Perhaps you're thinking of the Wikimedia Foundation.
There is plenty of space there for more volunteer editors to verify content, and likewise, WMF operates its own cloud platform where developers are automating tools that do maintenance and transformation on the human-contributed content.
Then, there is Wikidata, a machine-readable Wiki. Many other projects draw data from here, so that it can be localized and presented appropriately. Yet, its UI and SPARQL language are accessible to ordinary users, so have fun verifying the content there, too!
I don't think you understand what I meant by human verified, but I used a very vague term to express what I meant, I meant proving that some input or data that comes from a user was generated by a human (whatever we define that to mean) rather than an LLM or multi-modal image/video/audio model output.
Perhaps you're thinking of the Wikimedia Foundation.
There is plenty of space there for more volunteer editors to verify content, and likewise, WMF operates its own cloud platform where developers are automating tools that do maintenance and transformation on the human-contributed content.
Then, there is Wikidata, a machine-readable Wiki. Many other projects draw data from here, so that it can be localized and presented appropriately. Yet, its UI and SPARQL language are accessible to ordinary users, so have fun verifying the content there, too!
I don't think you understand what I meant by human verified, but I used a very vague term to express what I meant, I meant proving that some input or data that comes from a user was generated by a human (whatever we define that to mean) rather than an LLM or multi-modal image/video/audio model output.
You mean, like, the Book of Kells and a nice stage production of Lysistrata, vs. a PowerPoint and Star Wars? Interesting.
This issue in terms of cost is that if you want this to be truly human-verified for real, you're gonna have to dip into the real world.
Revival of the curators.