

4. Users who are already logged in with Knowledge will earn tokens, and thus the publisher pool will earn tokens.

We have two states of tokens, claimed and unclaimed. The not-logged in user will have the status of unclaimed tokens, but they are allocated, and could possibly reintegrate back into the meta-mining reserve if not claimed (details of which I cannot reveal at this moment).

The publisher is automatically in claimed status if they are accessing through our SDK/APIs, because they have gone through the KYC process already.

If later on when we decide to leverage the existing programmatic advertising infrastruture to show ads on effectively 90% of the internet’s inventory within our reach, those publishers may or may not be known and part of our private supply network. For those that are not within our network, the status is unclaimed.

At that point it becomes an effort to convince the publisher to claim.

Question from Interviewer:

Have you considered a scenario whereby users answer questions in the trivia app or other apps integrated into Knowledge ecosystem and get rewarded with Knowledge tokens. This increases the total amount in circulation with time.

What if there are no advertisers to put back KNW tokens into the system, just new users answering questions to earn tokens?

There would be a point where no tokens would be available to the users answering questions leading to the breakdown of the system.

If you can foresee this, how would you solve the problem?

You have to assume anything can happen, so please don’t tell me that there won’t be lack of advertisers or users that would purchase goods.

PS: This is a scenario where there are more users acquiring tokens than putting back into the system.

Answer from Steve:

The quick answer is that it all proportions itself out. Separately we will manage a reserve.

Response from Interviewer:

Lol. I knew that would be your answer. A solution would be to buy back the tokens when events like that tend to happen

Counter-Response from Steve:

Well, I can’t talk about how, due to our standard response to any of those types of questions regarding where.

Please note: We are not at this time able to answer questions about which, if any, exchanges our token may or may not be available on, nor are we able to comment on the price other than the sale price during the token sale. However, it is well known common practice and customary that most tokens end up available on exchanges.

Additional Comment from Steve:

Please keep in mind we also take transaction fees on the marketplace.

Response from Interviewer:

Good. But it would be nothing compared to the amount users get for answering questions.

For a system to be scalable, Reward received > Transaction fees

Counter-Response from Steve:

I can’t say what I want to say in response to that, but it might surprise even you.

Let’s just leave that one there, but I understand what you’re saying, and I believe we have a great solution.

When we do announce it, you’ll understand immediately.

Question from Interviewer:

Why did KNW token decide not to use IPFS for the database?

It’s more transparent than having a centralized database that can be manipulated

Answer from Steve:

We’re looking into all options. And we’re not trying to centralize the database. But some things cannot be transparent when dealing with sensitive info, therefore it will be encrypted through some process.

Many thanks to the Interviewer from https://twitter.com/simpleWhitepapr for the time spent researching the KNW token and Knowledge.io

To join the conversation, visit us on our Telegram chat.