Daniel sent us this one — he's been thinking about something we touched on a while back: that moment when you see someone down on the sidewalk and freeze, not because you don't care, but because you're running the calculus of whether calling nine-one-one actually helps this person or just drops them into a system that wasn't built for what they're actually dealing with. He's asking about non-emergency response lines — the numbers that exist between "this is fine" and "someone is dying" — and whether there's a safer path for civilians to interface with authorities without the full lights-and-sirens escalation. It's a good question, because most people don't know these exist.
The one everyone points to is CAHOOTS in Eugene, Oregon — Crisis Assistance Helping Out On The Streets. It's been operating since nineteen eighty-nine. White Bird Clinic runs it. They dispatch teams of a medic and a crisis worker, unarmed, no police uniform, in a van that very deliberately does not look like a law enforcement vehicle. They handle mental health crises, substance-related calls, welfare checks, dispute mediation. And here's the number that matters: they respond to roughly seventeen percent of all nine-one-one calls in the Eugene-Springfield area that would otherwise go to police.
Seventeen percent is not a pilot program. That's a load-bearing piece of the city's emergency infrastructure.
The cost difference is stark. CAHOOTS operates on about two point one million dollars annually. The Eugene Police Department's budget is around sixty million. So for a tiny fraction of the police budget, they're absorbing nearly a fifth of the call volume that police would otherwise have to handle — and doing it with better outcomes, because they're not showing up as armed officers to a mental health crisis.
Which is the core problem Daniel's getting at. The person on the sidewalk might not need a paramedic and they definitely don't need someone with a badge and a gun showing up. What they need is someone who knows how to talk to someone in psychosis, or someone who's coming down from fentanyl and disoriented, or someone who's just been sleeping rough for three weeks and needs a warm place and a sandwich more than they need a hospital bed.
That's the distinction CAHOOTS makes explicit. They classify calls into tiers. Tier one is the person is a danger to themselves or others, weapons are involved, there's violent behavior — police handle that, sometimes with CAHOOTS as backup. Tier two is the gray zone, where CAHOOTS goes but police are on standby nearby. Tier three — and this is the bulk of their calls — is purely social-service response. No police involvement at all. The person is distressed, intoxicated, disoriented, having a mental health episode, or just needs a ride to a shelter. Two people in hoodies knock on the van door and say, hey, what's going on, how can we help.
The hoodies matter. The absence of a uniform matters. If you're someone who's had negative encounters with law enforcement, or you're undocumented, or you're in a paranoid state, a badge doesn't read as safety — it reads as threat. And that means you might refuse help you'd otherwise accept.
There's a study out of Denver that looked at their STAR program — Support Team Assisted Response, which launched in twenty twenty, modeled partly on CAHOOTS. Over its first eighteen months, STAR teams responded to more than two thousand calls. Zero injuries to responders or civilians. The 911 dispatchers were trained to triage calls and route the non-criminal ones to STAR. The Denver Police Department reported a thirty-four percent reduction in low-level calls in the districts where STAR operated.
Zero injuries over two thousand calls is not a fluke. That's a design feature. When the responder isn't perceived as a threat, the interaction doesn't escalate.
This connects to something Daniel mentioned about the hesitation to call. People do this internal triage — is this really an emergency, am I going to get someone in trouble, am I wasting resources. That instinct is rational, but it's also a product of the fact that for decades, the only number we gave people was nine-one-one. One button for everything from a heart attack to a guy yelling at a lamppost.
Which is like having one tool in your kitchen and it's a flamethrower. Yes, it'll handle some things beautifully. Other things, you're going to regret even trying.
That's where three-one-one comes in. Most people know three-one-one as the number you call to report a pothole or a noise complaint. But a lot of cities have expanded three-one-one into a gateway for non-emergency social service requests. New York City's three-one-one system handles something like thirty million calls a year. You can request a wellness check, report a person in distress, ask for homeless outreach. It's not an immediate response line — the turnaround might be hours, not minutes — but it creates a record and routes the request to the right agency.
The problem with three-one-one is the branding. Everyone knows nine-one-one. Three-one-one sounds like nine-one-one's less attractive cousin who couldn't get into the same college. People don't memorize it. They don't reach for it in a moment of uncertainty. And the response time is all over the place depending on the city.
Yeah, three-one-one is inconsistent. Some cities have excellent systems with real dispatch capability. Others have what is essentially a voicemail box that someone checks on Tuesday. There's no federal standard. It's entirely municipal. So Daniel's question about "a safer path for civilians to interface with authorities" — the answer is yes, the paths exist, but they're fragmented and under-publicized.
Let's talk about the fragmentation, because I think this is where most coverage gets it wrong. The story isn't "Eugene has a cool van program." The story is that we have no national framework for non-police crisis response. Every city is improvising.
Denver has STAR. San Francisco has the Street Crisis Response Team. Albuquerque has the Community Safety Department. Los Angeles has CIRCLE. Portland has Portland Street Response. Olympia has the Crisis Response Unit. Austin has the Expanded Mobile Crisis Outreach Team. They all have different staffing models, different hours of operation, different dispatch protocols, different relationships with police departments. Some are twenty-four-seven, some are nine to five. Some handle substance-related calls, some don't. Some can transport people to shelters, some can only call a ride.
The patchwork is the point. This is the United States doing what it does with social services — letting a thousand flowers bloom and then not watering any of them.
The funding is a real issue. Most of these programs started with one-time grants or pilot funding. CAHOOTS is unusual because it's been baked into the city's budget for decades. But a lot of the newer programs are perpetually six months from running out of money. Portland Street Response, for example, launched in twenty twenty-one with a lot of fanfare and then got caught in budget fights. The program expanded, then contracted, then there were debates about whether it should operate at night — which is when a lot of these calls actually happen.
Cut the night hours. Because people only have mental health crises during business hours. Very courteous of them.
The scheduling piece is genuinely one of the biggest operational questions. If you're not twenty-four-seven, you're not a crisis response system — you're a daytime concierge service. And a lot of these programs started with limited hours because of staffing and funding constraints, and then got stuck there.
Let's go back to the bystander for a minute, because I think Daniel's framing was really about that person standing on the sidewalk looking at someone who's down. What do you actually do? What numbers do you call?
It depends where you are, which is part of the problem. But there are some general principles. First, if the person is unresponsive, not breathing, bleeding heavily, or clearly in medical distress — call nine-one-one. That's still the right call for a medical emergency. The hesitation Daniel described is about the ambiguous cases. The person who's breathing but not responding. The person who's conscious but clearly altered — talking to someone who isn't there, swaying, disoriented. The person who's asleep in a doorway and you can't tell if it's sleep or something worse.
The doorway sleep is the classic. You've walked past that person a hundred times. This time something looks different. Or maybe nothing looks different and you just can't shake the feeling. Either way, you don't want to be the person who walked past someone who needed help, and you also don't want to be the person who called an ambulance for someone who just wanted to be left alone.
That's where the non-emergency options become valuable. If your city has a mobile crisis team, that's ideal. If it doesn't, the non-emergency police line is a real option that people don't think about. Every police department has one. You call it, you say, I'm at this intersection, there's a person on the ground, they're breathing, I can't tell if they need help, can someone do a welfare check. The dispatcher will decide whether to send an officer, route it to a crisis team, or tell you it's not something they can respond to.
That last outcome is important to mention. Sometimes the answer is "we can't do anything about that." And that's not callousness — it's a reflection of the fact that being intoxicated in public, or being homeless, or acting strangely, is not necessarily a crime or a medical emergency. The system doesn't have a great answer for "this person seems not okay but not in a way that triggers any of our protocols.
Which is why the street outreach teams exist in the first place. They fill the gap between "call an ambulance" and "do nothing." In a lot of cities, these teams aren't dispatched through nine-one-one at all — they do proactive outreach. They walk routes, they check on known encampments, they build relationships with people over time. The call-in version is reactive. The proactive version is where you actually get people connected to services before a crisis happens.
The relationship piece is underrated. If the same outreach worker checks on the same person three times a week for a month, on that fourth week when the person is actually ready to go to a shelter or talk to a caseworker, there's trust. Versus a police officer they've never met showing up at three in the morning with lights flashing.
There's data on this from the Denver STAR program. One of their metrics was how often the same person appeared in multiple calls. They found that a significant percentage of calls involved repeat contacts — people cycling through crisis, stabilization, discharge, crisis again. The STAR teams were able to break that cycle for some people by connecting them to ongoing services rather than just resolving the immediate crisis and leaving. The police-only model tends to resolve the moment — the person is moved along, taken to a hospital, or arrested — but doesn't touch the underlying situation.
Resolving the moment versus addressing the situation. That's the whole thing in one phrase.
It's worth saying that police officers, a lot of them, will tell you they don't want to be doing this work. They're not trained for it. A typical police academy in the United States devotes about a hundred and ten hours to firearms training and eight hours to mental health crisis intervention. That's not a joke. That's the actual ratio in a lot of jurisdictions. So you have officers showing up to a mental health call with a fraction of the training that a crisis counselor gets, and then we're surprised when the outcomes aren't great.
You can't even learn to bake bread properly in eight hours.
Some departments have invested heavily in Crisis Intervention Team training — forty hours, which is better — but it's still a fraction of the force, and it's still police officers doing the work. The alternative responder model says: maybe the right person for this job isn't a police officer at all.
Let's talk about Good Samaritan laws, because Daniel mentioned the fear of "creating problems for the person." If someone is overdosing and you call nine-one-one, are you putting them at risk of arrest?
This has been a big policy push over the last decade. Most states now have some form of a nine-one-one Good Samaritan law that provides limited immunity from drug possession charges if you call to report an overdose. The idea is that the person who's overdosing and the person who calls both get protection from prosecution for simple possession. The laws vary wildly by state. Some cover only the caller, some cover the victim, some cover both. Some provide immunity for paraphernalia, some don't. Some require the caller to stay on the scene and cooperate, and if you leave, the immunity vanishes.
The answer is "probably, but check your state, and also don't leave, and also this only covers possession, not distribution, and also the immunity might not apply if you have priors." That's a lot of fine print for someone who's watching a person turn blue.
And the research suggests that a lot of people don't know the laws exist, or don't trust them even when they do know. A study out of Washington State found that among people who use opioids, awareness of the Good Samaritan law was relatively high — about sixty-two percent knew about it — but trust in the law was much lower. People were still afraid that calling nine-one-one would result in arrest, regardless of what the law said. And in some cases, that fear was justified, because police officers don't always know the law or apply it consistently.
You have a law designed to encourage calling, but the people who need to call don't trust the system enough to believe the law will protect them. The law on paper and the law as experienced are two different things.
This is where the non-police response model has an advantage that doesn't get talked about enough. If the person responding isn't law enforcement, the question of arrest never arises. You don't need a Good Samaritan law because there's no threat of prosecution in the first place. The crisis worker isn't running your name for warrants. They're not checking if you have an open container. They're there to stabilize the situation and connect you to care. That changes the calculus for the bystander too — you're not calling the police on someone, you're calling a service for someone.
Which is a much easier call to make. Emotionally, morally, practically.
There's an interesting model in Canada that takes this even further. Toronto's Community Crisis Service launched in twenty twenty-two, and it's completely police-free. The teams are dispatched through two-one-one, not nine-one-one. They respond to mental health crisis calls in several pilot neighborhoods, and they've handled thousands of calls with no police involvement. The teams are composed of a crisis counselor and a peer support worker — someone with lived experience of mental health challenges or homelessness. That peer component is significant.
The peer worker changes the dynamic completely. You're not being helped by someone who studied your situation in a textbook. You're being helped by someone who has been where you are and got through it. That's a different kind of credibility.
The early data from Toronto is promising. The teams are resolving something like eighty percent of calls on scene without transport to a hospital. Hospital transport is the default outcome of a police mental health call, and it's expensive and often not what the person needs. The Toronto teams are doing de-escalation, safety planning, and connection to community resources — and most of the time, that's enough.
Let's talk about the hospital default, because that's a huge piece of this. An ambulance ride and an ER visit for someone who's intoxicated or in a mental health crisis is thousands of dollars. If they're uninsured, that's a bill they can't pay. If they're on Medicaid, it's a cost the public bears. And the ER is not equipped to handle these cases — they stabilize and discharge, and the person is back on the street in six hours with no follow-up. It's the most expensive possible version of doing nothing.
There was a study in the Journal of the American Medical Association a few years back that looked at the cost of police response versus alternative response for mental health calls. The alternative response was something like a quarter of the cost per incident, mostly because it avoided unnecessary hospital transports and arrests. And that's not even counting the downstream savings — fewer court cases, fewer jail bookings, fewer repeat nine-one-one calls.
It's cheaper, it produces better outcomes, and it doesn't require armed officers to do work they weren't trained for. Why isn't this everywhere?
A few reasons. These programs are almost always grant-funded at first, and when the grant runs out, the city has to decide whether to absorb the cost into the general budget. That's a political fight every time. Two, police unions. In some cities, unions have pushed back against alternative response programs because they see them as taking funding or jobs away from police departments. Cities are nervous about sending unarmed civilians into potentially volatile situations. Even though the data shows very low injury rates, the fear of a single high-profile incident is a real deterrent.
The liability concern is interesting because it's asymmetric. When a police shooting happens, the city pays a settlement and moves on. That's a known cost of doing business. But if an unarmed crisis worker gets hurt, the question becomes "why didn't you send police?" The political risk is higher for the new thing than for the status quo, even if the status quo produces worse outcomes overall.
The status quo has the advantage of being the status quo. Nobody gets fired for sending police to a nine-one-one call. Sending a crisis team instead requires someone to make a judgment call, and if that judgment call goes wrong, there's a news story and a press conference and someone's job on the line.
Yet the programs keep spreading. Denver, Albuquerque, San Francisco, Portland, Austin, Olympia, Toronto, Eugene — these are very different cities with very different politics. Something is working.
The thing that's working is the basic insight that not every problem is a crime problem. A person in a mental health crisis is not committing a crime. A person who's intoxicated on the sidewalk is not necessarily committing a crime. A person who's homeless and disoriented is not committing a crime. But for decades, the only twenty-four-seven response infrastructure we had was law enforcement. So everything became a law enforcement issue by default.
The default is the enemy here. Nine-one-one is the default. Police response is the default. Hospital transport is the default. And defaults are sticky — they persist long after the rationale for them has evaporated, because changing a default requires building an entirely new system alongside the old one and convincing people to use it.
Which brings us back to Daniel's point about people not knowing these alternatives exist. Public awareness is a huge barrier. If you ask a random person on the street what number to call for a mental health crisis, they'll say nine-one-one. They won't say three-one-one or two-one-one or the local crisis line. And even if they know about the crisis line, they might not have the number saved. In a moment of stress, people reach for the number they've known since kindergarten.
Nine-eight-eight is supposed to change that, at least for mental health. The national suicide and crisis lifeline switched to nine-eight-eight in twenty twenty-two. The idea is to make it as memorable as nine-one-one for a specific category of emergency. But nine-eight-eight is a phone-based counseling service — they can talk to you, they can de-escalate, they can dispatch mobile crisis teams in some areas, but they're not a physical response in most places. It's a call center, not a van that shows up.
That's the gap. Nine-eight-eight is great for someone who's in crisis and able to make a phone call. It's less useful for the bystander who's looking at an unresponsive person on the ground. For that, you need a physical response. And the landscape of who provides that physical response, and how fast, and under what circumstances, is wildly inconsistent.
Let's talk about what a good non-emergency response system actually looks like, because I think Daniel's asking for the blueprint. What would you tell someone to push for in their city?
A few elements. One, a dedicated non-police crisis response team that operates twenty-four-seven. Not nine to five, not weekdays only. Crisis doesn't keep office hours. Two, a dispatch system that's integrated with nine-one-one so that dispatchers can triage calls and route the non-criminal ones to the crisis team. Three, staffing that includes both clinical professionals and peer support workers with lived experience. Four, the ability to transport people to shelters, detox facilities, or mental health centers — not just to hospitals. Five, proactive outreach capacity, not just reactive response. And six, stable funding — not a pilot grant that expires in eighteen months.
That's the wish list. What's the minimum viable version? If you're in a city that has none of this and you want to start somewhere?
The minimum viable version is a co-response model. Pair a mental health clinician with a police officer. The officer handles safety, the clinician handles engagement. It's not ideal — the police uniform is still there, the power dynamic is still there — but it's better than police alone, and it's easier to sell politically because it doesn't look like you're defunding anything. A lot of cities have started with co-response and then evolved toward independent crisis teams once the model proved itself.
The co-response model as gateway drug for full alternative response. I can see the logic.
The other minimum viable piece is just publicizing the non-emergency number. Every city has one. Most people don't know it. Put it on bus stops, put it on library cards, put it in the contacts list on city-owned phones. It costs almost nothing and it would redirect a meaningful number of calls away from nine-one-one.
The library card idea is actually good. Every library card in the city has the non-emergency number printed on the back. You're already carrying it. You see it every time you check out a book. It's ambient awareness.
See, you're a public policy mind when you want to be.
Don't tell anyone. I have a reputation to maintain.
There's another piece here that Daniel's question implies but doesn't state directly, which is the cultural shift around calling for help. A lot of people, especially in communities that have had negative experiences with law enforcement, have a norm against calling the police for anything. That norm exists for good reasons. But it also means that people in those communities are less likely to get help when they need it. The alternative response model potentially bridges that gap — you're not calling the police, you're calling a crisis team. It creates a category of help-seeking that doesn't carry the same baggage.
That's the real answer to Daniel's question about a safer path. The safer path isn't just a different phone number. It's a different kind of responder — someone who shows up without a gun, without a badge, without the power to arrest you, whose entire job is to figure out what you need and connect you to it. That's a fundamentally different interaction.
And we should be clear about the limits of this model. It's not appropriate for every situation. If someone is actively violent, if weapons are involved, if there's a clear criminal element — that's still police work. The alternative response model isn't about replacing police entirely. It's about taking the huge volume of calls that don't require a law enforcement response and handling them with a more appropriate tool.
The right tool for the right job. It's not complicated. We just built a system where the only tool was a hammer, and then spent thirty years wondering why everything looked dented.
The other limit is that a crisis team can stabilize a situation, but it can't solve the underlying problems. The person who's intoxicated on the sidewalk might need housing, addiction treatment, mental health care, a job, a support network. A crisis team can connect them to those things, but only if those things exist. In a lot of cities, the services don't exist, or the waitlists are months long, or the eligibility requirements are so narrow that almost nobody qualifies.
The crisis team becomes a revolving door. They show up, they stabilize, they refer, the referral goes nowhere, the person is back on the sidewalk next week, and the same bystander calls again. That's not a failure of the crisis team — it's a failure of the broader system. But the bystander doesn't see the broader system. The bystander just sees the same person in the same spot and wonders if calling even matters.
That's the burnout risk for these programs. The responders themselves can get demoralized if they're constantly stabilizing people and then watching them cycle back into crisis because the follow-up services aren't there. Staff turnover is a real issue in a lot of alternative response programs. The work is hard and the systemic barriers are frustrating.
The bystander burnout is real too. There's a concept called compassion fatigue that usually gets applied to healthcare workers, but I think it applies to communities as well. If you've called for help three times and each time the person ends up back on the same corner, the fourth time you might just walk past. Not because you don't care, but because you've learned that caring doesn't produce a different outcome.
Which is why the proactive outreach model matters so much. If the crisis team is building relationships over time, the bystander doesn't have to be the one who notices and calls. The team already knows the person. They're checking in regularly. The crisis is prevented or caught early, rather than escalating to the point where a stranger has to make a nine-one-one decision.
Let's circle back to something specific Daniel asked about — non-emergency police lines. What's the actual experience of calling one? Because I think a lot of people imagine it's just nine-one-one with a longer hold time.
It varies by department, but in most cities, the non-emergency line is staffed by the same dispatch center that handles nine-one-one. The calls are just prioritized differently. If you call the non-emergency line and say there's a person down on the sidewalk, the dispatcher will ask the same questions a nine-one-one dispatcher would — where are you, what do you see, is the person breathing, are they responsive. The difference is that your call might be queued behind active emergencies. The response time might be longer. But the triage is the same.
For the bystander, the experience is similar, but the psychological barrier is lower. You're not dialing the emergency number. You're not making the declaration that this is an emergency.
For a lot of people, that distinction matters. Calling nine-one-one feels like pulling the fire alarm. Calling the non-emergency line feels like asking for a wellness check. Same outcome, different emotional framing.
The fire alarm versus the wellness check. That's the whole thing.
Some cities have made this even easier by creating dedicated numbers for specific types of non-emergency response. Los Angeles County has a mental health access line. New York City has NYC Well. These are staffed by mental health professionals who can do phone-based counseling and dispatch mobile crisis teams. They're not police-affiliated. They're run through the health department or through contracted nonprofits. The branding is completely different — you're calling a mental health service, not a law enforcement agency.
Which is smart branding, because it signals to the caller what kind of response they're going to get. If I call a number called "NYC Well," I expect a social worker, not a cop. If I call "non-emergency police," I expect a cop who's not in a hurry. Those are different expectations.
Expectations shape behavior. If people expect a police response, they'll only call when they're prepared for that outcome. If they expect a social-service response, they'll call in situations where they'd never dial nine-one-one. Expanding the range of situations where people feel comfortable calling for help is the whole point.
To answer Daniel's question directly: yes, there are safer paths. They're fragmented, underfunded, inconsistently available, and poorly publicized — but they exist. The non-emergency police line is the most universally available option. The mobile crisis teams are the gold standard where they exist. And nine-eight-eight is the national number for mental health crisis, with the caveat that it's a phone service, not a guaranteed physical response.
If you're listening to this and wondering what's available in your city, the fastest way to find out is to search for your city name plus "mobile crisis team" or "crisis response" or "non-emergency police number." Save whatever you find in your phone. You're not going to remember it in the moment. The moment is when you need it to already be there.
Pre-saving the number is the difference between using it and not using it. Nobody is going to google "does my city have an alternative crisis response program" while standing on a sidewalk at eleven at night looking at someone who might need help.
One more thing worth mentioning — there's a growing movement to make these numbers more standardized across the country. The three-digit number nine-eight-eight is the mental health equivalent of nine-one-one. Some advocates are pushing for a similar three-digit number for non-emergency public safety response — something like three-one-one but with more consistent implementation. The idea is that no matter where you are in the United States, you should be able to dial the same three digits and get connected to the appropriate non-emergency service. It's not there yet, but the conversation is happening.
The three-digit number is powerful because it bypasses the need for public education. You don't have to teach people seven different local numbers. You teach them three digits. Nine-one-one for emergencies. Nine-eight-eight for mental health crises. Something else for everything in between. The cognitive load drops to near zero.
That's the goal, ultimately. Make it as easy to call for the right kind of help as it is to call for the wrong kind of help. Right now, the wrong kind of help is the easiest to access. Nine-one-one is universal, fast, and guaranteed to produce a response. If we can make the appropriate response just as accessible, we change the default.
The default, as we said, is the enemy.
The default is the enemy.
We should wrap this up. But I want to leave people with one thought. The next time you're in that situation — looking at someone on the sidewalk, not sure what to do, running the calculus — remember that calling doesn't have to mean calling the police. There might be another number. And even if you don't have it saved, even if you end up calling nine-one-one, you can tell the dispatcher what you're seeing and ask them to send the right kind of help. You can say, this doesn't look like a crime, this looks like someone who needs a social worker. The dispatcher might not be able to honor that request, but you can make it. And the more people who make it, the more the system has to adapt.
Now: Hilbert's daily fun fact.
Hilbert: In the early fifteen hundreds, the Ottoman Empire maintained a dedicated corps of palace food tasters whose employment contracts specified they would be compensated in lifetime supplies of salt if they survived accidental poisoning — a clause preserved on a single surviving payroll ledger from the reign of Bayezid the Second, recovered from a private collection in Kiribati in nineteen eighty-seven.
Your compensation for not dying is salt.
Presumably a lot of it.
This has been My Weird Prompts. Thanks to our producer Hilbert Flumingtop. If you want more episodes, find us at myweirdprompts dot com or wherever you get your podcasts. We'll be back soon.
