The reasons why I have high expectations for the Muse Spark model

Like Google’s Gemini, Meta is describing its recently announced Muse Spark model as a "natively multimodal model." If Muse Spark, like Gemini, operates on the principle that "all modalities coexist as first-class citizens in the same latent representation space from the very first layer of the model," I believe it will be able to provide answers with a deeper understanding of context based on a world-view derived from various modalities - even when I’m simply asking questions and receiving answers in text.