I've been doing a lot of the same thought process for what I'm trying to set up. Here are the conclusions that I've come to (some of which are very close to yours and some which vary tremendously).
1. People post their ideas into some layered set of systems that records them permanently (a wiki or three is fine for ideas initially as long as it maintains complete histories but code needs to go somewhere better protected). Self-suggested values are nonsense, however, and people have to be aware of the fact that the AGI will eventually do a search for prior art (i.e. they will get some credit for introducing the idea but not as if it were their own invention). People will *eventually* be rewarded either by the AGI (if such occurs) or by a consensus of active contributors (if rewards are necessary before an AGI occurs). 2. The project will be incorporated. The intent of the corporation is to 1) protect the AGI and 2) to reward those who created it commensurate with their contributions. 2.1 New ideas/algorithms/code can be submitted under a variety of arrangements. The more that the arrangement favors the corporation, the more the contributor will be rewarded down the line (or more immediately if the contribution is used in intermediate profit-making projects). The intent here is to negatively influence defection while making it possible for people with extant AGI projects to participate. Note that the corporation will be more than willing to accept contributions from (or exchanges with) other AGI projects and needs to offer good terms to attract such. 3. It is nearly impossible to determine the source of *many* ideas; however, code that is developed from ideas that are clearly developed within the corporation belongs to the corporation (but, obviously, the code counts as a major contribution by the author). All derivatives of the code belong to the corporation. The contributor or anyone else with the corporation can sell the executable of such code at the corporation's profit. Needless to say, however, the corporation will reward that person accordingly and nothing prevents that executable from being embedded in some other product (or AGI project) with a reasonable licensing fee (or other compensation). 4. Identity verification is mandatory. There will be several levels of access to the corporation's work/results and access to the source code of the various modules will be granted on a need to know basis (yet another advantage to a good modular design with good interfaces) as determined by the corporations Board of Directors or it's designee(s). 4.1 Idea bleed to other extant AGI projects is unavoidable. What we need to prevent is the harvesting of the corporation's assets for the benefit of another project with no return. As long as an individual/project has contributed sufficiently, access to the source of additional modules will be granted singly as necessary (although it is unclear to me that a single individual is going to be that interested in the low-level source of that many modules -- unless, of course, they're just interested in running through them all and improving the code -- which just makes that person an asset and someone that we want to get vested in the corporation). 4.2 NDAs and NCAs are fundamentally un-enforceable except at the largest and highest levels. The corporation will have "honor" contracts/agreements, however, and the corporation or the AGI can release information about these (or act upon them regarding eventual compensation) as it sees fit. 5. The software patent system is fundamentally broken. We need some way to quickly register any obvious innovations as "prior art" to avoid patent trolls but otherwise steer clear of the patent system (note: this emphatically does not mean stealing other's ideas, however). The entire point here is to make it beneficial for an individual and other projects to contribute to (or make equitable exchanges with) the project while attempting to reduce the probability of theft without recompense. The fundamental problem with Open Source, particularly during development, is that there is more incentive for defection and theft at a well-chosen moment than there is for remaining with a project. Thus, this is not going to be Open Source (though a committed, contributing individual will eventually be able to see all of the source). This is also not a one-vote-per-person democracy. Influence will be commensurate with contribution (and the best way to influence the direction of the project is to put in effort and contributions in that direction) although good suggestions are always welcome (regardless of source) and will always be implemented (with credit to the contributor) if appropriate. How's that? ----- Original Message ----- From: YKY (Yan King Yin) To: [email protected] Sent: Friday, June 01, 2007 5:33 PM Subject: [agi] Open AGI Consortium How about some brainstorming...? My proposal is this: 1. People post their ideas onto a wiki and discuss them, while carefully keeping a record of who has said what. Also, each person suggests an amount of how much the contribution is worth. If the amount is outrageous people can make complaints about it. 2. Suppose the group end up with some useful ideas / algorithms. Each result will be collectively owned by that result's contributors. 3. Suppose someone (a developer) wants to take a result and implement it? The developer will have to pay a license fee to the contributors, the fee being proportional to the total estimated worth of its "constituents". 4. Also, everyone who participates, must sign a non-disclosure and non-competitive agreement (NDA & NCA). There should also be some way to verify the person's identity. 5. I think this scheme can work for existing AGI projects like Novamente. It will not compromise the control over their ideas / intellectual property because of the NDA & NCA. 6. If something is deemed patent-worthy, the patent will be collectively owned as in (2). The licensing price will be set analogous to (2), so it won't be outrageous. How's that? YKY ------------------------------------------------------------------------------ This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?&; ----- This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=231415&user_secret=e9e40a7e
