That’s already capable using multiple methods and multiple AI GUI’s . A true open sourced multimodal model has just been released for beta access. It’s not great but since it’s open sourced it will be.
Can it search the internet (like Google) and summarize the results and then, keeping the context, refine the search? I also want video generation someday.
If all this does is handle documents, maybe it's not worth dealing with unless you need it for business. There's supposed to be an open source Perplexity type software that's better than Perplexity.