This fall, Radical Ventures launched the AI Founders Master Class, a four week program aimed at helping AI researchers build AI products and businesses. Over 200 students and faculty from the Vector Institute for Artificial Intelligence, the Alberta Machine Intelligence Institute, Mila (Quebec’s AI Institute), Stanford University, and Oxford University registered for the four week program. This week we are sharing an excerpt from one of the sessions.
The final week of the program focused on narrative development and mastering the founder’s pitch. Radical Ventures Co-founder and Partner Jordan Jacobs, had a candid conversation with Canadian entrepreneur, Co-founder and President of Clearco, and Dragon’s Den judge Michele Romanow.
The following excerpt has been edited for length and clarity.
Jordan Jacobs (JJ): For the past three weeks we have been talking with a group of incredibly talented AI researchers about how to make the transition from research to founder. Let’s start there. What drew you to becoming a founder?
Michele Romanow (MR): I graduated engineering school and figured out that the worldwide supply of sturgeon caviar was down by 95% because the world had overfished the Caspian Sea. I was crazy enough to graduate, move to the east coast and build a fishery from scratch. It was everything it sounds like: boats, fishermen, my hands elbow-deep in fish, the whole nine yards. That was my initial immersion and it was a great first business. It failed three months later as we went into the giant 2008 crash and I found myself being 21 years old selling the world’s most unnecessary luxury product.
JJ: Reflecting on that first pitch, when you were looking to get into the caviar market, what would you have wished you knew, or what would you tell yourself if you could go back and give yourself advice?
MR: There are a few things that I have learned that may not be as obvious as you think.
First, you have to have a pitch that has a vision. We first started pitching Clearco as building a bank for the freelancing gig economy and our feedback almost universally was that our idea was not big enough. I had never actually experienced exponential growth myself and so I could not speak to that with a lot of credibility. Thinking about how big the world can be is a really important part.
Second is creating a business model that is a win-win on both sides of the transaction. For example, I invested in the Airbnb of RVs. It was the simplest pitch to understand. There are two million RVs in Canada that are used two weeks of the year on average. You also have people who want to use an RV for only two weeks, but cannot afford to own one. Pairing these groups together, every transaction benefits both sides. These types of business models tend to be monopolistic.
Finally, founders have to lean into who they are and why they have the grit to succeed. Those are always the founders that produce alpha.
JJ: You have had an incredible career building many different successful businesses, plus being an investor in the most high profile way possible, where people can see your work in real time. What are you looking for in an investment?
MR: The first thing that draws me in is the founder’s story and looking for evidence of overcoming some difficulty in their life. Because, in the beginning, businesses need to pivot many times. For instance, at Clearco our Seed investors invested in a bank for freelancers, our Post-seed investors invested in a bank for Uber drivers, our Series A investors invested in a company that backed Airbnb hosts, and our Series B investors invested in what will be the largest e-commerce investor. We are now on that path, but companies pivot. This is not unique to Clearco. The core of every single business is figuring out these pivots. And so, in the early stage if you are overly consumed with market size and unit economics, there’s a chance that you may be missing a bigger point. You need founders that are smart enough to see the data and know that something is not working and to understand they need to pivot and have the guts to do so.
This idea of ‘total addressable market’ can be a misleading indicator for investors. My favourite example is a research document from an investor in Shopify that concluded there would only ever be 40,000 companies in the whole world that would ever need e-commerce software. I always say that if McKinsey has written a report about the market then there are already players worth billions and you have missed the opportunity to build that market. You need to be looking at markets that seem very small today. Looking back, the best decision I made was choosing to go into a fintech company in 2015.
In 2015 there was no Stripe and the existing companies in the market had no scale or size to indicate that fintech was going to be the next big space. You have to be OK going into an industry when it is in a bit of a trough and when people do not quite understand the value proposition. For example, when it comes to AI, people may not understand that this technology will impact every industry. You need to be looking for early litmus points, not just the hottest thing.
Next week, we will share Michele’s point of view on building a company and what it takes to succeed as a founder.
AI News This Week
-
Opinion: AI is about helping companies grow and create jobs (Globe and Mail – subscription required)
Radical Ventures Partner and former TD Bank Group CEO Ed Clark discusses the present opportunity for AI in small and medium-sized enterprises (SMEs.) Ed underlines the importance of supporting SMEs adopting AI, as these businesses will play a critical role in Canada’s economic recovery after the pandemic. Ed Clark is chair of the Toronto-based Vector Institute for Artificial Intelligence, which was co-founded by Radical Ventures Co-founders Jordan Jacobs and Tomi Poutanen, Geoffrey Hinton, amongst others.
-
Advancing mathematics by guiding human intuition with AI (Nature)
Mathematics still holds unsolved puzzles (or ‘conjectures’): ideas humans believe to be true, but for which solutions are as yet unproven. AI’s powerful pattern recognition abilities may have uncovered two discoveries in mathematics. A paper resulting from a collaboration between DeepMind researchers and mathematicians identifies a new formula in representation theory and a previously unknown relationship that helps us understand the mathematics of knots. Knot theorists proved the validity of the knot formula after machine learning guessed what the formula should be. Computers assisting mathematicians in discovering patterns and formulating conjectures is nothing new – most famously demonstrated in the Birch and Swinnerton-Dyer conjecture in the 1960s. This new paper illustrates the usefulness of machine learning in identifying patterns in data. AI adoption in pure mathematics would integrate machine learning into existing mathematical workflows and could lead to discoveries that may be beyond human comprehension.
-
Listen: AI Robotics alongside us today (The Robot Brains with Pieter Abbeel)
Pieter Abbeel, Professor of electrical engineering and computer science at the University of California, Berkeley and Co-founder of Radical Ventures portfolio company Covariant, launched the Robot Brains Podcast earlier this year. The podcast explores how far humanity has come in its mission to create autonomous robots and ranks in the top 1% of all global podcasts. In the first season, Pieter sits down with other renowned AI robotics researchers, professors, and entrepreneurs building robots that can see, learn, think, and react to their surroundings. In a snapshot of the first season, we see how far society has come in building machines that improve our lives. The conversations include robots delivering medical supplies to remote places, automating recycling, handling dangerous, repetitive warehouse tasks, running indoor farms, and helping nurses on the frontlines. Season 2 launches in early 2022.
-
An AI finds superbug-killing potential in human proteins (Wired)
A University of Pennsylvania research team has found 2,603 antibiotic candidates reported in Nature Biomedical Engineering. Along with vaccines and clean water, antibiotics are one of three ‘pillars’ that have let humans double our life span since the 1800s. The World Health Organization estimates that by 2050, 10 million people could die annually from drug-resistant infections as the efficacy of current antibiotics wanes. The researchers noted that the discovery speaks to AI’s ability to digest massive datasets. The research is the first time anyone has explored the human body for antibiotic candidates to this extent. By using AI to guide the search, the team stumbled upon another discovery: many of our proteins that are seemingly unrelated to immunity may have evolved to live double lives as protection against invaders. Read the full paper.
-
Electric cars are powered by rare metals. Can AI help find them? (National Geographic)
According to the International Energy Agency, to prevent the world from warming by 2 degrees Celsius (and to meet the goals of the 2015 Paris Climate Agreement) will require unprecedented production of commodities like copper, cobalt, nickel, and lithium – essential materials in the production of green energy infrastructure like batteries for electric vehicles. Geological models powered by AI may better predict where ores have formed, which could reveal deeper deposits of these rare minerals.
Radical Reads is edited by Leah Morris (Senior Director, Velocity Program, Radical Ventures).