I'd say research different stacks before you start the prototype, pick the best one from your initial perspective, and if it's not working then branch to an alternative and see if it's better for your use case. If you don't do this, then you might end up ramming a square peg through a round hole and will need to rewrite later.

Basically, stick to your core competencies, but don't be afraid to try new things that will clearly (hopefully) improve your product. Fail hard and fast instead of dragging it out 6 months :)

This isn't about learning a framework at the expense of the product, it's about building something new that you might not have experience creating solutions for yet.