Maybe a better analogy -- when I'm on call and get paged, my first response is not to immediately restart the servers. The immediate response should be to assess the situation, not jump straight to the most extreme approach.

Now you may quite reasonably point out that a page on a software system is seldom of the same weight as an emergency call to the police (true in my case -- money, not lives, have always been at stake in the systems I work on), but this ignores the reality that the "swarming with guns" approach _also_ can, and has, cost lives. So if anything, the higher stakes involved for police involvement is an even _stronger_ argument to take the extra time to properly assess the proportional response.

I guess I'm not disagreeing with this point of view.

Perhaps my conclusion would be, rather, that a SWAT team coming in to arrest someone shouldn't also lead to a heart attack situation. Maybe I'm being naive, and probably most people SWAT teams come to apprehend are super-dangerous, but it feels to me it shouldn't be a consideration.

In the age of the AI profiling a face in the crowd, it's very difficult to believe that some system to look up the caller ID cannot be brought to work.

AI is still dog shit for 99.9% of applications, we're not playing chess here. It would add an insane amount of complexity and points of failure for an edge case. If it even fucks up once it would need to be retired.

I know HN loves to think raw tech will save the world but come on... this is a societal problem, not a tech problem

Even if caller id had been used in this situation, he could have easily spoofed the number to appear to be coming from the neighbors house since they would have already had all the information necessary to do so. The Caller ID system you're thinking of simply does not exist. The current system is much to easy too spoof. Anyone with $15 and a VoIP number can do it.

Notably, the caller was identified, it seems. The problem wasn't identifying the caller, rather taking his accusation at face value.