Judah Taub of Hetz Ventures on Pushing the Boundaries of AI

Without realizing it, we humans often climb the wrong mountain, and later in our lives find it hard to move down. The facet of avoiding Local Maximums is one that humans are particularly poor at, while AI has learned to make decisions in such a way as to avoid the trap.

Artificial Intelligence is transforming industries at a breakneck pace, and the entrepreneurs driving this innovation are at the forefront of this revolution. From groundbreaking applications to ethical considerations, these visionaries are shaping the future of AI. What does it take to innovate in such a rapidly evolving field, and how are these entrepreneurs using AI to solve real-world problems? As a part of this series, I had the pleasure of interviewing Judah Taub.

Judah Taub is the Managing Partner & Co-Founder of Hetz Ventures. Before founding Hetz, Judah was Head of Data at Lansdowne Partners ($20B London-based Hedge Fund) as well as advising multiple young start-ups. Judah was also elected as one of Forbes 30 Under 30 for 2020. In his service in the Israel Defense Forces Judah served as an officer in a classified intelligence unit where he engineered a large-scale project to win the IDF 2014 Creativity Award. He has lectured widely, including throughout the IDF and at Wharton Business School, on time management and creative thinking, and wrote a book for new IDF soldiers published by Yediot, Israel’s leading publisher.

Thank you so much for joining us in this interview series. Before we dive in, our readers would love to learn a bit more about you. Can you tell us a bit about your childhood backstory and how you grew up?

Sure, happy to partake in this conversation. I was born in London, and at the age of one, moved with my family to Israel and grew up in Jerusalem, a diverse and dynamic place with a lot of personality and opportunity to try new things. Even in my school years, I always liked inventing things, and there were many failed attempts along the way. Examples include trying to create perpetual motion with magnets, and a metal detector which my best friend and I assumed we’d use to uncover lots of treasure in neighboring backyards. Some were more successful, such as a hovercraft I developed and a round sudoku puzzle that I created and sold to Israel’s largest newspaper.

Following high-school and a year studying and volunteering, I joined the military for five years, where I was exposed to people from all walks of life, and involved mainly in working in intelligence. The highlight of my service was innovating a new way of collecting intelligence which subsequently won the military’s top Creativity Award in 2014.

I still deeply appreciate the ways in which informal education — inventing as a teen, gaining perspective in the military — have enriched and enhanced my outlook in ways that formal education like university just doesn’t quite achieve. These experiences have informed so much of my career and the choices I’ve made since.

Can you share the most interesting story that happened to you since you began your career?

In early March 2020, as Covid-19 began spreading around the world, like most countries, Israel was in search of ventilators. At the time I was called back to the military to help support an emergency effort to tackle this issue locally. I was tasked with leading a group of engineers to develop a solution, and with the inspiration and ingenuity of my great uncle Joey, who at the time was turning 80 and had previously worked as both a surgeon and a pilot, we managed to help the military build a new ventilator in a matter of weeks. This ventilator was not only easy to mass manufacture, but could also ventilate up to eight adults in parallel. Luckily, the virus severity didn’t require the use of the ventilator, but the experience was an important one for teaching me about seeing beyond the limitations at our feet and what eventually I began to call Local Maximums, the subject of my book.

None of us are able to achieve success without some help along the way. Is there a particular person who you are grateful towards who helped get you to where you are? Can you share a story about that?

I’m very grateful that my father has been a recurring source of encouragement and example for me. A primary example of this is the very book I recently published, How To Move Up When The Only Way Is Down. After years working in startup building, investment and leading teams, I developed the idea to write a book that uses lessons from AI that non-technical people can use in their day-to-day professional or personal challenges. Humanity spends billions of dollars each year helping AI make better decisions (primarily the big tech companies trying to improve their cutting-edge algorithms), but along the way, some of these insights into decision-making is really relevant to all of us as individuals: what to study in college, what job to take, whether to settle down or continue dating, managerial business decisions and more.

Every few chapters I wrote, I shared with my father, who has even more experience working internationally across communities of people working on societal challenges. Having that perspective enriched the points I make about humans learning decision-making creativity and processes from AI.

Ok super. Let’s now shift to the main part of our discussion. Share the story of what inspired you to start working with AI. Was there a particular problem or opportunity that motivated you?

I’ve been working with startup companies building AI processes, solutions and infrastructure as the Managing Partner at my venture capital firm, Hetz Ventures. We invest in and support very early companies innovating in these spaces, since 2018. It’s obvious that this is the future of many if not most complex industries, even those we take for granted. It’s also clear that our developing artificial intelligence is having clear impacts on the evolution of humans, too. My interest lays in how we can enhance and improve our humanity by learning from the very AI we work so hard to build, train and transform.

Ok super. Describe a moment when AI achieved something you once thought impossible. What was the breakthrough, and how did it impact your approach going forward?

In 2015 I was working in London at a hedge fund called Lansdowne Partners. One of the perks of working there was that we would be visited by really interesting people working on cutting-edge ideas. One day, a man named Demis Hassabis, shared with us how he was leading a team that were training AI to beat the world champion at a game called GO. He described that this particular algorithm was learning the rules in a completely different way to how machines typically are taught sets of rules: Rather like a human, taught through examples. He claimed this type of algorithm training was going to change the world, which back then sounded far-fetchedת but the notion of teaching the algorithm to make its own decisions stuck with me. By the way, Demis Hassabis is now the CEO of DeepMind, which was eventually bought by Google and has led to breakthroughs in medicine, machine learning and many other areas.

Talk about why this change was so pivotal and how it shaped your view.

We used to think of a machine as something we give specific guidelines. In other words, we tend to imagine that there is an enormous decision tree to which you can trace back any decision the machines makes. With AI, this is simply not the case. The process AI follows is very similar to a human. We learned a cat is a cat, not because we were taught to cut off a bit of its DNA and wait for a lab result, but rather by seeing many cats and other animals that are not cats. Engineers have learned to train AI in a similar way.

This has huge implications. One is particularly important but often overlooked. As AI is not a simple set of rules but rather is learning to make better decisions based on what it has learned, this means that we humans can learn from the way it operates to make our own decision making better.

For decades, we humans taught computers logic; now, we can benefit from learning from computers as well!

Can you share an example of how AI makes decisions in a unique or different way to humans?

There is a trap in decision-making which I call hitting your ‘Local Maximum’. Imagine you are in a desert and your goal is to climb to the highest point in the hilly terrain. You don’t know which mountain is the tallest, but see one that looks promising, and so you climb it. As you reach the top, you realize you climbed a tall maintain, but it is not the tallest. There is a much taller one behind it. Now you are at a Local Maximum: you can’t move further up, as you’re at the highest point of the mountain you’re on currently, but you’re not at the highest or most optimal point for your goal. Your only option to progress even further is to go through the pain of climbing the whole way down, only then to climb the highest, and correct, mountain.

Think of a marketing executive who is paid well but ultimately doesn’t enjoy their field. Does she start over again in building a new career to achieve professional satisfaction? Or a sports team that includes one brilliant player but by playing him every time, the rest of the team’s growth is limited.

Without realizing it, we humans often climb the wrong mountain, and later in our lives find it hard to move down. The facet of avoiding Local Maximums is one that humans are particularly poor at, while AI has learned to make decisions in such a way as to avoid the trap.

Can you give an example of where AI does a better job than humans at avoiding Local Maximums?

In business and elsewhere, humans have learned to utilize a decision-making process known as A/B testing. The process involves comparing two similar options testing which gets a better result and then iterating again between the better option and a slightly different version of it. In fact, this process has become so prevalent across marketing campaigns, research labs, development studios and other areas, that it is often taught in university.

Although the process is really suited for a computer to follow, advanced algorithms have learned to alter the process further so that it learns to avoid Local Maximums. Rather than doing A/B testing, the algorithms operate an A/B/X process. While A and B are very similar to each other with one or two minor feature differences, X is a radically different option, forcing the machine to continually consider dramatic changes and new paths previously ignored. Amazon, Google, Netflix, Walmart, Nividia, Apple and all major tech companies worth their salt have incorporated Xs into their A/B testing as they recognize it helps them avoid getting stuck in Local Maximums. We individuals should learn to add Xs into our own decision-making, whether professional, personal or otherwise, as well.

When you think about the future of AI, what excites you the most, and how do you see your work contributing to that future?

AI has the potential to dramatically increase our productivity while also level the playing field for many communities or demographics that haven’t benefited much to date from the tech boom. However, the book focuses on a deeper level on how we can learn from AI so that we make better decisions ourselves.

Not so long ago, the greatest chess players learned and improved their strategy by reading books about extraordinary game play and record-breaking, historic moves. Then, one day, chess grandmaster and world champion Garry Kasparov lost to Deep Blue, a chess-playing expert system run on an IBM supercomputer. Yes, the newspaper headlines were focused on Kasparov losing the throne to lines of code, however there was an even deeper change that took hold: chess players from all over the world started improving their game by using software. Even today on free mobile apps, a player can ask the software to recommend what move he should have considered. Do the chess players always listen to the software? No, of course not. Sometimes humans can think of the better or more creative move. But even chess grandmasters will attest to the fact that using software to improve your game is a key tool in the training toolkit.

The same can happen with our daily decisions and activities. Do we always need to listen to the software? Of course not. But not asking what an AI algorithm would suggest is simply sub-optimal, and a wasted opportunity to improve overall.

Is there a person in the world, or in the US with whom you would like to have a private breakfast or lunch, and why? He or she might just see this, especially if we tag them. :-)

I’d love to have lunch with Adam Grant and Malcolm Gladwell. When I started writing the book, it was very technical — it even included lines of code. Some of the Harvard professors I worked with challenged me to share these insights in a way that non-technical people will not only understand, but enjoy reading. As part of this process, I reread most of both Grant’s and Gladwell’s books as I think they are both masters of teaching and sharing insights through captive storytelling. In many ways, even though I have never spoken to either, both played a part in inspiring the style of my book.

How can our readers further follow your work online?

Check out www.judahtaub.com/ for further insights and content, and you can find the book in various formats at Amazon.

Thank you so much for joining us. This was very inspirational, and we wish you continued success in your important work.