Are there policies in place for when invasive surveillance technology such as stingrays can be used? Is there any oversight or accountability or is this entirely handled within the police department and the public is expected to trust them to police themselves?

As Mountain View leads the way, can the thought police be far behind?

Surveillance technology may be a slippery slope, but if you're not doing anything wrong then it shouldn't be a problem. Most of us are good citizens who lead a vanilla life. My life is enjoyable for myself with a family, a dog, a job, vacations, etc., but would most likely be boring to anyone else. Most people have better things to do than watch me. Except the bad guys. I would rather give MVPD the ability and tools they need to keep up with the bad guys. I got stuff to hide from the bad guys, I don't want them looking into my life.

I used to work for a fairly major predictive policing startup. There are a lot of issues with the technology and it's not a good investment at this point in time.

- First: It's not big data. Even a huge city's crime data is pretty small. The less data you have, the more suspicious the results are from a model. Mountain View is a small, safe city. Send your patrols through Rex Manor and Monta Loma / South Rengstorff. And that Bayshore thing :). G'day.

- Police departments keep sloppy crime data. It leads to a lot of issues which lose the officer's faith in the product. They're already very skeptical as it is. Police chiefs burn a lot of political capital pushing the tools.

- You'd be shocked at the security bugs I fixed in my first 2 months there. You're giving these companies your crime data, down to the street address and event descriptions. Public crime data is anonymized to at least the city block level, sans the free form data. Imagine dealing with that data breach?

- Bias: These tools are marketed as reducing bias, which is a great goal. But you know who's less likely to get the police involved? Minorities. Predictive policing simply disenfranchises differently, because it only works using crime reports.

- The statistical models are based on white papers with minimal review. It's then applied to all crime types uniformly. Do thefts, robberies and assaults have identical patterns? Just kidding: officers already know where assaults happen (at the night club). Telling them where to go look for that on a map will get a well-deserved eye-roll.

- It can be useful for property crime / car break-ins / etc. But the Bay Area has changed so dramatically, crime data from 5-years ago likely looks much different from today. And again: officers know which areas are prone to having issues. Predictive policing can enhance accuracy, but it simply depends.

- When you notice issues, they'll make some thumb in the air changes for you. Now your results are based on a deviation from the original white paper. Nobody knows what they're doing. And yes, salesmen sell things. You won't know what's up.