"Initial Reputation Staking"

Feedback Feature Requests
1

I uploaded this to reddit but realised this would be more fitting.

AI talent is a scarce commodity as big business take all the cream. If AI talent could be guaranteed a stable income for a period of development time required, this would incentivise individual talent toward fostering a positive AI future within the SingularityNET.

I propose a way to finance AI talent by the use of Reputation Crowdfunding. The community could crowd fund for projects and developers by staking AGI based upon their ideas, achievements, qualifications and reputation. This could be groups of people or talented individuals.

Community stakes AGI on the future work of said talent in the Initial Reputaion Staking engine which is then paid out in installments through smart contract for work achieved by their own proposed roadmaps. As the roadmap hits its milestones the smart contract releases a portion of AGI to said developer ongoing. The communties staked AGI would be held in a smart contract but issued 1:1 ā€œ'Devā€™tokensā€ (each project a different token) which could be traded back for AGI to new community stakers at any time. The ongoing reputation market decides the value of the "Dev"Tokens based upon possible future returns. As the milestones are hit, early stakers are rewarded with increased AGI per "Dev"Token based upon the timing and Reputation of their initial stake. When the roadmap has been completed the devtokens become the early staked AGI tokens in the Reputation Market for that particular project. It would also allow for the whole AGI community to learn and participate in the Reputation Markets as new projects enter into the space and create profiles.

I believe this would quickly foster SingularityNETs Ecosystem to develope industry products faster than any other project which would solidify its market share in Decentralised AI products and services.

Particulars and projections would need figuring out to implement anything like this.

All ideas welcome. :slight_smile:

4 Likes
2

Good concept that reflects the practical process of a decentralized democracy. It has been discussed on the telegram philosophy channelā€¦ An idea worth stakingā€¦

2 Likes
3

@akolonin @bengoertzel what do you think? We will have a request for AI portal, and a reputation systemā€¦ I think this would add a new dimension which might be interesting in the long-term. Not sure about the Dev Tokens though, but I like the idea.

1 Like
4

Hi FractalUniverse, really like your idea and thanks for proposing it. I was initially bit confused by what you were proposing but I think I get it now. So just to clarify:

Letā€™s say John, whoā€™s an AI developer, has a new AI project in mind - lets call this project MED - but rather than trying to seek funding for it through other routes, John seeks funding via SingularityNET.

So John estimates he needs $100,000 for the creation, training of Algo, hiring of talent, marketing, building of website and to meet all other costs for the MED project.

The supporters of the MED project fund the project via AGI. The total funds raised are $100,000 - so they get MED tokens with initial value of $100,000. The AGI raised are locked in a smart contract, and as John advances in his projects via already declared milestones - the MED token holders vote to release the next installment of AGI.

As the project advances, the market decides the price of MED Tokens and the value of the MED project.

After project is launched and AI node becomes live on SingularityNET - the MED Tokens are no longer traded but become staked tokens for the MED Project - so rather than AGI being staked on the project, the tokens of the project can also be staked on the project? Converting the tokens into AGI right before the launch of project will not work, as someone will have to buy the tokens with AGI - and why would someone buy MED Tokens if theyā€™d be worthless after project launch. So the tokens themselves will have to be staked. And token owners will get staking rewards.

If my understanding of the model is correct, then one problem I see in the model is that there might not be enough of an incentive for people to fund projects.

In this model the tokens of the projects will not be utility tokens, but will derive their value entirely from the fact that they will be getting the staking rewards for the project.

According to the White Paper, there are 40 Million AGI Tokens reserved for staking rewards over the next ten years - 10% of that 40 Million is to be released each year - so 4 Million AGI rewarded each year. And these will be for all the projects. Iā€™m not sure exactly how staking will work but Iā€™ve been an investor in the crypto space for some time and though I may be wrong, my personal belief is that if tokens only get the staking rewards, people might not be attracted to invest in projects.

Also, having custom Tokens act as stake along with AGI on project might prove messy to implement.

Perhaps a different way to go about all of this might be to implement the ā€˜Buy Back and Burn Modelā€™ - as in the smart contract running the project node or the team of the project shifts a fixed and already agreed amount of earnings of the project to buying back the project Tokens and burning them. In this model, the tokens will still be traded after launch of project and anyone holding the tokens can expect good returns if the project is a success.

1 Like
5

Thatā€™s basically the plan: Reputation System Design for SingularityNET | by Aigents with Anton Kolonin | SingularityNET
To implement this plan, few options are possible:

  1. Wait for some past-alpha release of SingularityNET to have this supported in some way - in 2019 as it is not on the roadmap yet;
  2. Consider independent Reputation Agency created to serve purposes of SingularityNET as well as other purposes - this has been discussed with earlier but not on the plan either;
  3. Use current functions of Aigents API and the reputation subsystem being developed now, as presented in this video Personalized social connectivity and reputation with Aigents - YouTube and this one Aigents News and Reputation Graphs - YouTube - so some reputation computation functions may be employed this year for any application like in the case of Reputation Agency above.
    Note, I am personally kind of cautious about ā€œstakingā€ component of reputation because it can be easily turned into ā€œgamblingā€ which is not desirable, from my point of view. But as long ā€œstakingā€ is part of composite reputation scheme, it could be balanced carefully, possibly.
    @bengoertzel + @cassio - what do you think?
1 Like
6

The gambling aspect should be addressed asap. Our platform ( blockchain) governance framework should address this by creating a high degree of certainty in staking. First things first. A Constitutionā€¦ Then a formal governance framework. I have roughed out a blockchain governance framework for snet if it helpsā€¦ We could roll out and tune this as part of our beta releaseā€¦

2 Likes
7

Here is the latest design: [1806.07342] A Reputation System for Artificial Societies

1 Like
8

The model contains security conceptsā€¦ What is the best transmission method do you think?

9

Sorry, missed your replyā€¦ Reading nowā€¦

10

OK, interesting readā€¦ The constituted quality management (QA/QC) internal and 3rd party processes seem closely aligned to proof of reputation. Data bases would be held for non individuals by Lloydā€™s, Det Norse Veritas etc. I am not sure how digitised or available they are. For nation states POR, WEF/World Bank. POR frameworks is UN,/ISO. I do not think they collect any rating data. They would have a perspective on the value of independent ratings agencyā€¦ Kay Firth-Butterfield is the Head of Ai/ML at the WEF. She is involved with AIforgood, as is Hanson roboticsā€¦ Hope this helps , sorry for the confusionā€¦

1 Like
11

Thanks, I think this is rather insightful for PR advances on this side of the Reputation Agency project for @Mitchell & @ibby

1 Like