Well as usual this is implying that all of the money sunk somewhere but in order to get the wheels rolling in the govermental branches you need to actually have multiple commities etc.<p>Of course you can replicate your prototype for one day of work. There are even muliple libraries which you can just drop in place, but none of them have gone trhough law approval (sorry i dont know how to call it better in english).<p>At least in europe most of that money would have gone to all the bodies that are testing the solution and are actively trying to poke holes into it. Therefore yes consultants and lawyers as always. But the slight difference is that your system will actually adhere to law.<p>Or differently put: its not so much about recognizing the plate but more about what happens after, before and inbetween with the data.<p>The cyberpunkesque solution for sure just sends your plate to a central server that will log any way it came into contact with that plate, but a data protective view of this will download only the license plates that it hunts for and will beep when its in the local database.<p>I am aware that this also has a local database but I do not mean the locality of the data but rather the number of plates you need to actually scan for at any given time. Just like how you would like the top 50 on the most wanted list found, but not to be interrogated on the street if you know someone you never heard of.
I do a lot of work as a contractor to the public sector and I can tell you that very little of the $86M will be spent on software development. The vast majority will go on hiring a massively bloated team of non-technical people to go through the motions of whatever delivery methodology they use, producing a huge amount of documentation and getting various people to sign it all off. Assuming it’s outsourced the vendor will have its own team producing similar quantities of paperwork. The ratio of developers to paper-pushers will be lower than 1:20.