Safe: Set maximum capacity for any room, floor, or building | Hacker News Reader

Safe: Set maximum capacity for any room, floor, or building

density.io

135 pointsjordanmessina81 comments

The idea is definitely cool (and important/impactful at this time if successful). That said, in the midst of a global pandemic where human density is inherently dangerous and social distancing is espoused, naming a product "Safe by Density" gives the exact opposite initial reaction you're probably hoping for.

Yeah. Should have been "Safe by reducing Density"

"Safe from Density"

I've replaced the submitted title ("Introducing: Safe by Density") with a more specific phrase from the article.

I guess the old "Safety in numbers" adage is going the way of the dodo?

Safe by scarcity

It is a bit Orwellian. War is Peace, Ignorance is Strength, Safe by Density.

This is a cool concept, though it appears to only care about the total number of people being below a set threshold regardless how people interact (i.e. maintain social distancing)

Shameless plug: two friends and I put together a small app that uses security cameras feed to calculate real-time density + estimate "safe" space capacity:

Demo video: https://www.youtube.com/watch?v=RTgxhptePxM Demo site: https://flou.fyi/ (takes a bit to load)

I think it could work. Heavens know, I will now have to actively avoid places that are not at least careful about how they reopen.

I am mildly concerned about people, who see doing the opposite as a political expression, but some things cannot be helped.

The bet is that by giving data back to people to let them make their own decisions about safety, they'll self moderate.

Except for the workers, paid predominantly minimum wage or near it, who effectively have no choice.

Actually, many deployments are going into factories, fulfillment centers, and meat processing plants specifically to help keep workers safe.

> I am mildly concerned about people, who see doing the opposite as a political expression

Those people are a very small but vocal minority, being amplified by social media.

All the data says that most people are complying with safe practices, even when their governments aren't requiring it.

Doesn't help when things are pressured into opening up and then you're forced to make the decision to put yourself at risk by going to work or get fired.

> Those people are a very small but vocal minority, being amplified by social media.

I disagree. Tens of thousands of people in my state don’t obey any stay at home order. People take off masks once in stores.

As someone who also goes out a lot right now I haven’t seen anyone be asked to put their mask on or leave.

Which state? And how do you know tens of thousands are ignoring it?

Georgia reopened weeks ago and yesterday hit their lowest levels of cases and deaths since the outbreak began at 22 and 3 respectively: https://ga-covid19.ondemand.sas.com/

Georgia reopened 14 days ago. The incubation is 14 days.

If their numbers stay low in the next few days then great!

But if they start going up, they will have sacrificed the lives of Georgians to prove that it’s not safe to reopen.

I hope the cases stay low.

No, they started reopening on April 24th: https://abcnews.go.com/Health/wireStory/kemp-shuttered-georg...

So yeah the models were wrong.

It was a very limited reopening on April 24. The full reopening didn’t happen until April 30.

I love this idea and I love how it looks. The design is top notch. Really hope to see this near me soon.

The stylized design makes my brain recognize is as just another random inspirational poster or an add, not something I should pay attention to. We have standard and widespread symbols already for stop/go which should be applied here in some form.

That's an interesting point. What would you recommend? Follow street sign structure?

Having seen the way people 'queue' when waiting outside an eatery implementing social distancing guidelines - this only solves part of the problem.

Edit: Perhaps integrating it with some sort of QR Code based ticketing system for entry would alleviate the issues of people queueing (or huddling) outside the store?

You're right. "Notify me when it's quiet" is a thing the system supports. In related news, load balancing humans is a two-sided thing.

So ... how does it work, given that 2m is not enough, and the distance of 10m is thought to be more likely? :)

My first thought is this puts way too much stock in the idea that rooms are safe as long as they aren't too packed. To me it seems like this 1.5m rule and person limits aren't actually backed up by real research and are just numbers pulled out of a hat when something had to be done fast.

We shouldn't be finding ways to get people back in to shared buildings as soon as possible because infections don't just give up because the TV on the door said OK.

I was just thinking that the difference in fatalities between NY and SF might be that NY has a colder climate, so airtight buildings with more heating.

A virus like corona with a high R0 would spread like crazy in a building with active ventilation piping. Or an airliner.

Corona is really a non-event in SF.

I could not find whether this integrates with NFPA etc. to help interpret blueprints (or simplified diagrams). While determining capacity is relatively easy, it could definitely be easier.

Interesting, but I'm skeptical it could make some semblance of normal life "safe" purely through occupancy counts. It also seems like a likely very expensive solution to something that could be solved with a handheld clicker. Corporate offices might splurge for it, but they're also more likely to just have people working from home.

You think a software solution is more expensive than someone on the payroll with a handheld clicker?

Depends mainly on whether you already have employees working as receptionists/greeters/etc.

More difficult with multiple doors when you have radio between to reconcile count.

I think people think that shooting down ideas makes them sound intelligent and “in the know”, but in reality it just makes you sound rigid and not creative enough to see the merit in a idea.

This seems to be a common attitude on HN, but it’s less than useful. Genuine critical thinking is great... dogmatic cynicism is useless.

I think there are some valid concerns here. Someone running a clicker is already common procedure at some busy places, and it's often a hostess or someone else who is already "doing stuff".

Likewise any sensor is going to open up some questions about the data it collects and it's method of action - what's the error margin?

I think there's arguments for systems like this as well, but it's on you to make them, not just complain about others being skeptical.

Accuracy is fundamental. When installed properly, a 99%+ f1 score is achievable.

First off - this is really neat, especially the focus on privacy and the use of AI at the edge.

It seems like the occupancy limit problem for small rooms is much more challenging than the building-density problem, since a single error could DOS the room.

How do you account for that?

Def. We call that the OB1 problem (off-by-one). If either event detection (did something happen) or event classification (entry or entrance) is wrong, you can be off the whole day. To solve that, we have a different approach / non-threshold approach for boolean occupancy. Not implemented here.

What data does the sensor collect? How visible is the data it sends back to the platform?

It's a custom sensor in the lidar family. Uses infrared lasers as illumination and generates depth data. Essentially millions of height values. When depth is rendered to be human legible it look looks like greyscale silhouettes (dark grey is far away, lighter gray is nearer). Sensor processes those values on the edge and published +1/-1 and telemetry data (system health).

I think people think that criticizing others by over-applying perceived HN tropes makes them sound intelligent and "in the know".

I wasn't shooting it down, I was just expressing skepticism toward the complexity of the solution and opening it up for wider discussion.

Skepticism on this stuff is valid. If we're going to distribute large scale infrastructure that measures human movement we ought to be critical of it. So many systems purport to work but don't. Others are way over engineered. And a lot market as if they're sensors when they're really just cameras taking pictures of behavior.

It is not creative to count the number of people in a room.

I’m currently working on a similar proprietary system that is similar to the one described in this article, and I can tell you that there is much more involved here than “counting people in a room”. Claiming someone with a clicker in front of a door is a better solution misses dozens of other use cases where this isn’t at all possible or cost effective...

But you know, everyone on HN is an expert in everything.

This. People are weird. They do strange stuff and it's really hard to count accurately.

Definitely not creative just really, really hard. Took us nearly half a decade.

Nets out to ~$85/mo. For a small place, that's substantially less expensive than staffing it full time.

disclosure: co-founder here.

I didn't see the pricing; that's more affordable than I would've thought, especially with the rent-instead-of-buy model.

In an ideal world, we just give this stuff away and the applications on the platform become the primary value. Some people are open to that others are less used to having a widget and not owning it. Will take time.

The name bothers me a bit. Social distancing does not make us “safe” just “safer”. It worries me that people might think things like masks and possibly this tool are more effective than they actually are and become complacent.

It's definitely going to take a village of tools and methods. Don't think it will make it more dangerous, though.

A phone app could replace this right? Using the front facing camera you could run some image recognition to count the flow of people and show the statistics on screen. Put the phone on a stand and leave it running. Much cheaper.

Their whole selling point seems to be that their hardware + software can count people accurately without needing a camera. The premise being that cameras spook people.

Cameras are already commonplace enough where I work, for security purposes. So not sure how easily cameras can be eliminated.

The dashboards they provide don't seem to be have to be coupled to the kind of ($850) hardware they are selling. Like you said, cameras should be able to do the job.

I don't think cameras will be displaced / eliminated. It's just a question of whether or not they will be accepted as a form of active surveillance or remain a method of security. Today, they are largely security but it's entirely possible the world moves toward mass surveillance.

Many have / are trying this. Harder to pull off than it seems. Privacy is a big part but distributing the infra to do this at scale is nontrivial. A number of years back Placemeter used to pay you 50/mo to install your Android device on a window sill so they could understand movement. Never quite took off.

Or you can just fingerprint the wifi/bluetooth signals from people's phones.

MAC address tracking is one approach but it's imprecise and with the proliferation of "things with antennas," you have to do a lot of reconciling on the backend to not count 1 person as 3 when they have multiple devices on them. Euclid analytics tried this. It's a common but flawed approach to count (use depending).

How do these devices actually work?

Here's an interview from a while back: https://softwareengineeringdaily.com/2018/07/16/counting-peo...

"Indoor Densitivity"

Viruses don't care about how many people are in a room, and this doesn't seem like something people are going to worry about much 10 years down the road.

At the same time, I'd much rather shop in a small neighborhood grocery store with 5-10 other patrons instead of 200.

Trader Joe's and other retailers (at least in the bay area) have to delegate a full-time employee to the front door to throttle incoming patrons. This would free up that employee.

There are grocers who have required all corporate employees to work at least 3 days a week in a store bagging groceries, cleaning carts, and managing lines. Side not, essential businesses are the only real experts in safety right now. Everyone WFH is just guessing.

>This would free up that employee.

In my area Walmart has started making each aisle one way and counting people coming into the store. Just yesterday I went in and it's already devolved into no one following the directions on the aisle, no one seems to be in the slightest bit considerate about trying to give others space to go around. People were regularly reaching around someone getting an item and coming into close contact instead of waiting all of 5 seconds for them to get their item and move.

There are already plenty of news articles about customers doing increasingly outrageous things and attacking employees because they dared to tell the customer not to violate store policy around purchase limits, senior hours, mask policies, occupancy limits, etc. The fancy sign telling someone to wait will not free up employees because there will inevitably be a chunk of the population who ignores it outright and when they just walk in the door, others will follow. These policies are meaningless unless there's some actual enforcement.

I agree with this. It's compliance will be contextual / regional / etc. This is why the system also support real-time alerts and safe analytics. In the event a place like Walmart wants to more actively enforce they can. Display is just the public facing feature.

Agree. Intent is to kill the product once this is all over and use the infrastructure to design better cities.

They do "care".

First of all, if that number is in the interval [0,1], it's a qualitatively different situation than if the number is in [2,∞].

Jokes aside, there is a statistical relationship between the number of people in a room and the distance they can/will keep from each other. More density means more close encounters on average. Less density doesn't guarantee nobody gets too close to someone else, but it makes it less likely.

Good example of this is in elevators. People fill the available space like water.

I am sad about how far apart we're growing as people due to virus that is no more scary than the flu.

Funny you say no more scary than the flu when the flu is literally the cause of the last pandemic that killed 3% if the global population.

Sure, to be more correct I mean a mild flu season, not 1918.

you mean to imply covid-19 is no more scary than a normal seasonal flu ?

Yes. Are we to live our lives always like this now - always trying to stay 2m away from each other?

if not keeping a 2m distance if going to get us killed, then I guess many would not have a problem with keeping the 2m distance. Ultimately it is all about survival.

This is all way overhyped. The virus is not as deadly as you're being conditioned to believe. It's about control

Relax dude, it is a pandemic, we have enough data which tells us how contagious it is, and what needs to be done to control the spread. It's going to pass, here in India, we have opened up almost all zones where there are no cases, and activity is almost back to normal.

Contagiousness is different than mortality. It seems quite contagious, but low mortality. But I am not in India right now - I am in the US where hysteria is everywhere and they are talking about continuing lockdowns for months longer, or basically indefinitely.

indeed contagiousness is different from mortality, but think of it like this, if more humans are infected at the same time, it'll overload the hospital capacities, this will 1- make it impossible to ensure proper care, since the hospitals will be overwhelmed. 2- also impact patients who require treatments for other illnesses, since systems will be busy and overloaded with the covid-19 patients.

Hospitals in the US are so lacking with patients that they are furloughing and laying off doctors and losing money. They never were overwhelmed and would not have been

Mortality rate for the coronavirus is about 0.3% - the flu is about 0.002%. These are not comparable numbers. For every person who catches the flu and dies from it, 150 people who catch the coronavirus would die.

This is with the best treatment we can afford given the current contagion rate - if more people were infected, we could not afford treatment for them, and the mortality rate would climb.

Reply on news.ycombinator.com