I just want to say that I fully trust dirks experience and judgement.
From what I understand from following this thread, he made extensive and thorough R&D to come up with the project as it is.

Imho, making substantial changes at this stage is not such a good idea.

Changing the position of the magnets may appear simple but, e.g., adding new components (like glue, epoxy, ...) and the combinations of the components (cover <-> glue <-> magnet) imho add significant error potential and would require to repeat quite a lot of R&D or even longer additional testing.

Similarly, injection molding may sound cool. But there are many uncertainties as well, e.g., the molding process may vary, the molding tools may not fit on the first try, or the quality of the material may not match the expectations.
Furthermore, the costs for iterations in molding, e.g., when the tools have to be changed are in the magnitude of thousands of euros compared to about tens of euros for 3D-printed parts. Plus the advantage dirk mentioned that the 3D-printed parts can be re-ordered by anyone who has access to the respective files.

Last but not least, the backers at kickstarter had a very good overview to know what they subscribe to. Changing this afterwards could be problematic and may lead to frustration and disagreement.