Guys, I have an idea and I need your help.
The problem:
Currently, there is a lack of a cross-interaction system between different GPT-s add-ons for interesting projects.
For example:
On a livestock farm, a person might notice certain symptoms of illness or strange behavior in an animal, or even a group of animals. They could then launch a complex of different GPT-s that would, from video or photo + saund:
Identify probable forms of disease.
Suggest actions the person should take immediately before a professional veterinarian or animal specialist arrives.
Requirements:
The response model GPTs hallucination should be strictly within the framework of documents and rules from add-on’s.
It should only record specific behavioral patterns, photos, or video analysis (especially the latter).
It should be a strictly regulated multi-agent system, focused on assessing situations from a list of probabilities.
If there are no answer options, it should request an extended version of the answer with a broad match outside the context and go beyond the standard analysis branches.
Additional notes:
The system should be able to handle multiple concurrent requests form few sources.
It should be scalable to handle large numbers of animals and farms.
The system should be user-friendly and easy to use.
I believe this system has the potential to revolutionize the livestock industry and improve the welfare of animals.
And I’m also interested in the possibility of flexible connection of additional modules if necessary, for example, the situation that has developed on a farm is already incorrigible and the farm owner needs, based on this information, to create a block for preparing legal documents for investors and for insurance companies, and so on, and it would be possible right inside this a tool to connect an additional source of legal advice in this direction, and so on.
If you are interested - say something.