Revise the graphics used in README for vizro-ai and vizro-core so they render better in PyPI

[stichbury]

This is a good issue for someone with design knowledge or who knows how to optimally set out graphics to work on different GitHub themes and on PyPI. It touches docs but doesn't change any wording. No knowledge of Vizro is required.

https://pypi.org/project/vizro/

and https://pypi.org/project/vizro-ai/

both look a bit rubbish right now because of the graphics choice in our package README files.

We need to revise these to work better, or make separate choices for PyPI. This is a potential good first issue for a contributor to Vizro if they have previous experience with PyPI page submissions. The remit is that the graphic/GIF should work in light/dark mode and on GitHub plus PyPI.