I think all he way trying to say that there are ways that a LIDAR gun could give a false reading, and there's no current incentive for law enforcement/manufacturers to ensure accuracy. In fact there's an incentive for both the manufacturer to emphasise accuracy based on successful prosecutions, and for police to emphasise accuracy to prevent large-scale contesting of tickets.

They are used for law-enforcement, but presumably aren't available for public scrutiny, especially around 'proprietary' algorithms. As mentioned, they can get firmware updates that change behaviour, yet that doesn't make it out into "does this mean it was mis-behaving before, therefore I should dismiss tickets issued in situation X".

It's similar in some ways to vote counting machines. If they are used in public elections, then the inner workings should be available for introspection to ensure correctness.

He doesn't have to prove his innocence, he just has to convince a judge that there is reason to doubt that the LIDAR accurately reported his speed.

It's not even half science, it's just a big slide deck of FUD padded with a healthy amount of entitlement.