Microsoft unveils AI model that understands image content, solves visual puzzles | Ars Technica

This seems like the obvious next step and something I eluded to earlier in reference to Wolfram Alpha working with ChatGPT to each do their respective part best. The same way the human brain has different regions to do different tasks, why wouldn't we use different AI models in the same way.

While the media buzzes with news about large language models (LLM), some AI experts point to multimodal AI as a potential path toward general artificial intelligence, a hypothetical technology that will ostensibly be able to replace humans at any intellectual task (and any intellectual job). AGI is the stated goal of OpenAI, a key business partner of Microsoft in the AI space.

It seem like a very hard problem to transition concepts between different models. The models are neural nets themselves that do not have a readily decomposable way to describe themselves post training. How does this thing send an object around to different models? Maybe it doesn't need to? They can all just guess and then an executive model can render the final thing together?