Any Humans Assessing Images Today will be Augmented in the Future
/Zavain Dar is a Partner at Lux Capital. He invests in companies that are using machine learning and AI to augment or replace physical-world functions including biology, language, manufacturing, and analysis. He looks for entrepreneurs that can use software and data to hone a philosophical position on where the world is, and how to direct it for the better.
Zavain was sharing his knowledge on trends and investment opportunities in visual technologies as a panelist and an ECVC judge at the 6th Annual LDV Vision Summit.
Evan Nisselson, General Partner at LDV Capital asked Zavain some questions about his experience investing in visual tech.
Evan: You have entrepreneurial experience as a founder and computer scientist before joining Eric Schmidt’s Innovation Endeavors and now a Partner at Lux Capital. You were early at Discovery Engine which was acquired by Twitter and also a cofounder of Fountainhop, one of the first hyper-local social networks. Which aspects of your expertise do you believe helps you empower entrepreneurs to succeed and why?
Zavain: There's no singular answer here. At various times in my venture career, I've drawn across all moments of prior experiences. Not just the "LinkedIn highlights" but perhaps less intuitively, the lowlights: grinding it out working odd jobs as a teenager to cover tuition for private HS while living with a single mom who herself was a graduate student; getting laid off from my first job (door to door salesman for solar panels in high school); my work advising the "Trust the Process" era 76ers where at the moment we were just about public enemy #1 in Philadelphia and on many vanity metrics arguably the worst team in the league. Seeing not only the ups but knowing how to traverse the inevitable downs whilst keeping a cool calm head speaks volumes to most entrepreneurs. As an investor, your energy and calm (or lack thereof) can often be contagious to your entrepreneurs and learning to both be aware and in control of that is just as important (if not more!) than any prior experience or axis of expertise itself.
Evan: Lux Capital and LDV Capital are co-investors in Clarifai. What inspired you to invest in Clarifai and why?
Zavain: I'd known Matt since 2013 shortly after he single-handedly won ImageNet and had just started Clarifai, and while I was still at Innovation Endeavors. Tracking Matt over the years gave us a keen sense of confidence in his unique ability to not only build world-class technology but build and lead a team. In a lot of ways, Matt's been emblematic of many of our entrepreneurs where he's really seen success in everything he's put his time and energy into (I'll have to share more about him scoring 70 points in an HS basketball game another time). We've learned to see historical horizontal success as a leading proxy for future entrepreneurial success. Thematically, I don't think anyone can argue our computers in the future won't be able to see and understand what's around them. With all that's going on in and around privacy with the large tech incumbents the question for us was could there be an independent franchise that could truly deliver the best of class technology and products while retaining neutrality and customer protection against larger tech hegemony. Around our partnership, we were an emphatic yes.
Evan: You invested in Recursion Pharmaceuticals and their co-founder/CTO Blake spoke at our 2017 LDV Vision Summit. We recently published an in-depth report on Nine Sectors Where Visual Technologies will Improve Healthcare Over the Next 10 Years. There are many visual technology companies working to develop drugs. How and why did you choose to invest in Recursion Pharmaceuticals?
Zavain: Venture's a game of balancing long-tail serendipity hunting alongside acute deep dives to build theses and strong POVs to have a prepared, informed, and agile mind when opportunities do arise. Frankly, we got exceptionally lucky with the process on Recursion. I'd been teaching a seminar at Stanford for a number of years at the intersection of AI and Philosophy largely on how philosophical shifts underlying our practice of AI were redefining how we put to action the Scientific Method itself. In the past, I've playfully called this "Radical Empiricism". When I first met Recursion founders Chris and Blake in 2016 within 5 minutes I knew I wanted to be partnered with them, as what they were working on was an almost perfect instantiation of what I'd been talking about in my seminar at Stanford. You'll have to remember that in 2016 this wasn't a hot space and applying computer vision and deep learning to assess the morphological changes of various perturbations to myriad human cell lines wasn't an obvious idea. "Would there be enough signal to noise?" "Why hadn't others done it?" "Cell screening has been around for years, how is this different?" Over the last few years, we've been able to answer all of these questions from what I call "informed skeptics" with increasing conviction based on data and tangible preclinical and clinical results. Candidly I don't know of any other companies then that were applying computer vision to drug discovery as a foundational substrate to a larger platform, but I suspect Recursion's ability to show exceptionally high signal to noise from its approach has encouraged a handful of upstarts to join the fold.
Evan: In the next 10 years - which business sectors will be the most disrupted by computer vision and why?
Zavain: The above on computer vision in and around biotech I think still holds true in spades, and more broadly any industry that has historically relied on humans for any sort of image analysis should be empowered and augmented with computer vision. Technologically it's no longer controversial to say computers can "see" across more data and per datum quantify a larger number on nonlinear features and relationships. That is computers can both see more per image and see across more images, so it's reasonable to assume any humans assessing images today will be augmented in the future. From a business perspective, this touches everything. I candidly spend more time thinking about how we'll integrate computer vision across the nuances and workflows of heterogeneous industries in ways that are sympathetic to problems unique to individual industries but productize across various industrious while retaining efficiency with scale.
Evan: LDV Capital started in 2012 with the thesis of investing in people building visual technology businesses and some said it was “cute, niche, and science fiction.” How would you characterize visual technologies today and tomorrow?
Zavain: Visual technologies (from self-driving cars to facial recognition for mobile security) have undeniably arrived. The question is longer "are these science fiction?" or "is it possible?" but rather what will the path to pervasive distribution look like and commercially who will the winners be?
Evan: You frequently experience startup pitches. What is your one-sentence advice to help entrepreneurs improve their odds for success?
Zavain: Take time to understand the investor, their portfolio, their areas of interests, and their particular lens on the world; for any company in their portfolio try and understand how that investment came to be and how your path could compare.
Evan: What are you most looking forward to at our 6th LDV Vision Summit?
Zavain: Learning what everyone else is up to, what problems everyone is facing, and how different teams are finding creative and disruptive solutions.