Please provide sources supporting your statements.
In other words, you think that the moderator function is enough to allow you to add rules allowing you to decide who practices disinformation and vandalism.
Start by justifying your statements concerning Joe Faust, who is an AWE pioneer, a hang gliding professional, a world class champion in high jump.
To help you, here is his wikipedia page:
Joe also looked hella sexy in a soap commercial 
On Joe Faust: Questions about Moderation - #156 by Windy_Skies
Wikipedia has the resources to combat it. You would be the main contributor to the wiki. If even you don’t see the problems with unchecked and unsourced AI content, the thing is cursed from the beginning.
Again, if the AI correctly copied and summarized the contents from the sources, the person who wants the add the content to the encyclopedia should have no trouble finding it, checking its accuracy and adding the source. Then it is no longer AI generated, but checked and correctly sourced.
I don’t see any supported statement about on what you accuse Joe Faust of, concerning the Wikipedia page of “Airborne Wind Energy” which @Rodread is referring to. I only see generalities, or references on peripheral subjects, or even those that have nothing to do with AWE.
So I am waiting for: Joe Faust’s quotes from his “Wikipedia edits” in the concerned Wikipedia page, then how they would break certain rules, with supporting evidence.
Below is the revision history of “Airborne Wind Energy” Wikipedia page. We can see that the last Joe’s contributions date from 2011. I don’t see how old contributions could invalidate a site whose subsequent contributions have been numerous.
I can see the error of my ways now…
I should not have published the AI content I did as an
AWES Ground Station wiki
Before compiling all of the relevant initial references.
The problem with trying to work backwards is the strongest links to the text are the text itself.
It becomes garbage in
The method may be better used in developing the IEA architectures documentation
I have now deleted all of the wiki ground station topics contents
We can do better on the next version
I would assume you have a folder of AWES literature that the AI looks at? But that must exceed the token limit? I don’t really know how you would do it otherwise. You can’t have it look at the internet because the papers are not open access so shouldn’t have been part of the training data, and even if they were, the references it would give you would be hallucinated so you would need to divine the correct references. You could only really use it as a brainstorming help then I think.
Did you train your own LLM? That’s what I would do. Then do a fuzzy search for the things it generated on the source texts to check if it didn’t hallucinate, and then quote or reference the source.
What you are doing is very similar to writing a chapter for a book. We already have the AWES books. Maybe you could petition for and volunteer to write a chapter for the next edition, or cooperatively do research for the chapter.
Perhaps I’ll look at the AWES page later @PierreB. Though I didn’t limit my comment to just that page. I haven’t seen evidence yet where he didn’t misunderstand how to edit Wikipedia pages, so finding that would be refreshing. His own page is ample evidence to the contrary. Can you not understand how to do something on one page and understand how to do that same thing on another page? I don’t know how that would work.
That was nice of him and was good. It was also nice that there was a forum.
There are so many methods for large language model integration in a workflow these days.
Yes, I have copies of the AWES books and open access literature which I trained an LLM on locally… But that’s going to remain a bit useless until I build a colossal computer and can inference across many more layers of network.
However, new offerings from large AI providers let you train on your own documents using their compute resources. You’re right, you very quickly hit token limits.
However there are also ways to get API calls systematically passed into RAG Retrieval augmented generation ai systems like Perplexity or Gemini
Those were the ones coming back at me with the answers being - what we had already written…
What we created was a really good SEO optimization system because instantly we were at the top of the lists for those searches
Yes, there is a risk of tautology. I saw this before on wikipedia citing some articles which were based on wikipedia… At one point, something is becoming true out of nowhere. I saw this a few times in my previous work experience where some assumptions were used to derive some values. I was asked to use these values to check the original assumption… and was blamed when I told that I had to do some modelling, check some experimental data, whereas for some a simple formula to return to the original assumption would have been enough (i have to admit that this process can help to invalidate some assumptions by absurd reasoning, but not to validate them).
Maybe that process informed the design of AWES.
This is a kite - this is how energy works - this is how a kite does energy.
A kite is how we do AWES
So 
Let’s just automate the process of funding engineering instead
Apparently we just have to create a DAO
Set a bunch of performance milestone contracts
Let the money, equity churn while the tech progresses
Darnit
I had another go at [Wiki] GS
however
An error occurred: Body is limited to 32000 characters; you entered 45803.
Yeah, waaay too many lines of generic format and no filler
However it’s been trimmed down and is ready to format at
https://forum.awesystems.info/t/wiki-ground-stations/2838/2
I don’t expect anyone has appetite to do that.
I don’t think it’s a sensible format for a wiki
It claims to be a
reference guide for engineers, technicians, operators, and other stakeholders involved in AWE system deployment and operation
OK
There is an academic AI based web service for this task already
e.g. generate a Wiki / literature review based on an academic topic
https://storm.genie.stanford.edu
Just about to test it out now
Storm Genie Stanford a929277b-fa02-4d6e-834f-1ac942bfd791.pdf (127.4 KB)
A bit vague so far and not quite so specific and nailed down…
You can definitely see where I tried to steer it with rotary relevance
There are a lot of much more relevant references
Including to this forum 
Love that the last reference is Joe Faust’s Energy Kite Systems .net 
Research journals are now accepting AI content in submitted papers with the following rules…
as long as you
Make it obvious and declare where AI has been used with a description of the method.
Only use it for text formatting works not original research.
Take responsibility… It still has to be you that checks and claims your work in the first place.
You have to make sure your presented work is credible - AI is not an author.
You can’t use it for generating peer review
The same guy on youtube has further recommendations on using AI for research
This is indeed a good reference. AWE’s situation has been stable for years. This reference therefore has staying power.
What good easily accessible search engines still exist… These I know, are there any more?
https://link.springer.com/advanced-search
https://worldwide.espacenet.com/advancedSearch?locale=en_EP
WIPO - Search International and National Patent Collections
I don’t know of a good search engine for the internet. If you had that it would be much easier to check the AI generated content.
For checking the AI, I’d probably do an advanced search like that on Springer, or similar. Or maybe I’d put all content I wanted to search in a folder and search the content of the files. I use FileLocator Pro for that for example.
And, I don’t know if this website is the right place for wikis, Energypedia and Wikipedia seem better choices, but you could just use wiki software on here if you want better wiki formatting. You’d create wiki.awesystems.info for example.
Skimming through a few articles on Energypedia though, is it an encyclopedia?
Access to Modern Energy - energypedia This section for example doesn’t have references. It seems to be someone or someones advocating for something. Who the author advocating for this is I can’t see.
Tricky to classify Springer and patent searches as “Good easily accessible”
When they represent closed IP or paywalls.
I’d have thought good easily accessible was more to do with merit and open reputation.
Thanks for the hint on FileLocator Pro, Yep I can see that’d be handy for the kind of more exacting document referencing we tried on the Wiki generation task.
What is it you are checking the AI for ? - Like generated content validity ? yeah always needs done.
Good idea - that could overcome the post size limit which excluded our initial attempt.
One problem however - I’m only a moderator - Not an Admin - The ones with the real heady powers and responsibilities
Oh goodness me that FileLocator Pro was a revealing tool thanks @Windy_Skies
So I searched an AWES resource folder for the terms “groundstation” and “Ground station”
The document with by far the most hits (167)
Modelling and Analysis of Rotary Airborne Wind Energy Systems - a Tensile Rotary Power Transmission Design
April 2021
Thesis for: PhDAdvisor: Hong Yue and Julian Feuchtwang
Authors:
Oliver Tulloch
University of Strathclyde
@Ollie knows his stuff
The AWE book 2 only had 141
| Author | Resource Type | Title | Hits |
|---|---|---|---|
| Oliver Tulloch | PhD Report | Modelling and Analysis of Rotary Airborne Wind Energy Systems - a Tensile Rotary Power Transmission Design | 167 |
| Roland Schmehl Editor | Book | Airborne Wind Energy Advances in Technology Development and Research | 141 |
| Matteo Milandri | Dissertation for Masters of Science in Engineering | Design of a Kite Controller for Airborne Wind Energy | 97 |
| Uwe Ahrens Moritz Diehl Roland Schmehl Editors | Book | Airborne Wind Energy | 95 |
| Andreas Klein-Miloslavich B.Sc., Simon Bolivar University, 2016 | A Thesis Submitted in Partial Fulfillment of the Requirements for the Degree of MASTER OF APPLIED SCIENCES | Modeling, Simulation, Hardware Development, and Testing of a Lab-Scale Airborne Wind Energy System | 68 |
| Google Kenny Jensen, Berkeley, CA (US); payon Vander Lind, Alameda, CA | United States Patent | METHODS AND SYSTEMS FOR ESTMLATING AN ORIENTATION OF A TETHERED AERAL VEHICLE RELATIVE TO WIND US 9,389,132 B1 | 64 |
| Davide Todeschini and Lorenzo Fagiano and Claudio Micheli and Aldo Cattano | Paper Milan | Control of a Rigid Wing Pumping Airborne Wind Energy System in all Operational Phases | 50 |
| Antonello Cherubini | PhD Thesis in Emerging Digital Technologies | Advances in Airborne Wind Energy and Wind Drones | 48 |
| Uwe FECHNER | Masters Thesis | A Methodology for the Design of Kite-Power Control Systems | 45 |
| Florian Bauer and Ralph M. Kennel | Research Article | Fault-Tolerant Power Electronic System for Drag Power Kites | 42 |
| Karel van Hussen, Enno Dietrich, Job Smeltink, Koen Berentsen, Manel van der Sleen, Robert Haffner, Lorenzo Fagiano | European Commission PP-05081-2016 European Commission Directorate-General for Research and Innovation | Study on Challenges in the commercialisation of airborne wind energy systems | 39 |
| Dylan Eijkelhof | Master of Science in Aerospace Engineering | Design and Optimisation Framework of a Multi-MW Airborne Wind Energy Reference System | 38 |
