1. 26

Sensationalist title, but it made me consider how civic hacking could benefit communities in non-obvious ways.


  2. 10

    The $86 million number seemed really high, so I did some research to get more context. Here’s what I found:

    • The pilot was rolled out in 2014, and the report to expand it was in 2015. It was just an estimate and was never rolled out. No idea how good image recognition libraries were in 2015. Earliest prototype was 2012, where I’m confident saying off-the-shelf image recognition was pretty poor.
    • BlueNet also includes the cost for cameras, drug testing equipment, in car video, laser scanners, etc, not just the image recognition software.
    • I haven’t been able to find the original Deloitte paper that proposes an 86 million budget. Is it just for BlueNet, or does it include other expenditures? Even if just BlueNet, is that just the equipment or also service and training?
    1. 2

      Well, in case it includes service and/or training, depending on the kind of equipment provided, the budget may seem more reasonable. Thanks for the information!

    2. 2

      The solution proposed in the post is interesting and functional, but the government budget seems outrageous. $86 million? The author found an open-source Automatic License Plate Recognition library with a little research, which is something I assume the government did not. I am no expert in image recognition hardware, but that budget (for 220 cars) seems unacceptable to me.