I don't really agree with you. I think there were a million things to do, and in how many you can try something new and untested depends on how much buffer you

I don't really agree with you. I think there were a million things to do, and in how many you can try something new and untested depends on how much buffer you have. Some of these new things come in batches, like the automation thing they tried with Model 3 that didn't work out and was reverted. And even this was probably a calculated risk.