[apwheele]

Spotcrime just allows you to sign up for email alerts to crime nearby an address of your choosing.

I am generally of the mind even if it results in negative externalities, knowledge is good. So even if it on average increases fear of crime, knowing the reported crime nearby your home is a good thing.

[Zigurd]

The quality of the data matters. Over-policed minority neighborhoods don't provide the quality of data that supports rational decision-making. LA has a lot of problems with quality of policing.

[snapcaster]

"knowledge is good" is such a naive take. Trivial example: You only have knowledge of crimes committed by immigrants but zero knowledge of crimes committed by citizens. How is that good?

[wang_li]

Having partial knowledge is good, even in your YOU'RE RACCIIISSTTTT!!!! example. Let's explain.

Consider you have to perform a task that in some way can interact with something in the environment. You have two choices of where to perform this task. In the first location there are 20 red things in the environment. In the second location there are 20 red things and 10 blue things. You know that 1 in 10 of the blue things have a negative interaction with your task. You know nothing about the red interactions with your task. You obviously choose the location with no blue things.

[snapcaster]

You're not modeling this adversarially and that's your fundamental error here. If you look at your "knowledge supplier" as an entity actively trying to deceive you and give you an incorrect world model you'll realize i'm right and you're living in a fantasy theory world

[wang_li]

No. You are trying to change the rules to a different game and then use those rules to invalidate a completely unrelated point. Having bad information doesn't make the fact that partial information is better than no information. It just means you were deceived by the information and your attempt at making decisions was sabotaged. If you are so wound up in "everything and everyone is against me" you have a mental disorder. If you are just picking and choosing who you define as adversarial based on which information disagrees with your priors you are just close-minded.

[_vertigo]

I could see it being good if it helps you estimate crimes committed by citizens. If you know where the gaps in your knowledge/data are, you can attempt to account for them. And that’s better than nothing.

I think misleading information is obviously bad, incomplete information is not necessarily misleading though.

On the other hand, it might be better to remove incomplete information if it is actively being used to mislead people.

[snapcaster]

You're not modeling this adversarially and that's your fundamental error here. If you look at your "knowledge supplier" as an entity actively trying to deceive you and give you an incorrect world model you'll realize i'm right and you're living in a fantasy theory world

[jvanderbot]

So, continue this train of thought - If only partial data is available, then no data should be available because the partial data might induce incorrect assumptions in the general populace.

Apply this to:

Vaccination / disease management

Housing availability ("if they only know of these areas, will those areas become swamped and drive up prices?")

Price of drugs / medical services, or even medical test results (how many more suicides "might" occur if someone gets a possible cancer diagnosis)

Climate change

or anything else.

I think you'll find you're quickly concentrating knowledge dissemination into a central authority who decides what is "right" and that is much more dangerous than incomplete information.

[Zigurd]

We're not talking about "partial data." We're talking about tendentious data that propagates existing known bias, produced by brutal problematic low quality policing. At the very least, people making apps based on crime location data need to acknowledge and flag such problems and inform their users of the dubiousness of LAPD and LASD data.

Surveillance tech and cop tech generally don't contribute to society because of these problems.

If you wouldn't trust RFK Jr. about vaccines, you should also be skeptical about what many PDs tell you. LAPD is just a particularly notorious example.

[jvanderbot]

> We're not talking about "partial data."

You might change the subject away from partial data, but the comment I replied to _was_ talking about partial data, and my rebuttal _is_ about partial data and the judgement of whether that partial data is worth releasing based soley on how we imaging people might react to it. Have you read "The Unthinkable"?

I wouldn't trust RFK Jr. about vaccines if I didn't trust his data. But establishing a body that was in charge of disseminating data about vaccines is in high likelihood going to be taken over by RFK Jr types. Such a body shouldn't exist. Such a body would write "Turtles all the way down."

> "knowledge is good" is such a naive take. Trivial example: You only have knowledge of crimes committed by immigrants but zero knowledge of crimes committed by citizens. How is that good?

That counter-logic is so fundamentally flawed b/c it rests exclusively on the prejudgement of others and prediction of their use of the data while "I", the good thinker, can determine that it is bad for "them" to have access to this data. That is just a very bad way to think and is precisely what RFK-types do all the time.

[throawayonthe]

> even if it results in negative externalities, knowledge is good.

i can respect[0] that, perhaps the data should be available in some form; but would regardless call 'crime alert emails' a disservice.

[0]i'm skeptical of LAPD producing knowledge in this area, others have noted overpolicing etc. also this is tracking police activity/knowledge, not crimes directly (and tbf it's impossible for it to complete even with best effort)