You've nailed it. Two FCCs can't meaningfully vote on two AoA vanes, no matter how everything is configured. And the task of reconfiguration to get anywhere near an effective solution (maybe toss in a third, synthetic, AoA input) . . . well, how can that not mean an enormous testing and certification burden, followed by significant new training requirements?
I wrote in March this year on the Lionair thread:
Add a third AOA with a different software programme for both and you have solved the issue of how to make a poor aircraft design at least flyable.
I can see a lot of companies cancelling their options on this flying pig.