SurfaceUI brings a lot of useful features to Phoenix Liveview. The technical risk is low due to this project green field, so everything is new.In my personal experiencie using SurfaceUI helps people that comes from Javascript frontends such as React.js and quickly learn it.
Tackling the shortest time to market does not depends on the technilogy used (assuming Elixir is chosen), depends more on the stakeholder priorities and knowledge of the business domain.
If you have a good and deep knowledge of the business domain, then chosing a new technology stack is doable. If you are learning both the domain and technology stack then may be a risky move.
At the end of the day it only depends on who will participate in creating the project and the available time and resources for the project. But in my opinion chosing Elixir is a great choice.
True, ideally if you know where to look for great elixir engineers, you will get a much better value for you buck compared to most other stacks on the market. The only exception I would say is if another ecosystem provides most of what is needed as already implemented functionality, then you would never beat that when building it from scratch.
Better how? These are 2 completely different tools aimed to solve different problems. If you are using live_react, then that arguably is no longer a liveview project, but something that leans towards having a classical 2 stack web application.
LiveReact and LiveSvelte are good choices. But they must download the client js code. So for certain use cases, for example when you need to have less than 1 mb of JS downloadss to improve the Lighthouse metrics, then Liveview and Surface UI can be a better choice.
This is real example, once of my previous projects they migrated from Next.js to SurfaceUI due to reaching better Lighthouse metrics and less JS bloat.