What if marijuana improved your driving performance?

I ran across this intriguing bit from scientists in Heidelberg, Germany. It’s not a proper study — merely an anecdotal incident based on one individual, but it’s still fascinating.
Turns out that they discovered a 28 year old man with ADHD who had a history of ADHD-related driving problems, improved dramatically after smoking pot. The scientists

conclude from their investigation that “it has to be taken into account that in persons with ADHD THC may have atypical and even performance-enhancing effects.”

It’s not enough to prove anything, but it does point out the fact that different people are affected by drugs in different ways.
It reminds me of a good friend of mine in college many years ago. Brilliant man. Theatre major. Unusual brain. As an actor, he never said the same lines exactly the same way twice, yet they were perfectly fitting and he could actually ad-lib Shakespeare in iambic pentameter. We once lost a performance space for a mainstage production the day before opening night and in one day he re-lit the show in another theatre with a totally different configuration and ran the lights for the show by feel without a light plot. He did all of these things while under the influence of marijuana. If you ever ran into him when he was not, he was incoherent.
Everybody’s different. I had another friend in High School who would pass out for 12 hours if she took just half of a simple cold capsule.
This points out the potential problem with determining driving capability by what’s in a person’s blood. It may be better than no method at all, but it’s not an accurate measure of actual impairment.
So what would be?
Here’s my idea, but I need someone out there to invent/develop it (technologically, it seems that we’re pretty close if not already there). Modify regular police cars so that you can, with a switch, change the driving controls to become a full-fledged driving simulator with the windshield turning into a video screen. The simulator would test reflexes and measure driving ability.
Seems to me that this would be a much more accurate measure of impairment than breathalyzers, urine samples, or blood tests. Wouldn’t it?