• Arzhang Kamarei

Coronavirus Exchanges to the Rescue?

Updated: Apr 13, 2020

Problem: Here is a problem I would love to get the tech world focused on:

We do not have the intelligence infrastructure for a species-wide coronavirus collaboration.

I'm seeing a tremendous amount of effort in the tech and philanthropy community to organize, but it's being done on very shaky technology and organizational infrastructure.

[UPDATE: New coronavirus exchanges are starting to pop up - see bottom of article for list]

For example:

  1. We are coordinating coronavirus crisis responses on social media platforms such as FB and Twitter typically used to share cat videos. This is not, for instance, a good way to get 20MM N95 masks organized and shipped from Asia to the US. I’m seeing confusion, repetitive questions, and coordination problems in all these early efforts.

  2. Otherwise, we are using tools like Google docs to organize relief efforts - tools which do not scale. Check out this PPE effort which abandoned their Google doc. The CEO of OpenAI gave up on his Google spreadsheet and went to email. The CEO of OpenAI had a tech problem.

  3. There are over 25,000 coronavirus projects on GoFundMe (as of 3/28/20) - none of these seem coordinated. How is this any kind of reasonable, rational effort?

  4. Articles like this (3/31/20) describe the chaotic business of N95 masks, none of which is on a centralized or transparent exchange, allowing for price gouging and chaos.


We need to build coronavirus organizational infrastructure - fast.

If we can make any improvements in 1, 2, or 3 - not the perfect solution, mind you, just improvements - then we can accelerate our response to coronavirus.

Here is image of what I think is happening (nodes = intelligent organizations of people):

So to get us out of Scrambling, here are 5 basic ideas and then one big ambitious idea (Exchanges or "Matching Engines").

[Note: this is the business I used to be in and we built some of the most efficient exchanges in the world, so I know this is all possible.]

5 Basic, Common Sense Tech Things We Should Build to Coordinate Better

1. Global Needs Database with Subject Matter Expert (SME) voting & contact info, search, filter, share functionality

Who needs what and how can we reach them? Can we see all that info in one place? I have a friend who was trying to reach Governor Cuomo to give him N95 masks but there was no way to get in touch with him. How frustrating is that? We literally couldn’t *give* N95 masks to the state that needs them the most because of… contact info? And now Michigan needs them… I think. But we don’t know. Since there is nowhere to report these needs.

2. Global Assets / Helpers Database with SME voting & contact info, search, filter, share functionality

Who is trying to help and what can they offer? Is there a list of all the foundations who want to help in one place? Why is everyone asking and answering this question 100x in my FB feed instead of getting organized?

3. Central Coronavirus SME directory with self-organizing social graph, SME rankings, and contact info

Who are the coronavirus Subject Matter Experts (SMEs) and how do we get in touch with them or their teams? What do they think we need to do? Why aren’t we asking them? Is it because we can’t get in touch with them? Why do I see people on FB saying “I have resources to deploy, what do you think I should do”? Why are we employing “Move Fast and Break Things” here? This is analogous to war. Do you know what Navy SEALS say? “Slow is smooth and smooth is fast.” We need the experts involved.

4. App / Website ranking tool to show the most used coronavirus sites/apps, traffic stats, SME rating, & contact info to connect with / help them

What is the coronavirus tech everyone is using, the best rated stuff, and is there an easy way to find that out? This 18 year old kid’s website seems faster to publish data than Johns Hopkins - why don’t more people know about it?

5. Virtual, Roving Tech Support Teams to help coronavirus related tech platforms solve their tech problems

There is a bunch of tech out there run by Subject Matter Experts that is slow or under-engineered (for instance the Johns Hopkins website was super slow for weeks - finally got better). These SMEs on the front line are trying to share data and instead of helping them with their tech (which we could be good at), many of us are trying to guess at what new apps might be useful for them. Are we that good at guessing? Because none of us really guessed that this disaster was gonna happen. Why not combine our awesome tech with their awesome subject matter expertise?

That’s Not Complicated Enough to Keep Me Interested

Now let’s assume that we don’t want to only do the basic, useful stuff above, but would rather build something more ambitious and more complicated.

Now, if you take the top four items on the list above and combine them what does that look like?

  1. Global Needs Database

  2. Global Assets / Helpers Database

  3. Central Coronavirus SME directory

  4. App / Website ranking tool

If you have any experience in quant finance, you may recognize that this looks very familiar - and potentially very useful - a Matching Engine or Exchange.

In fact, the main big thing I would suggest we build are Coronavirus Exchanges (aka "Matching Engines").

A matching engine helps combine a Need on one side with a Resource on the other. Where does the concept come from? Matching engines are a fintech tool based on the architecture of stock exchanges.

I’m not a UI designer and the below is crappy (and could be incredibly different), but these are just some starting ideas. You can see there is a public “ticker tape” at the top and the matches are below.

The goal of exchanges is the decentralized coordination of resource matching in new networks.

So if you are a hospital and you have a regular supplier you are talking to? You don’t need this. If you are a hospital and need to find 10 beds in 24 hours and none of your existing network can help? You need an exchange.

In other words, exchanges are a decentralized way to make sure that:

(A) problems are identified fast, locally

(B) resources are identified to help them and make a match

(C) experts / voters can score the value of problems and resources,

(D) everyone can see what everyone is doing.

In other words, it’s a way for someone to say “I need help damnit!” and someone else to say “I’m here to help!”... in a way that is more efficient than the batshit way we are trying to get this done over social media or Google docs. It’s the last two points - C & D above - that can really make a difference in the long run.

Why do we need this now and in the future? Because coronavirus is straining all our available networks and breaking networks we didn’t even realize were vulnerable.

Now, some may question this approach and say “Well, would such an exchange be *much* better than doing this on Google docs or Slack? Those are free and easy!”

Good point. It’s probably not better when you are trying to match 20-30 people. Why?

  1. Because exchanges are not easy to build.

  2. The rest of the world probably does not give a damn about the 20 “I Need” Projects and 20 “I Help” Offers that you put together.

However, if you are talking about 20k+ projects, then yeah, a few things happen:

  1. Google docs don’t scale well so you have to build something that can scale.

  2. Other people in the world probably *will* give a damn about what you are doing, since you are doing a lot.

On top of that, there is a risk that you face by trying to do this all in a private Google doc - and that is that no one who knows better than you gets to vote on how smart or stupid your matching is. Should we be making videos to train non-ICU doctors on how to help with Covid-19 patients (since elective surgeries have been canceled)? I’m not sure. Maybe it would be a lot better to create a training video to show grocery store workers how they can reduce risks of contamination. Either way, wouldn’t it be great to be able to see doctors and epidemiologists scoring the utility of these projects before their funding is decided?

In other words, you know that moment in the holiday charity drive when people finally say “Stop giving us canned tuna, we are up to our eyeballs in the damn stuff, how about some nice cashmere coats?

That’s what I’m trying to avoid - especially as everyone rushes forward to help without any coordination amongst them. And I find it excruciatingly painful to watch folks futz around with funding ideas that seem good to them but which may not be needed. We are in a triage situation. We need to triage money also. OK, rant done.

So - let’s go a bit deeper...

If we are going to go down this path, I really think we should build an SDK / platform for exchanges / matching engines so that everyone can customize them for their own specific communities and we can connect them all on the backend.

This is also useful because of: (a) different spoken languages; (b) we can’t even imagine the needs we will have in the future + some customizability may really help with niche communities; (c) matching engines are hard to build but have basic underlying functionality; and (d) we need them on a common platform since being able to collect data from / coordinate across multiple exchanges would be *insanely* important for global responsiveness.

The publishing of data feeds helps make the exchange viral because it makes the exchange itself a source of news about what we are doing as a species.

What are examples of sites or apps that are really just matching engines? FoodCloud, GoFundMe, and TaskRabbit, to name a few. But these are all independent platforms. Can you imagine instead that there was a generic technology so that someone could create a site like FoodCloud (1) super quickly (2) with a lot of powerful tools under the hood, and (3) reporting all the data from their "FoodCloud" to a central network so that we could globally see how, say, 1000 Food Clouds around the world were doing? There could also be 1000 "Mask Clouds" and 1000 "Get a Nurse to My Hospital" Clouds. This is the power of a Exchange network.

So - some things a Coronavirus Exchange platform probably needs:

  • Built with assumption that you will have at least 1 million Needs and 1 million Offers of Resources. Don’t underestimate the scale.

  • It has to be fast at scale. Really fast.

  • Super easy process for inputting needs / projects.

  • Solicitation tools to invite people to input Needs or be Helpers. People in the middle of a crisis may need to be prodded to join.

  • Access control to let someone start an entry and someone else edit / complete it.

  • Ability to visualize *aggregate demand across related projects* - e.g., I need 10,000 N95 masks and have my own project, but my project is tagged to the aggregate N95 mask projects so that we can all see total demand on the platform is 100MM masks.

  • Medical people will flip out about having the right data taxonomy tags. We need *serious CS brainpower on this* - the tagging taxonomy *must be adaptive*.

  • Flexible matching criteria, i.e., “I need to match on some / all of the following”: dollar amount, labor needed, logistics needed, due date, etc.

  • Ranking / Scoring of “I Need” projects by SMEs so resource providers know what the best projects are.

  • SME “expertise” scores and/or peer rankings (maybe pulled from Linkedin with Google Scholar h-score equivalents) so that non-experts don’t flood system and we know who is a qualified SME project evaluator.

  • Conditional matches (I will supply X resources if other people supply Y resources) - *this builds momentum* - you can do *really powerful stuff this way* (we created huge Wall Street transactions based on conditional participation - e.g., I will put in $100mm if you get another $900mm).

  • Full transparency and public data feeds that publish all (i) projects, (ii) resource offers, and (iii) matches - essentially everything that happens on the site is published to everyone - even if they are not a member of the site.

  • You want to design to maximize info entropy in the world (see here) - stock exchanges do this by publishing their data to all the interwebs.

  • You want these app data feeds to be captured in a central command / data hub so that they can be aggregated and published for social insight. For instance, if people use the platform to create a Food Supply Exchange and a Trucking Supply Exchange and a Respirator Exchange, all of these different exchanges should have their data available for centralized distribution.

  • Easily searchable history of everyone who entered a project, resource, or created matches.

  • Everything shareable on social media.

  • Source code is visible to all to allow for collaborative effort / debugging (maybe).

  • Publish traffic stats so we know which apps using the SDK are useful and which are not.

  • Do not focus on encouraging stupid comments, debates, social media-ness - matching engines are a tool - let them be them and let social media be social media.

  • Don’t make it look like GoFundMe where you can only see like 8 projects per page.

  • No big pictures / videos to help market your project or needs (well, maybe within reason). But please, for love of God, do not make this a site for grabbing attention.

Again, *do not confuse this with a social media project* - social media *is terrible* at creating prioritized matching. This is why stock exchanges are not set up like social media sites.

Who might want to use this platform / SDK? I can imagine several different use cases that may need matching. For instance (there may be more than one, but hopefully all using same underlying technology to be mergeable if needed):

  • Food, Medical Supply, Transportation Exchanges

  • Scientist / Hospital Funding Exchanges

  • Peer-to-Peer 911 / Helper Networks

  • Help a local person who has XYZ problem

  • Support a hospital with XYZ items

  • Get volunteers for XYZ emergency in your neighborhood

  • Fund this research project

  • Find people with coronavirus immunity to volunteer for your dangerous mission

  • Raise money for XYZ

  • Find survivors to be interviewed

  • Get help with solving US govt regulations / customs problems

  • Get data processing help for ABC organization

One More Time Why Is This Different Than Social Media?

Here are some additional background from a related proposal:

Who Should Build This?

Well, a team who can work fast and get it up in 1-2 weeks would be great.

Ideally, they would be backed or affiliated with people who are really good at distributing data, like Reuters, Bloomberg, NYSE, Nasdaq, etc. I don't think these company can individually move fast enough unless they set up hackathons.

So - which company has the presence of mind to tackle this?

UPDATE: Coronavirus exchanges have started to pop up. Here is a list as I find them. None are a platform / SDK - if we had that, this would be faster, but awesome anyway.

This Article - 3/28/20

  1. Valence Medical Exchange (here) - 3/30/20 (for profit)

  2. (here) - 3/30/20

  3. (here) - 3/31/20

4/1/20 and after (to best of my knowledge - many from Covid Accelerator):









  9. MutualAid.World (here)


Author Background / Contact

Website and Linkedin.