Our theory was that if that team is capable of getting across the finish line, even if it doesn't sell well, they'll now know what it feels like to finish. And they will get feedback from the sales and the public as to what they did and didn't do right and so that team will be more strong, the individuals will be stronger, and we'll be able to circle back and do a more competent game next time.
Good point, but also back then games weren't that expensive and long time to make than ever before. The AAA industry first needs to get the development cost under control, and I don't mean by using Ai to make cheap games faster and cheaper than before.