# Proof of Concepts and Successes | AGI Laboratory

- Canonical URL: https://wefunder.com/feed/147125
- Entity ID: wefunder:feed_item:147125
- Published at: 2021-07-29 01:42:27 UTC
- Updated at: 2025-07-09 03:20:36 UTC

## Author
David J Kelley

## Subject
AGI Laboratory

## Content
For the past 7 years or so, AGI Laboratory was just about research.&nbsp;For the most part, all of that research was generally based on working code examples.&nbsp;Someone mentioned they would want to see more of the proofs and concepts and productization we have done.&nbsp;&nbsp;To review all of our published research, you can look at those papers here (most of which is peer-reviewed):&nbsp;&nbsp;https://agilaboratory.com/research/&nbsp;Now that being said, a fair percentage of this is behind paywalls. If you see a paper you want to read, feel free to contact us, and we will send it to you privately.&nbsp;So with that background, let's look at some of the proofs of concept.&nbsp;The most important of which is the mASI test system.Another example of proof of concepts is the application of collective systems to e-governance.&nbsp;This study is still in progress but will be wrapped up by September.&nbsp;If you are interested, the experimental framework is available already and passed peer review on its own.&nbsp;&nbsp;https://agilaboratory.com/wp-content/archive/mASI%20E-gov%20Experimental%20Framework.v5.pdf&nbsp;The most important example of our technology is the main research system, or mASI.&nbsp;We did a cognitive study with the help of the University of California.&nbsp;The results of this study are here:https://agilaboratory.com/wp-content/archive/Preliminary_Analysis_Results_of_ICOM_mASI_v11.pdfThis first screenshot is the administrative dashboard showing the state of the machine.&nbsp;There are some videos floating around of this and of the mediation process, but this screen gives you a high-level view of the research system.&nbsp; If you look closely this dashboard gives you a direct view into what the system is experiencing emotionally at any given time both at a conscious level and subconscious&nbsp;level.&nbsp; Additionally there is a section showing composite emotions.&nbsp; These are the emotions that are composites of any two emotions next two each other on the Plutchik emotion model enforced on the system.&nbsp; Additionally there is the list of state values including the hard numbers of the emotion models that tell you about the condition of the system, for example you can see how many models are in memory, how many mediation login from the last one and other details about the working system.&nbsp;Now many people have asked for code, and we wrote a book that we can send you, but it won't be public tell all the patents are done.&nbsp;But looking through that book can be misleading in that it only shows the most straightforward path, and that takes 70 pages.&nbsp;So here is a screenshot of a test solution of the main codebase.&nbsp;You can see on the right that there is a lot to the system.&nbsp;When mediators log in, they see the following screen.&nbsp;Responses or models are generated before mediators see them.&nbsp;Mediators essentially are able to add 'tags' and emotional valences.&nbsp;And this doesn't necessarily change the model directly, but all of the mediator's input is averages against what the machine set first and all the other mediators, with all of the tags being associate with a given model.&nbsp;after all the mediators have audited or mediated a given model the system may choose to hard reprocess where it would go back to mediation or do a minor reprocess and skip additional mediations.&nbsp;Here is what the mediators see in the research system.From this screen, we can see how mediators add emotional data in this screen:&nbsp;After this screen, they also add tags, and there are a few other analysis tools they can use if they choose, but it is not mandatory, and they can't actually change anything in the model directly.&nbsp;&nbsp;While this video is more than a year old and based on an older version of the research system, it is still generally valid. It does a complete walkthrough of the system at that time from a mediators standpoint:&nbsp;One of the first things after the Wefunder is done to do a proper UX study and build UI for the open-source voting framework.&nbsp;We have had some design work on new screens and the look and feel, which shows a bit more of where we are doing from a design standpoint, but this doesn't include all the types of voting and policy analysis that will be built into the product tool:Keep in mind these preliminary designs were based on the more general collective intelligence. In contrast, the open-source voting system will be more focused on just voting, certifying on a blockchain, Policy analysis, and so forth.&nbsp;