There is a lot of overlap here because programming is also formal logic at the end of the day. The whole idea of having a reasoning engine that can operate on symbolic is a really important step forward in my opinion. This will allow models to actually explain the steps they use to arrive at a solution, and would also open up potential to training through instruction. If a model gives you a wrong answer, you could explain to it which step in its reasoning is incorrect and it could fix its logic going forward. This would allow for similar style instruction we do with humans where we can explain to students how to solve problems.