This is a very good idea; and it could probably help in more ways than anticipated.
Experts have been arguing (very well I might add) that government agencies have been modeling and analyzing the pandemic wrongly. This is resulting in data that seems pretty bleak when it’s not likely so bad. Here’s an explanation by a PHD specialist in mathematically modeling epidemics.
A researcher who has been modeling economies as though they were epidemics for decades, and has a list of degrees as long as your arm, has been pointing out that the data isn’t backdated or corrected for testing ramp up.
Another AI researcher who has a bioinformatics company that models and analyzes epidemics, among many other things said that the data is really bad; back in January.
An application like that could easily time stamp data; and give detailed information about transmission. It could help with many of the problems that we have been facing for the past few months.
The reason that I suggested adding accountability in the EULA is because those who would not use the application properly could be black listed from using it. They tend to be much more clever at doing such things than hiding the fact that they are doing it. My intention was to add another layer of protection for the user.
You are familiar with my work. When I created bot framework, relevancy and simulation concepts, I was not aware of Agent based modelling and graph theory and all that shit. To be honest I still don’t know anything about them except they existed since 80s. However agent based stuff admittedly has a few things common as bot framework.
Agent based modelling and simulation may be superior to other modelling techniques, however we need to understand that a model is nothing but an approximation. They are useful in some sense yet fundamentally wrong. I don’t think we can simulate complex systems accurately where human behavior is involved and other unexpected things.
For example, you can never implement what I am proposing as counter solution to lock-down using simple modelling and simulation, It can not be achieved without participation. There are just too many variable and unexpected things happen. For example, in the middle of crisis here, some communal tensions building up. How can you possibly account that? In the end, you end up with more dangerous situation with these approaches and resort back to total lock down. Nothing gets solved…
Apart from that, modelling and simulation on its own is not solution, they are just tools. People using and building these tools must realize that these are unreliable by nature. Real solution is participation and contribution and how well it’s implemented.
Also relying on statistical modelling is actually very dangerous. Because it gives a false sense of some truth but it’s only as true as your input. I would not rely on assumptions based statistical modelling for things like pandemics where consequences are very real.
Even for logistical purpose as wrong simulation result can create scarcity of essential things where it is required most. Historically we have seen that such tools never seems to work beyond some basic predictions.
THE BACTERICIDAL ACTION OF PROPYLENE GLYCOL VAPOR ON MICROORGANISMS SUSPENDED IN AIR.
It has been found that propylene glycol vapor dispersed into the air of an enclosed space produces a marked and rapid bactericidal effect on microorganisms introduced into such an atmosphere in droplet form. Concentrations of 1 gm. of propylene glycol vapor in two to four million cc. of air produced immediate and complete sterilization of air into which pneumococci, streptococci, staphylococci, H. influenzae , and other microorganisms as well as influenza virus had been sprayed.
Short story? Vaping might help protect you and others from viruses such as Pneumonia and the Flu which can help mitigate your chances of catching them during this time. Why do I say Vaping? Because most Vape Juice is a 70/30 mix of Vegitable Glycerine and Propylene Glycol.
We did the math and determined that a 10’x10’x8’ (LxWxH) room is roughly 2-4 million cc of air volume. 1gram of propylene glycol is equal to 1 cc of volume. Thus, since vaping literally vaporizes the mixture, which in turn makes it an Aerosol, vaping may help sterilize an area as well as prevent lung infections.
Every single slide looks like something from a social score program. All the way down to a constantly changing Code system. Look at the Bottom left corner. CODE: 64MCN
The LAST thing we need in this day and age is a social compliance program scripting our daily lives based on compliance. Also, HOW exactly is this program a SOLUTION for Covid?
Well it’s not every slide… 64MCN is a code/numbering standard that is used for simulation. it’s not social scoring… and if you are going to assume things without understanding then I don’t know what to say.
you are late to party… you gotta read it… I can’t write same thing again just to explain you…
So, you are saying that no one in authority would want to know your “simluated” score? And No one in authority would block you from doing certain things or going places based on the simulated score?
this thread is ABOUT Corona and Pandemic Solutions…I submitted a potential mitigation for COVID-19 and associated illnesses. Something that can actually be used now.
I appended that link to the end of your original post and made a notation that I did the edit. If the link is wrong or you feel there is something more appropriate, let me know
Thanks… I think I will have to write a proper summary. I have not been getting enough time and this has become a giant mess… I’ll update it as soon as I get time. I think I am gonna make a video summary or something this weekend.
“If implemented correctly” there is no perfect implementation, there are always bugs and issues, and sometimes backdors. Just saying.
“It will be almost impossible …” Isn’t enough today, to be bulletproof, it has to simply be impossible.
But how do you even proof or disproof that it is or isn’t?
The specification of a design is of cause a or rather the base for the implementation.
I don’t think anything what you have written so far can be called a design or spec.
That might be a good point to start actually, how about you make a spec?
There are many tools and guidelines out there to do so.
I had to do a lot of UML in uni, might be worth a look for you.
To nitpick the “trust” part, one aspect of the buzwordy “Block chain” is to eliminate the need for trusting a middleman, like a bank.
I’d say there is a Trend in CS to eliminate said need of Trust.
Why could that be? Isn’t the concept of trust automatically implying that there are ways to abuse it?
But to get away from all the nitpicking, sorry for that, how about we try to focus on a small and rather easy to figure out, aspect or usecase of the system?
Something we can figure out a mini spec for.
So it shows a timeline of locations and activities, specifically what zones where passed through, assumably on foot, which uber rides where taken, and which people where in close proximity.
It then also shows a “risk?” status which in this specific case is rather high for a person tagged 34MCA.
The app on my phone has access to my GPS, it can therefore know where I was and am, but where is the information about the zones coming from?
There could be a public download of a map with the zone info that the app could grab and then look through with the position info.
issues with that
Though this would potentially leak the home and location of all infected people and even if they behave good like they should, they could be discriminated against, or shops in the area would loose income because noone wants to go there anymore.
Or the app sends off my location data to somewhere that checks all the info and then tells me about the zones and risks.
Any other idea how to do that feature?
Uber would need to have an API available to get the ride information, or I would need to manually enter that into the app.
GPS data and speed of travel could also warrant guessing on the app side but well, that isn’t great.
But how is the information about 34MCA acquired?
Does his app broadcast his status over WiFi and Bluetooth so the app just has to listen in?
Or does that also happen in the cloud? I mean, it could be done in the cloud but of cause with the obvious downsides.
Both aren’t that great in my opinion, any other options that I’m not smart enough to see?
ISPs can triangulate positions too
from their antenna data. “Funkzellenabfrage” in German and here at least, the isps are sharing that data with the state to “know” how well people stay away from each other.
So beside WiFi and Bluetooth or the cloud, that is another place where position data can be farmed.
General issues with that
How stringent is the detection and hence restriction?
If its not made strong enough, the pandemic will probably still spread.
If its made very strong, it is at a minimum bordering on a lockdown.
And then there is the question of inequality, when just a single house, a block or city is locked down but nothing else.
This is dangerous territory.
Maybe its not a cognitive bias.
It sadly seems as if China already implemented something like that, with a focus on surveillance and control of cause, I assume.
Just read that the people in Wuhan had to have a app on their phones.
A mandatory App.
Onto the garbled mess that your “simulation” part of the topic is, reality is a dirty mess.
If you had access to the Matrixes APIs, sure, everything you say about that is true, then you could do some cool and crazy shit.
Great. I like your questions… Give me sometime to answer them. Thanks
update…
I agree. I am not looking at a perfect implementation. There are bad, good and great implementations. I am talking about great implementation that is built around usecases and not having hidden motivation to garble up data. So basically I am talking about the intent of implementation which comes before you start thinking about technological implementation.
Yes, almost impossible isn’t enough. But I also don’t want to claim it’s bulletproof mainly because it’s just a higher level vision and you can’t be sure of things at this level. Most you can do at any higher level vision is to avoid obvious failures. It’s upto engineers who are designing it.
I would but we are not there yet… First thing is create a higher level vision that makes sense, then share ideas and get some feedback. See the obvious drawbacks before you start designing specs of the system. That’s what your feedback is doing right now… I find it very stupid to start developing technical specs before actually understanding ideas, feedback, vetting it etc.
My strategy is ideation which comprises higher level vision, then abstraction where you start going deeper into granular things. It has some advantages and drawbacks. Advantage being, you kinda see a low resolution picture and possible paths to explore into. Drawbacks are that it’s an iterative process and you need to keep changing and refining initial ideas.
So trust I’m talking about is more on the line of scientific methodologies. Honestly I am not yet very sure of the implementation but the general idea is, we need to be able to work with generalization without having deep expertise in things, it could be achieved with help of science and statistics hence companion thingy. Trust is highly associated with the source of the information, which in turn creates an opportunity to exploit. With scientific methodologies, any transactions could be validated, like validate information independent of its source. People will benefit from it greatly as not everyone is an expert. People don’t always have good judgement on things that may be important in case like Covid 19…
I was hesitating to go into details because it would create a lot of confusion and a lot of questions from a security point of view. But for now, I want you to imagine that despite the hyper-connected ecosystem I’m suggesting here, the security and privacy is robust.
So imagine a framework which gives identity to most of the objects that matters like houses, shops, atms, cars, roads along with people. I call it Bot Framework, so each object is a Bot which is kinda aware for its purpose. CarBot, HomeBot, ShopBot, SignalBot and PersonBot etc…
All bots have memory function and update their states as interactions happen. Depending on bot type, there could be some IO devices for each of them, for example cars could have GPS. People have phone and vital tracking bands or something. Shops and houses have some detection devices. These devices interact with each other and create timestamps, simplest interaction would be that devices came in close proximity… In the case of Uber, you extrapolate as you travelled along with the car and the driver.
So it’s a mesh network of devices that are interacting with each other connected in a secure Bot Framework or I like to call it as Logistics of Intelligence Management. I really want to make this clear that even though all sorts of interactions are happening, the data is blackboxed for each bot. Users have full control of their interaction and control sensitivity so the interactions are limited to essential things, meaning they can turn down the sensitivity of interaction if not participating in some problem solving use case or simulation studies. In case of COVID 19, the sensitivity would be much higher as we have to track the spread.
You own the intelligence of all your devices. So you have Honda, all the intelligence of the Honda CarBot is yours. So data and timestamps are being recorded but not being send over to anywhere but your ComputeBox instance… and when you want to participate then you agree to share whatever data required for the participation. Like sharing 3 months of movement of your car for some traffic simulation study. I am not good at explaining but I hope it gives you some ideas.
So no data is being leaked. It’s just recorded. You ComputeInstance will process all the data.
It does not expose all the infected person to everyone, depending on abstraction and who should be informed and with your permission. For example you interacted with an infected person then you would be informed. Your local responders would be informed.
Also you obviously design your algorithms to counter obvious biases and account for the nature activity over time. like if an infected person interacted with an ATM machine, the ATM machine would not forever be banned. The state of the ATM will be renewed based on several actions, like it has been cleaned or enough time has passed to be infectious.
Most automated information will come from mesh network of devices on bot framework as stated above. Others will be self reported by users as in manually enter. Some information are extrapolated… There could be more and better ways someone else can come up with.
Mesh again… So for example, if your device recorded the interaction with someone else’s device, like you came in close contact with another person, that interaction data will not be accessible to you. you only have a timestamp and transactionID that is common with interacting party, the actual data is encrypted and only accessible based on abstraction policy and user’s permission. Also the interactions will be forgotten with time as set by owner of device. and no one can get your data without your permission… same privacy laws as phone… These are just higher level ideas as part of how intelligence framework would work…
Think ComputeBlocks connected in the network. Each person owns a computeblock and all their data from SelfAI and devices bots are stored in it…
I am pretty sure it is, the reason why I am saying this because a lot of people rely on their default intuition. So any idea that remotely resembles their worst nightmare is rejected even before knowing what it is… Historically we have seen that with everything… People don’t like change especially if it’s a lot… On normal days no one sees a need for such things. It’s only when disaster like this happens and then you end up with some crazier implementation like China.
Yes, but I would appreciate it if you didn’t compare it with Chinas implementation because it’s nothing like that. The whole system is thought around individual security and individual empowerment and problem solving.
I hope your perception is different on simulations I was talking about after reading the explanation with Bot Framework and ComputeBlock network and devices mesh.
Also you wanted to know who’s adding and updating the data?
It depends on abstractions, All kinds of people will engage in platform and frameworks on personal and professional level… People who are responsible for the abstractions will do, like making some region as high risk zones will be done by local authorities.
PS: I like your questions. But I also remind you that I am not some mutant or something that can do everything from conceptualizations to system design to technical specs right away. It’s a process of abstraction. I like to discuss things first and understand peoples opinion before diving deeper…
Yeah. That’s the point. Real time, real world data, that is timestamped, is just more useful for complex systems analysis than even the best linear model. The models aid in understanding the system; but the real time data tells you exactly what is happening. It doesn’t get everything; unless everyone participates, but the raw data, presented in the right way can help quite a bit with analysis. Experts know what to do with the raw data; and many of them couldn’t get the data they needed.
