Teams test their designs and set-ups using computer simulators and have done for years, but I've often wondered how accurate these are and how much they gain from them. A lot of the effects that they need to test for would come from multiple precise physical interactions and even a tiny initial error or approximation could lead to completely inadequate results. Apparently none of the teams predicted the porpoising from their simulations (I read that anyway), and also it even seems that quite often the teams have absolutely no idea whether the hard tyres are going to be any good in a given race until they actually try them out on race day, and it seems surprising that their simulations can't give them some idea. And I'd say that was one of the more basic things, so if they can't do that, what are they useful for?
And also, it seems a bit wasteful getting drivers to test out things on the simulator. Wouldn't it be better to get an AI to do laps at e.g. 1000x speed so that what would be a day's work for a driver could be done in moments?