How to Make Moonshots into Moon Landings & Other Tales in Deep Tech Entrepreneurism
/Fireside Chat: Dr. Mary Lou Jepsen, Founder & CEO at Openwater, and Evan Nisselson, General Partner at LDV Capital
LDV Capital invests in people building businesses powered by visual technologies. We thrive on collaborating with deep tech teams leveraging computer vision, machine learning, and artificial intelligence to analyze visual data. We are the only venture capital firm with this thesis.
In 2014, we started LDV Vision Summit – a premier global gathering of people in the visual tech sector. It’s an annual event where participants can meet the world’s brightest technology innovators and learn firsthand how their visions will transform businesses and society.
At our 8th annual LDV Vision Summit we virtually welcomed hundreds of attendees from 15+ countries. We had a fantastic lineup of speakers who covered the latest advances in computer vision, computational imaging, machine learning and AI in healthcare, web3, climate, space, and more! We had the privilege of hosting a fireside chat between LDV's Evan Nisselson and Dr. Mary Lou Jepsen, an impressive serial entrepreneur recognized by Time Magazine as one of the 100 Most Influential People in the world.
If you missed our 8th Annual LDV Vision Summit, this is your chance to watch this interview. You might also want to check out our 5-question interview with Dr. Jepsen that was published leading up to the Summit. This interview was edited for length and clarity.
Dr. Mary Lou Jepsen is the founder and CEO of OpenWater, a startup working on novel imaging systems that leverage lasers as well as optoelectronic & holographic systems to see deep into the body and brain with the detail of a high-resolution 3D camera. In other words, the OpenWater team is building a new generation of portable, real-time neural diagnostic and therapeutic systems.
With a Ph.D. in optical physics, Dr. Jepsen has previously served as an engineering executive at Facebook, Oculus, Google[x], and Intel in addition to founding 4 startups including One Laptop per Child where she was CTO, chief architect, and delivered the $100 laptop to mass production.
Her startup CEO experience includes the world’s only fabless display screen company which was based in Taipei. She has been a professor at MIT and is an inventor on over 100 published or issued patents in the last 5 years alone.
Evan: As a younger woman you loved math and art. You grew up on a farm. How did you get into wanting to be a Ph.D. in optical physics?
Mary Lou Jepsen: When you grow up on a farm, you have to figure out how to fix things, like a tractor, with nothing. You have to pull that tractor out of the swamp or plow the... whatever the thing is that I had to do. I got practical. My dad was an auto mechanic, and I also learned how to fix cars before adding fixing computers into the mix. I loved optics. My parents wanted me to be able to get a job so convinced me to give engineering a try in college. I fell in love with art too and took enough classes to get a degree in art as well. This is at Brown and RISD in the mid- '80s.
As a freshman in college, I fell in love with making holograms with lasers and big optical tables and things that capture not just the intensity, but the phase of light so that you get these magical things. When the MIT Media Lab was opening up, they had gotten one of the best holographers in the world as a professor so I applied and got in and found myself in graduate school at MIT.
Evan: Was that when you unveiled the first holographic video system?
Mary Lou: As a master's student, I co-created the first holographic video system. If you remember in Star Wars, R2-D2 projects out, Princess Leia said, "Obi-Wan Kenobi, you're my only hope." We did that for real! That was a special effect in the movie because you're seeing it on a movie screen in 2D, but we did that in true 3d in thin air! It was computer generated, the actual interference structure on a Connection Machine massively parallel super-computer. I made a screen with one-micron pixels in the '80s. It was hard, it was impossible, but it worked.
Evan Nisselson: One of the biggest challenges a lot of people have when they want to start a business or businesses — and it’s a long journey, is that they think things might happen sooner. In 1989, did you believe that holography would be mass adopted much sooner? It's been about 33 years since then and yes, we're now making more progress in augmented reality but it seems like it's taking much longer than people thought.
Mary Lou: It was a big step in augmented reality. What we have with the Oculus headset is like a Jaron Lanier headset from VPL that actually works!
Two people demonstrate the EyePhone system, developed by VPL Research, three decades before Oculus introduced its $599 Rift VR headset.
Photo Credit: AP Photo/Jeff Reinking
There's a lot we can do with augmented reality. Do we even have to put anything on our faces? Why can't we just exhale the screen or why can't the screens be on every surface and come to life in full 3D?
During the pandemic, I've been working on a project that I call “my Venice Biennale” project. I was thinking, why can't everybody have a million-dollar view? I happen to live in a houseboat in Sausalito, and I have this fantastic view. Why can't everybody have that? It's a small thing but there's so much more that can be done and that has been done, but there's a new generation that's come into it.
There are four big problems for VR:
the content,
the interaction,
the actual headset or however you get the image – you can do it through VCI, then you don't have to wear anything!
the sensing in the room.
There's been a lot of progress on all four of those.
Evan: Another one of your unbelievable earlier projects as an entrepreneur was your One Laptop per Child initiative when you were at MIT. You were telling people that you were going to build something that nobody believed. I've heard stories of you talking about Bill Gates, Steve Jobs, Michael Dell, and Craig Barrett saying it was impossible. What went through your head when people were telling you it was impossible? It's probably a recurring theme for you as an inventor. I've heard many times that I'm absolutely crazy. How do you deal with that? Tell us about that one example of how you delivered and how it became a thriving business, helping kids around the world.
Mary Lou: They didn't want it to happen because it threatened their business model.
The first time I gave a public talk ever (not the laptop but something in holography), somebody stood up - who a lot of people in the room thought should have won a Nobel Prize, that's how great his work was - and told me it was impossible and it'll never work. It felt like it went on for 15 minutes. It was probably just a few.
I went up to him at the cocktail party that night and said, "You've done all these impossible things! Can't you let it hang in the air, think about it for a bit and tell me an actual reason it won't work?" To his credit, he sat down with me and worked through the problems, and became a friend and collaborator. Of course it wasn’t impossible and I got that project to work in the end.
So when Intel and Microsoft say it won’t work… it might be because you're not using their chips or their software. In the early 2000s, laptops cost ~$2,000 plus the ~$1-2,000 of software on them. Those companies knew how to sell something at that price. If we were successful, that was a threat because what we were trying to do was make something that could go to every kid in the world at the cost of textbooks and their software and chips were too expensive or power hungry to use for this purpose at that time.
Today, 100% of students between the ages of 6 and 15 who attend public schools in Uruguay, along with teachers and students from private schools in low-income areas, have a connected laptop and free access to the internet thanks to Plan Ceibal and the One Laptop Per Child education program.
I'd never designed a laptop before, Even though I had been a group CTO at Intel until the end of 2004. I co-founded the first liquid crystal on silicon chip company and was it’s CTO as well. We pioneered a new kind of display technology that was used in the early VR and AR headsets in the '90s and also wristwatch-video, and early large screen TVs, and lots of things like that. We were making the fundamental core technology for that so I knew a lot about making stuff at scale. Electrical engineering was my first undergraduate degree. And I knew software too - I had been a CS professor previously as well.
People in hardware didn't usually consider the screen in architectural choices - even though actual laptops were enabled by the screen… Remember, we used to draw a computer as a big CRT? And then it got flat. It enabled a true portable. That's this huge innovation that people don't realize has happened in the 1990s. The screens got thin, but they were still the most expensive component in any portable. They still are today, and they're the most power-hungry - a big deal in a portable. I approached it as a screen designer. I designed the $100 laptop from the screen backward so that it could also be low power. Why: My customers, the children of the world, at least half of them, lacked steady access to electricity. I designed a different type of laptop that had a retinal resolution on it before Apple coined the term, the lowest power laptop ever made. Still is. You can hand-crank it to charge it up. It wasn’t a stripped down laptop - it was a very clever architecture.
The One Laptop Per Child laptops can be disassembled and repaired using only a screwdriver.
It has a sunlight-readable screen that is color and a motherboard which turns off when no pixels are moving on the screen and back on so fast it’s perceptual to the user - these two things lowered the power a lot. We shipped the first of these laptops in 2006, the beta ones, which are still working in the field and they're fixable by the kids. We went to mass production in late 2007.
If you can have a different perspective, maybe from an adjacent field, you can be disruptive if you come in with a set of expertise.
I partnered with my much more famous co-founder, Nicholas Negroponte, and we were able to get attention, lots of it. I was invited to go meet with the chairman of a large consumer electronics company. And what would usually happen... well this sort of thing! I'd walk in and the chairman of this gazillion-dollar company would be flanked by four EVPs on one side and four EVPs on the other side. It was just me on the other side of the room. They started laughing: "$100 laptop. Ha, ha, ha." The women in the audience will get this: I pulled out a notebook and asked, "Tell me why it won't work." and I dutifully started taking notes for the next ~hour and a half. Then I said, "Okay, you've listed 23 reasons why this won't work. We’ve handled 16 of these but these new 7 are good. I'm going to take these back to the team." The team then was just me, but they did not need to know that.
You pick any Asian city – 60 years ago they were very poor. These guys know (and they were all guys) how computing changed their lives, bringing them and their countries out of poverty. They weren’t opposed to this happening. In between the lines I could see that they wanted it to work and were explaining what they needed to say yes. I started reaching out and going to people who could debug my architecture because they could think through all of the problems before spending tons on prototypes. I fought the urge to make a prototype till I'd done more of that - this by the way breaks all the rules of engineering..! Eventually, I made a prototype that Kofi Annan unveiled at the World Summit on the Information Society. After that, every head of state in the world wanted laptops en masse in their countries, because we had solved this complex multi-dimensional problem by getting the feedback of the most accomplished people in the area.
Evan: It was a classic example of a moonshot that landed!
There are two different “you are crazy” though. One is, that they're saying you're crazy because they potentially have an ulterior business objective, and they're going to say whatever they want to say to benefit themselves. But then there's also “you are crazy” that covers the technical aspect like they are saying that's not possible. That’s what I've come to accept over the years and that’s where our sweet spot in visual tech investing is. We love when during diligence, other people that we talk with say, "That's technically impossible to do, but if it works, it's going to be a huge opportunity if that team is the right team at the right time building the right solution.” This “you are crazy!” doesn't scare me anymore, but it scares a lot of people. How do you deal with that statement of, "Hey, it's not going to work”, but you have the vision that it's going to work?
Mary Lou: Let's face it, the people that rise to the top of the most powerful tech companies today - the ones that get to make the big decisions - are good at optimizing click-through revenue for ad sales, for example. A single vice president can kill your program because they don't understand it, and their intuition is good at the thing that is the core part of the company and you're doing the new part of the company...
It's harder to do a startup within one of those companies than to just do a real startup with 1/100 of the funding. At the levels I can get to in the big tech companies, I get a lot of funding but it's easier and I can go faster and more directly to the problem with a startup structure.
There are technical details in all of that and if you don't understand them... It's fine to not know something.
Evan: Just like that example when you were taking notes with all the reasons why they said it wasn't going to work, or talking to that person who stood up and said, "It's not going to work" at an event. You left Oculus in 2016 to focus on Openwater. Tell us about your mission and where you're at with Openwater.
Mary Lou: I love VR and I did a lot of great things in this space for decades. I believe in other things that are coming out. Finally though, circa 5-6 years ago all of this talent was going into VR and AR and I think, "Great, they're going to do it. They're going to do it and make it great. I wish they'd go faster and waste less money, but it’s going to happen now.” What was being missed was too large though for me to let go of.
I didn't think anybody else saw that with the advances in components and manufacturing processes developed for next-generation high fidelity, virtual and augmented reality, plus LiDAR development. I didn't think anybody saw diagnostics and therapeutics in healthcare could be transformed.
Our body is translucent to infrared and also sound. In the case of Infrared light that's why the invention of fire was so important. You sit around and it warms you because the fire is infrared light. It goes through your body. It warms your belly and ribs.
Openwater’s novel & interdisciplinary approach is possible now due to silicon advances from the steady march of Moore's Law + slipstreaming on LiDAR and AR/VR advances. Image credit: Openwater
The other thing that happened was a big change in camera chips. I knew they were going into smartphones in a couple of years because I was running advanced consumer electronics at Facebook and Google. The key thing that nobody was using is the fact the pixel sizes were approaching the wavelength of light. There's all this discussion of Moore's law, and that's a disconnect in Moore's law, in what you can do in physics because then you can capture the waves of the wavelengths of light or the phase of the wave. That is fundamentally an information theory thing. That's what the invention of holography was initially. It’s how you capture not just the intensity of the light, but the phase through interference. I thought that we could make holograms of our bodies and transform healthcare cost and access using the manufacturing processes developed for next generation AR/VR and LiDAR.
Evan: This connects the early days!
Mary Lou: That and having a brain tumor, which we skipped over. An MRI saved my life in 1995. I was going to go home and die... Nobody knew I had a brain tumor. Somebody sprung for the cost of an MRI and I was happy to have a diagnosis. When you have a diagnosis, you can solve it. I have to do all kinds of things every day to this day to care for my brain, but I've made myself better. I did biohacking before it became a thing. I had to just to live.
The thing about it is I look at those MRI machines today - they're the same cost, the same size as they were 27 years ago, when they saved my life. That's crazy! Could you imagine that in consumer electronics?
What do you do with the two-ton magnet? I mean, a lot of people have tried. There's a company called Hyperfine, they're making it a little smaller, but the resolution is not good (Editor’s note: Dr. Matthew Rosen, the co-founder of Hyperfine, gave a keynote speech “Building an MRI Scanner 60 Times Cheaper, Small Enough to Fit in an Ambulance” at our annual LDV Vision Summit back in 2017). There are some trade-offs. I thought it was time to look at what we could do using the trillion-dollar manufacturing supply chain to make healthcare both better and more affordable. The pandemic of course accelerated decentralized healthcare. Why can't we have these devices at home to be used for diagnostics and therapeutics, without drugs? I don't mean to dismiss how you're using your cell phone for digital therapeutics – that's important, but we might be able to cure Alzheimer's, cancer, brain disease, as part of many other things that are going on as well.
Evan: There are a couple of aspects that you have to reinvent or advance on the chip side, correct?
Mary Lou: Sometimes, you get lucky! Apple and Google and others decided to ship a chip that you want to use exactly as it is and it costs a dollar - it’s in there smart phones and beyond. That's where we are. Although we get charged more than a dollar for it because we don't have the same volume.
We had to invent an impossible laser that pulsed with coherence at the right wavelengths so that we could make holograms onto this camera chip through your brain. We've now substantially reduced its size. I've got a tiny prototype ready to move to mass production.
What people probably don't realize is that in both their iPhones and their Pixel phones, there are 250 lasers in it. IR lasers that scan over your face to record with structured light, a 3D vision of your face so that a picture of you can't unlock your phone but a face scan can. That's a use of some of technology and Moore’s law - 20 years ago, those lasers were $100,000 a piece.
Dr. Mary Lou Jepsen and the team at Openwater developed a device that allows seeing blood flow throughout the brain. Image credit: Openwater
The lasers we're making now we are also putting into silicon processes. There's no reason their cost structure has to be much different than the lasers in your phone. What we can see with these lasers and camera chips and our software and AI, etc, is extraordinary detail of blood flow. That ends up being important for something that gets reimbursed for $1,000 per diagnosis, but now usually when it's too late. That's for the #2 killer in the world even in the mid-COVID, which is a stroke.
We can detect strokes quickly and that matters. Every minute counts when the blood flow is cut off in your brain. Your brain starts to die. There's a 90% chance of no neuro deficit whatsoever for something that right now either kills or severely permanently disables 60% of the people that have it in the case of a the severe strokes - they are called LVOs or large vessel occlusions.. That's 30-40% of all strokes. It's huge because that's the #1 cost of long-term disability. Every minute matters to save a life.
We're detecting strokes first. Then we've got a system where we think we can start to loosen the clot, but that unit also does neurostimulation where we seem to be growing nerve cells for Alzheimer's on a microscope chip and also bursting fast-growing cancer cells. We're working on things that can be very low-cost, that we're developing with an eye to what can go into a high volume mass production and not have to be in the hospital. It can be a low-cost structure, but also replace existing therapies that are flawed.
Right now, if you have Alzheimer's, you go 5 days a week to a hospital for 6 weeks to get therapy. That's hard for an Alzheimer's patient. Can we do something at home that's more accurate, focused, and productive? The early results are promising so we'll see how it pans out this year.
Evan Nisselson: We’ve been investing a lot in proactive healthcare. That’s about 40% of our portfolio! I've got friends and relatives that are doctors and they say, "Everybody's been talking about proactive healthcare for a long time." I think that we're finally at a cusp. What you’re sharing with us falls under that category too. When will the initial product be ready for the home?
Mary Lou: We have five products in our pipeline right now. The first one is a stroke detector. We're testing it in hospitals now, comparing our data to the MRI data so that a paramedic or point of care place can use it because every minute counts. When you have a stroke, your urgency is the same as somebody with a gunshot or a stabbing victim. It's just not apparent that you have that. That data will go to urgent care, EMS, etc. Then we plan to add something that starts to loosen the clot on the wave therapy.
With all that data, why does anybody have to have a stroke at all? Can't we predict it before it happens? Did you ever take fluid mechanics? I did. You get turbulence when something is blocking before it blocks completely. Think of white water rafting and the rocks. You can see the water around them. That's the turbulence. We think we should be able to see that and predict it prior. That's a whole suite of things for stroke!
It turns out the cell membranes of a fast-growing cancer are brittle and they've got bigger nuclei. What we've found, and the literature supports this, is that we can selectively kill fast-growing cells with brittle cell membranes and keep the healthy, slow-growing flexible cells in your brain that you've had for most of your life. A bit like a bridge can withstand an earthquake. As an opera singer can, with the right note, burst a wine glass but nothing else in the room. We can burst and kill those cells. There's no known cure for glioblastoma or these fast-growing cells, where it's like an octopus in your head.
That unit that can do that at a different frequency seems to be growing nerve cells in Alzheimer's. Here's the thing: there are ~250 drugs that have been tried for Alzheimer's. Each cost on average $1 billion, and none of them work. It seems crazy to NOT try a non-drug solution.
We can stimulate neurons too. That is also potentially extremely useful. We're starting to explore this with some clinicians for mental diseases. Repetitive negative thought leads to anxiety, OCD, PTSD, and ultimately depression, which are huge problems. We can objectively look at the brain and see what's going on. Rather than DSM (Editor’s note: DSM stands for Diagnostic and Statistical Manual of Mental Disorders) where you answer a series of questions that diagnoses you with depression or not. There are questions like: have you gained weight? Are you not sleeping? Do you have thoughts of suicide? If you answer “yes” to those and others, you're clinically depressed and you're no longer depressed when you answer “no” to them. Instead, can we see what's happening in your brain? Like if you come out of a coma, you need some stimulation on your muscles to get your muscles working again. We need some stimulation in our brain to get our brain working again. It's a bit similar to how psychedelics are being used, but without the psychedelics. Maybe that's less fun. Maybe it's more fun! It's more targeted and objective and doesn’t completely reorganize your brain… because we can target the exact part of the brain at issue and stimulate it.
Evan: Each one of those is a huge opportunity, let alone 4 or 5 different product tracks! Of the ups and downs, not only pushing the envelope technologically and your passion and determination, like in the lowest of lows, how do you get to the next stage or get yourself and your team to the next stage without being overwhelmed or stopping? Are there tricks that you have in the normal ups and downs to help manage it?
Mary Lou: Yes. No. I mean, I've been there with a blanket and a bottle of tequila crushed in failure. Except I don't like tequila but it's a metaphor.
Evan: Sure. We've both been there.
Mary Lou: I have to take a dozen medications every day to live. Every single day. I don't have a pituitary gland. Adrenaline is a key hormone. Cortisol is a key hormone. You die with out them immediately. I've had to fight so hard to live. What do you want to do with the time you're alive? Sure, you can take a week and wallow in a huge setback but ultimately, try to figure out what could have been done better and keep moving.
There are macros you can't control, but what can you do next time? What gives you enough passion to throw yourself back at it, because we're alive? We get to live now! We get to be here! There are so many things that can be done!
It's fun to work on something seemingly impossible. Make it possible, and get the people around the world that want to also make the thing possible and work with them. It's the most fun way to spend time!
Evan: Everybody has a different perspective and there's no right or wrong, but there is such a tremendous impact that your Openwater has plans to hopefully improve many aspects of healthcare and people’s well-being!
Mary Lou: Amidst the great resignation, you have people who think they have enough. Their bellies are full, so to speak. They've got abundance. Maybe they should just take it easy. If you've been working your 100-hour weeks for a decade, maybe you should take a little break. What seems to be true across the studies is that you're just not as happy resigning. You're not as fulfilled. Taking a break, being with family, switching up to a new career, or learning something adjacent or completely opposed, might be the way. Trying 5 things on for size, seeing what fits - going with that might make sense.
Evan: You've started several companies. Are there certain types of people or personalities that you think are most valuable to find for you? I mean, everybody's different. But what type of people are the ones, "Oh my God, I have to work with this person," as one of the first 3 or 4 people of a new company? How do you decide?
Mary Lou: I do hard tech so I don't go for the business people early on. I'm technical but I look for people in areas of my weaknesses where we can communicate well.
They have to fight fair. That’s key. Not fight to win but fight to find the best way. I remember working with one of the best guys I’ve ever worked with. I'd come in the morning and say, "I have an idea, and if we could do this thing we could make a much better impact faster" while we were getting our coffee and he'd say, "There's no way in hell that's ever going to work. Such a stupid idea!" But by lunch, it would've gotten into his head and he was good at things I wasn't. By lunch, he would say, "You know, the thing you're asking? There might be a way!" By dinner time, we'd have a working idea of how to do it because he'd understand the impact we could have if we could invent in this area together. I love working with people who are good in other areas.
We share these spheres that we cover so we can do more as we span out. Because the key again, that guy could debug in his head all the reasons it wouldn't work. A bit like an executive at a big company in consumer electronics, so you can go faster.
Evan: What would be your advice for that deep technical person who's a researcher now, or in school, or has been working in a couple of big companies, but wants to start a business, but hasn't jumped? What would your advice be to them, to help them decide if and when to jump to start a business themselves?
Mary Lou: They probably think it's safer at the big company, for some reason. At Google, people change jobs every year. They last less than 2 years in a role.
Evan: You mean within Google, they switch to different groups.
Mary Lou: They switch to different groups every year. The longevity of the purpose of a project is longer out of startup statistically. If there's something you want to do, you're probably going to get switched around from project to project at a big company and not get to do it for long. It’s OK - the projects meet the corporate goals and can be interesting. It's good to learn. That's a fine way to approach it. But if you really want to a thing outside the core of a large company as startup is the best way.
I started my first startup a month after having brain surgery... With my co-founders, I was like, "I can't do this. I need health insurance." And what they said is, "You seem to know a lot about health insurance. Why don't you pick the health insurance?" so I got the most expensive health insurance, and I still do, because why not do that at a startup? There's more risk because you don't know if it's going to work or not and it's not like you can switch to something safe. You might be able to pivot, but you're not going to get reassigned to some other group if it fails - if it fails the company goes under. But when you get to stick with the project, you know what's going on, and you have a lot more visibility in what's going on if it's a well-run startup. Everything should be shared with you, the board decks, the finances. You should know how much money is in the bank every week. With more information, you can make better decisions in the early stages.
Evan: You can likely have more impact as well! That's fantastic advice to close on, Mary Lou! I wish we had much more time, and I appreciate you sharing your insights and thoughts with us. We're looking forward to Openwater being a tremendous success and improving people's health around the world!
Mary Lou: I'm so impressed with all you do at LDV! I've learned a lot. Thank you for believing in visual computing and all the incredible people you've brought together!
Evan: Thank you very much!
