GG Weekly Podcast Ep. 6: Behind The Scenes of Rootclaim with Founder Saar Wilf
Global Guessing Weekly Podcast

GG Weekly Podcast Ep. 6: Behind The Scenes of Rootclaim with Founder Saar Wilf

Clay Graubard
Andrew Eaddy
Clay Graubard, Andrew Eaddy

In the sixth episode of the Global Guessing Weekly #Podcast we sit down with Saar Wilf, an Israeli Entrepreneur who has sold companies to PayPal, Yahoo!, and Oracle among others. Saar is also the founder of Rootclaim, a probabilistic analysis website that we've covered in previous podcast episodes and a Metaculus Monday's volume.

In our conversation with Saar, we discussed his entrepreneurial background, before focusing on the story and method behind Rootclaim. Later on, we question Saar on certain assumptions made in some of Rootclaim's more controversial analyses, including their one on the Syrian chemical attacks. By the end of the podcast, we make a series of predictions and forecasts on alien life before discussing an AI-driven mosquito-detection start-up Saar's working on which he claims will soon be able to zap 🦟 dead.

Episode 6 of the Global Guessing Weekly Podcast with guest Saar Wilf is now live. Enjoy!

YouTube

Spotify

Podbean


Opening Transcript

[Below is a transcript of the first 11 minutes of the podcast, which has been slightly edited for clarity and repetition.]

Clay Graubard

Welcome everyone to the sixth episode of the Global Guessing Weekly Podcast: The podcast on all things forecasting and geopolitics. My name is Clay Graubard and I'm joined with my co-host, Andrew Eaddy. In this week's episode, we are pleased to be joined by Saar Wilf, a serial entrepreneur and the man behind the website Rootclaim. We've talked about Rootclaim before on this podcast as well as on our website, using it in one of our more recent volumes of Metaculus Monday. We were very pleased to have Saar on our podcast. Saar, welcome.

For those of you who do not know Saar, he is a serial entrepreneur and a company builder from Israel. He's an expert in startups and he has seated or chaired the board of numerous companies, including some of which were sold to blue chip names like Yahoo, Oracle, and PayPal. He has been featured in publications such as Cnet, Vox, Haaretz, and Forbes, and is generally known for driving innovation across industries and geographies. He was kind enough to agree to sit down with us today to discuss his career, and one of his exciting projects Rootclaim. Without further ado, let's hop right into it.

Saar, we would appreciate hearing about sort of your run in terms of seeding and sharing companies. What were you doing before you entered the entrepreneurial space? I read that on top of being an investor and an entrepreneur, you are also a Semi Pro poker player. What was your focus then and what is your focus right now?

Saar Wilf

Before I was an entrepreneur, I was a child. I started my first company when I was 21. That was in the payment space. My second company was called Fraud Sciences that was acquired by PayPal and became what is today PayPal Israel. It's a major part of their AI security operations today. Since then, I've been doing founding companies investing in companies usually like deep technology, AI sensors, computer vision, optics. A lot of interesting companies, and I'm also devoting quite a lot of my time into projects that are not necessarily supposed to be profitable. Could be, but if they're not, that's fine. And companies  I think should have a significant positive impact on our world.

Clay Graubard

And what sort of companies are those? Is that Rootclaim?

Saar Wilf

Rootclaim is a good example.

A lot of it is related. I'm doing a lot of work recently on promoting better treatments, and better analysis for the pandemic, I’m conducting a clinical trial on a few substances that don't have any financial incentive behind them. So...

Clay Graubard

Is vitamin D that sort of thing you're talking about?

Saar Wilf

Yeah, there are a few substances and if you want, we can go into them all. But it's a very interesting issue. Most of the substances in a world that could be effective have no financial incentive to have them explored. Basically, as a society, we said one year ago, we're going to look in solution only in this small space of patented drugs that can be sold at a high price or alternatively, vaccines that can be sold at mass amounts. But there are so many other solutions, and especially just no one is doing it, there was no incentive to do it.

Clay Graubard

Have you done any sort of research and I'd actually like to hear about some of the trials you're doing there, have you done any research into fluvoxamine? I mean, it was featured in 60 Minutes recently, it's a super cheap antidepressant that they found was really good at preventing severe diseases in COVID-19. Is that this type?

Saar Wilf

That's one of the subsystems that seemed to be very promising. It's not an our trial, my trial is only focused on very  safe substances that are not regulated at all. Basically stuff is classified as nutritional supplements. I know there are people doing research on many other substances. But yeah, that's a good example.

Clay Graubard

What brought you to want to focus on doing that type of research? I assume you don't have a sort of background? Is it out of public wellbeing in a pandemic, trying to find sort of really cheap and safe solutions? What was the driver behind that?

Saar Wilf

I did make one investment in a pharmaceutical company, but other than that not really my world. It’s something I've been studying since I was a child, but it’s never been serious work. During this pandemic, it became a major issue for everyone, I started looking into how can I help? How can Rootclaim help? How can the assets, the capabilities we developed, be helpful even if there's no financial benefit?

It appears my dog is snoring. Do you hear that? You want me to wake him up? I mean, you don't hear that?

Clay Graubard

I mean, only if we can see him.

Saar Wilf

Can you see that?

Andrew Eaddy

This will be a great clip. He looks so comfortable. That's awesome.

Clay Graubard

What's his name?

Saar Wilf

Leo.

Clay Graubard

Leo, adorable. That is the first pet on the Global Guessing Weekly Podcast. That is the first dog on the podcast. I'll have to go run off and get Hector upstairs.

Andrew Eaddy

Your projects have spanned across a lot of industries and for the viewer's own edification they span even more than this. You have PointGrab, which is in the workplace optimization and facility management space, InitiativeQ in the payment network space, Bzigo is in the mosquito detection and location space, and as you mentioned, there are sort of common threads with optics and AI and machine learning, but how did all these different projects? What got you into the forecasting space to begin with?

Saar Wilf

Rootclaim has its origins before most of my other companies. It's something I've been researching for over 10 years. It's a combination of my general interest in global problems, global challenges, and my interest in machine learning, statistics, probability theory. At some point, these two passions started to converge and I thought there could be a lot of value in helping to understand better through probability theory, better mathematical modeling of complex issues.

Andrew Eaddy

And just for our viewers, would you mind just giving an elevator pitch, overview of what Rootclaim is and how it works? I think a lot of our viewers are familiar with Metaculus and some of the other sort of crowdsourcing forecasting platforms, but Rootclaim is very unique in the way that it approaches answering certain questions, and I think they are definitely interested to hear about that.

Saar Wilf

We don't really look at Rootclaim as forecasting, and the crowdsourcing element is pretty negligible. Rootclaim is more about a methodology in which you take a factual question with a few hypotheses, and you want to examine them. The hypotheses need to be exclusive, distinct, and small in number.

It's more about a methodology in which you take a factual question: Something where you have evidence and there's a few hypotheses that are exclusive, distinct, and small in number. As a basic example, it could be either Person A is the murderer, or Person B, and you want to see how the evidence supports each hypothesis: What is the likelihood of each hypothesis?

These are the kinds of problems we are very good at. We don’t do forecasting. We don’t do anything where the space is very varied--where you could have thousands of possible answers--or anything that involves dynamic systems you need to predict. Only in cases where there's a distinct list of hypotheses and a list of evidence that you want to tidy them up and see how they affect each other. That’s the goal. It could be famous murder cases, cases in wars and political disputes, it could be medicine, or an issue like what is the source of a virus. Anything that has a distinct hypothesis and can be analyzed this way is what we take.

That approach is to try to map all the evidence and the relationships into a probabilistic model. The probabilistic model itself is not too complicated--it is purely mathematical. It is a formula. There is no machine learning, no training, no algorithms. It is not something that can be improved on. Once you decide that these are the inputs, there is only one mathematical solution.

The question is how do you map the world into the inputs of the formula in a way that is as accurate as possible--and that is most of the work. There is a lot of work on methodology here on how you go through the hypotheses, how you find them, how you find evidence, how you find the dependencies between evidence, how you find the probabilistic effect of the evidence on the hypothesis. It is a lot of that methodology that is able to be very unbiased, very accurate, very comprehensive, and get results that are much better than any human can do by themselves.


Continue the podcast on Youtube, Spotify, or Podbean. Thank you!