They're not insane, just speculative. Tesla has a lot of potentially very lucrative new technologies in development, that could completely disrupt a lot of markets if brought to fruition. For instance, if they actually crack full self driving, today's valuation will look very conservative. But those are all big ifs, so it depends on how much you believe any of that stuff will actually happen. And Musk's behavior has certainly not inspired confidence in recent years.
I'll be highly suspect of self driving for anything but say rural highways for a very long time even under the best of circumstances.
Given Musk's disastrous takeover of Twitter which highlighted how little impulse control or forethought he has, I'd sooner eat my own stool than get in one of his self driving cars. Even if Tesla engineers delivered a viable product I have no doubt musk would somehow fundamentally fuck it up before releasing it to market
I'm not convinced the guy who repeatedly lied about full self driving and recently said "lol LIDAR is useless" without expanding on why, save for trying to peddle some AI bullshit along the way, is to be trusted.