Legal in AI

This all started when I was starting to index content for an engineering manager assistant that would help our original thesis for Path.dev, in the hope to raise funds cause we were running out of money. Which we did, story for another time.

I started to design a system that would scrape and index content like a search engine from respected blogs on engineering that I respect.

But this then raised an important question. Is that allowed? Should I be able to index knowledge like that now? It's not search, it'll be a model that will almost steal that knowledge.

I was trying to put up an indexer.ai file on my blog (link to twitter) that would define to the models out there that they can't index my ramblings.

But really that isn't the right thing to do, I'd like to give my insights and knowledge away but make revenue on it. So how would we do that in AI.

I believe that we can do this if we model the AI engine differently. If we return to search indexes and contextual predicates to structure knowledge, think of it as a tree. Then we'd be able to own different parts of the tree a little more easier. Allowing for folks to know when their knowledge is used, extended, collaborated with other folks, etc. I'll leave my thoughts on how we could build something like this building on search based models rather than neural networks for another time. This type of model will open up many other doors, such as music and other creative applications.

I haven't researched it, but I believe that OpenAI and other companies out there are indexing open sources on the web (still working on my offline/online approach) and now generating revneue from that knowledge. Is that fair? Do tokens map to the source of knowledge and they build up a financial line to give to them?

Time will need to be invested in this area to truly work through it all, but as a scientist we must and should consider how that will work in the design of the implementation. Otherwise it could make or break a company as legal catches up.

The approach I outline here is just the start and there are many holes in it. But I'm sure it's one of many out there that would work for our time today.