Comment by alain94040

Would QA solve that problem? As in, "we replayed 1.4 trillion accident scenarios and our AI behaved correctly on all of them". Safety in numbers.