Unfortunately there isn't much overlap between danger addicts and people with the skills, foresight, and pragmatism necessary to make a mars colony work.
That's not to say there are no good candidates for colonists, they just aren't like the guy that you're talking about. They are more like our current U.S. astronauts for whom getting a Ph.D. is just something they do along the way because they have a ridiculous amount of discipline and aptitude. There are many more qualified applicants than there are spots for them to fill, and there would probably be even more if the space program were more real. Interplanetary viability of the species is something that even very pragmatic and intelligent people are willing to risk their lives for.
I think the concern is that the Tesla system may not have even identified the semi trailer across the lane as an obstacle at all because it is not clear that the Tesla's brakes were applied prior to the collision. Any defensive human driver would have taken note of the truck in the turn lane before it even started its turn and prepared to take action if it turned unexpectedly, even if they were unable to stop or swerve in the end. Tesla's system was definitely unaware of the truck until it entered the lane, and was most likely incapable of determining the actual structure and size of the obstacle.
But the big problem everyone is worried about is right there in the first sentence of your link. The article calls it a "self-driving car" which even Tesla (except for their CEO in promotion mode) takes pains to say it is NOT. Treating a Tesla on autopilot as a self-driving car is extremely dangerous to the driver and all motorists around them. The sensor package and software is not designed for that usage.
Of course they use the more direct size measurements when they are available. It sounds like those 102 are the control sample, in which they have both WISE measurements (IR light) and a more direct measurement of size. They don't need a thermal model for those objects. And when just reporting object size they should report the size inferred from previous radar or occultation if that is available.
It sounds to me like this guy developed a thermal model that doesn't work for objects for which the size is known from other methods. The obvious conclusion is that the thermal model is not good, but apparently he has instead decided that it's a vast conspiracy to hide the true sizes of asteroids. I'm sure the WISE and NEOWISE teams would be happy for someone to develop an improved thermal model if theirs is only accurate to 15% for objects with known size. However, this appears to be a case of this crazy outsider trying to push a worse thermal model through outrageous and insulting press releases. Not exactly how science is supposed to work.
e-credibility: the non-guaranteeable likelihood that the electronic data you're seeing is genuine rather than somebody's made-up crap. - Karl Lehenbauer