Key Topics of Episode:
- Genetics & Cancer – understanding more
- Stories from the early days at Google & related hacks
- Re-connecting with your calling in career and life
In this episode, we talk about your genetic code, stories from the early days of Google, accelerating feedback loops, and why you shouldn’t follow others without first understanding who you are.
Ryan was an early engineer at Google, and currently holds a senior position at Color, the genome sequencing company. He has dozens of patents under his name, and graduated from Stanford with a Master’s in computer science. He holds black belts in various martial arts. Ryan is also a very close friend of mine, starting when we were at Google during the same time.
Ryan: Thanks for having me, this is going to be fun.
Jeremy: Excellent. Let’s jump in. Among the many amazing things that you’ve done, at a young age, you did something that’s pretty unusual. You won a National Lego Competition, an early sign of your problem-solving prowess. Can you tell me more about that and how it influenced you, going forward?
Ryan: That was fun. I spent a lot of time building with Legos as a kid. I think I knew early on in life that I liked building things, I liked taking things apartment and putting them back together, that kind of thing. My sister and I are both very lucky in that sense. You hear a lot about quarter-life crisis. You know people or you maybe have this experience yourself where you go through college, you’re out of college, you work for a few years, but you don’t really know what you’re going to do with your life. There’s nothing clear burning in you, where you feel like, “I want to do this.” Some people have that but not everyone. My sister and I were really lucky we had that. You can try a bunch of stuff but it can be really hard to figure out what to do with your life. That can be difficult in a bunch of ways. I feel very lucky. It doesn’t necessarily have to be software or anything else. But I’ve always known that what I wanted to do with my life is build things. My sister always knew she wanted to draw or paint things. She’s an artist. We were lucky.
Jeremy: It is very valuable like you mentioned, it’s something that a lot of people, along the way, are getting back the roots of who they are and what makes them tick. Now, you solve powerful engineering problems at Color Genomics. It’s something where you’ve taken your skills throughout your career and applied them to a burgeoning field. Can you tell us more about the company and your role?
Ryan: Color is a health tech company. Right now, we focus on genetic tests. You spit in a tube and send it back to us, we see what’s in your DNA, and we tell you if you’re at higher risk of certain diseases that tend to have hereditary components. Right now, we’re focusing on cancer and cardiovascular disease. I spent a long time out of school. I worked at IBM Research part-time during school but my first full-time job was at Google, where we met. I was there for about ten years, doing mostly infrastructure engineering. I learned how to throw machines problems. When you’re programming a computer, it’s like a toddler. If you’re lucky and you get the instructions right and it’s in a good mood, the computer will do what you tell it. But that alone is pretty difficult because there’s so many ways you can tell it’s slightly wrong and it’ll just be confused. Once you try to do something larger scale where there are multiple computers, like Google where there’s millions of searches per second for example, they need lots of different servers working together. Think about your average preschooler kindergarten class. You have toddlers running around. One kid’s just peed her pants, another hits some other kid and she’s crying, and another is smearing chocolate all over the walls and floor. Getting them to do anything together is hard. Google was a great place to learn that. I felt like I really grew up in my career at Google and it was a great place to do that. So after ten years, I still liked it there. I learned how to throw machines a problem and get them to work together. But can I do that in service of something?
For so long, healthcare has been symptom-driven. When you feel fine, you generally don’t go to the doctors. Maybe you go for a check-up but you’re not doing any deep dive. When you feel bad, when you know something is wrong with your body, that’s when you start going, “Do you know what’s wrong with me, doctor? Can we do some more tests?” It’s not great. Your body is complicated and doing a bunch of stuff at once. By the time you feel something wrong, stuff inside your body has been going wrong for a long time. It’s gradually been building up. By the time you feel it, it’s pretty far down the road of not working right. So we are starting to have more methods to learn about your body, whether it’s biomarkers, blood panels, FitBits etc.
There’s a lot of opportunity there to start noticing things that might go wrong with you or could go wrong with you in the future, way before they actually start to. Those early interventions can be really powerful and you can end up with much better outcomes. Long-term, that’s the kind of thing we want to look at. We started with DNA but we want to do much more.
Jeremy: That’s interesting. When you and I were first talking about the world of genetics, I remembered some entertaining stories about purple cats and some other aspects. Are there any other unusual stories like that that you have or something fascinating in the world of genetics right now that you think are compelling?
Image of a glow-in-the-dark cat that was genetically engineered by scientists
Ryan: There are a number of interesting things going on. There are lots of recreational uses for DNA and genetic engineering, and non-health uses too. Glow-in-the-dark animals or animals with glow-in-the-dark fur, cross-breeding plants, those are definitely interesting. Craig Venter famously is working on creating the first cell from scratch.
Jeremy: What’s happening with Color Genomics?
Ryan: We started with cancer and we expanded to cardiovascular.
Jeremy: So that means you’re able to detect potential cardiovascular issues before it happens.
Ryan: At least risk. Most disease and other things that might go wrong with your health have some hereditary component and some environmental component. To grossly oversimplify, cancer in particular comes from DNA mutation, when your DNA has some change in it that’s not good. Often, you have one of these pathogenic or bad mutations. You got it from your mother or your father, or just randomly when the DNA was copying itself. And then, you get another one later in life environmentally, from smoking or something else. If I’ve learned anything at Color, it’s that biology is freaking complicated. We think we build complicated technology but it’s nothing compared to the human body. Most pathogenic mutations, you’re born with. Your body has enough redundancy and safe guards that it can protect you against them. They’re in your DNA forever, what’s called germline DNA. The DNA you’re born with is your DNA forever. But your body can still protect you from bad things that are in it. Often, once you get two of them and they happen to be in the same biological pathway in your body, then it’s harder for all the redundancy and safeguards in your body to protect against them. That is often how people end up getting cancer. Not always, it’s a gross oversimplification.
Jeremy: If I understand correctly, Ryan – and this is something more of a generic biology question – we all have cancer in our bodies. It’s just a question that it becomes expressed in a way that is problematic or not. Is that correct?
Ryan: This is still a very active area of research. In medical research, people are starting to come around this idea that cancer is not a singular disease you might get. Most people or maybe everyone has these micro-cancers, say five, ten, fifteen, or twenty of them in your body.
Jeremy: And they’re based on earlier definition that they’re simply mutations.
Ryan: No. Mutations in your DNA are what we call differences in DNA. There are a bunch of differences in my DNA from yours, for example. The vast majority of them are fine. They contribute to something visible like taller or shorter, or they don’t contribute to anything visible. But, they won’t hurt us. If a small fraction is pathogenic, that might actually cause something bad to happen in your body.
Cancer itself is when clumps of cells in your body start to multiply and grow without bounds. You have a bunch of genes that are called tumor suppressor genes. Basically, they control how fast your cells grow. So if some of those genes stop working because they have pathogenic mutations that turn them off, then cells in some parts of your body may become cancerous and start growing without bounds.
So there’s this idea that we all may have these micro-cancers in our body, that these little clumps of cells are growing faster than they should. But maybe they are growing only one percent faster than they should. Or maybe they are growing a lot faster but there’s only five, ten, hundred, or thousands of them. So they’re either small enough or slow-growing enough that you’d have to live to 200 years old or thousands of years old to notice.
Jeremy: That’s a problem that we may face at some point in the distant future.
Ryan: Yes. I think this is a fascinating example where in health and medicine right now, technology has advanced really fast to the point that a lot of the imaging we use, whether it’s x-rays, CAT scans, MRI, etc., they’ve gotten really high quality and resolution faster than we fully understand the new things we can see. Ten or twenty years ago, maybe we couldn’t see small enough detail to see these micro-cancers. But now we can.
For so long, TV was just SD. Now we have HD, and then 4K, and it starts ramping up really fast. It’s the same with medical imaging. But our understanding isn’t expanding at the same rate. So we’re seeing these things before we fully understand them. It’s exciting but it’s also scary because if it’s used to saying, “Oh, that’s cancer. We have to cut it out, burn it out, poison it out with radiation or chemo,” but it turns out it’s never going to hurt you, that’s a problem. The community is actually trying to figure out, “Do we understand these? Do we have to change what we do about them? We’re still figuring it out.”
Jeremy: In terms of where genetic technology is evolving and so forth, what are some of the ways that you think are pretty exciting in terms of either the next couple of years or the next decade? How do you see things shaping up?
Ryan: The biggest news has been the incredible drop in the cost of genetic sequencing, famously much faster than Moore’s Law. With human genome project 15 years ago, it cost a billion dollars to sequence the first human genome. Fifteen years later, we’re looking at well under a thousand dollars, closer to probably a hundred dollars at this point to sequence your whole genome at some reasonable depth. That suddenly brings it into the range of affordability for normal people and for health insurance, where health insurance will start to say, “The benefits outweigh the cost. We should maybe do this for everyone.” This is what’s called the population scale genomics.
It used to be that only when you already knew you had very specific diseases that were prevalent in your family, then maybe your insurance pays for it. Now, it’s much more accessible. When Color launched three or four years ago, the biggest differentiator for us was that we were affordable. We were $250 which was like the [inaudible 13:59], the out-of-pocket cost for one of the other labs, and we were much more accessible. We worked with an independent network of doctors who would order the test with you but you could just come to the website and buy the test. One of those doctors will work with you, instead of having to convince your GP. That brought industry shift, where it’s like, “It’s so much cheaper. Should we bring it to more people? Can we use more uses?”
Jeremy: I imagine that has a positive snowball effect when you have more people getting access to their DNA and companies are also giving more access to it. Then there is more data for you to analyze and there’s more scientific findings that can occur as a result of that.
Ryan: We think so. We firmly believe that everyone who wants to, should be able to learn this kind of thing. You have to do it responsibly. A fraction of people who get this kind of test, you’re going to tell, “The average person has a 2 to 3% chance of getting this kind of cancer. You have a 50, 60, or 70% chance or higher.” That’s deep news. What we felt strongly about and we’ve done in the beginning is with every test, we include a one-on-one session with a human genetic counsellor who can really help the average person understand this stuff and figure out what to do next.
The technology can outpace the understanding. You’ve got to make sure you bring the understanding to bear. If you see something new with new technology, you don’t jump to conclusions before you fully understand what you’re going to do.
Jeremy: That makes sense. We’ve all done searches online where we see some type of health information, we immediately in our minds, jump to the worst case scenario and spend however much time researching it.
Ryan: Google is both the worst doctor and everyone’s first doctor now, which is frightening.
Jeremy: It’s funny. You were at Google in the early days. What was a story from there that you’d like to share?
Ryan: When I showed up for work on my first day, I met my manager. He said, “Welcome to the team, glad to have you. I’m going on vacation for the next two weeks. You don’t have a computer yet but here’s your desk. Hopefully you’ll have a computer maybe later in the day, maybe tomorrow. Here are your teammates. Enjoy!” He was out the door.
Jeremy: It was sink or swim at that point.
Ryan: I think that was very emblematic of much of the engineering culture, especially at the beginning. One of the best engineering managers I’ve ever worked with, [inaudible 17:05], he was a very hands-off manager. At his peak, he had something like 120 or more direct reports, which is totally unreasonable and not how you should be a manager.
Jeremy: What do you say the average number of reports a manager has in the corporate world, traditionally? It’s probably anywhere between three and maximum, ten.
Ryan: Maybe in Silicon Valley we’re a little more, bottom-up or self-driven, you’ll get five to twenty or something. But one of the jokes in the early days was your job is to automate yourself out of a job. Find some problem and write some code or figure out something that solves it and get that thing working on its own. Then, go find the next one
Jeremy: That’s very different than the typical corporate motto of “Let me do what I can do to be irreplaceable.” In this case, what is irreplaceable is the ability to essentially automate yourself out of the job.
Ryan: Yes, especially finding the right opportunities and problems, solving them, and then moving on to the next one. It’s usually valuable and not necessarily obvious. We’ve all heard that there’s value getting outside your comfort zone. But the problem with that is it’s uncomfortable. So you need to accept that. Constantly find new things to do and not stay in the same place forever is uncomfortable. It’s okay that it feels uncomfortable because it’s going to create more value. That’s important. There’s this idea culture around here where, “Everything’s awesome! I’m crushing it, everything’s great, and everyone’s awesome.” I think one of the dirty little secrets that’s not so secret about starting a start-up is it’s really hard. One of the best descriptions of this comes from Jason Calacanis who said, “If there’s any way you can do anything other than starting a start-up, do it,” because it’s hard on your body, mind, your wallet, your relationships. It sucks in a bunch of different ways. So if you can survive and not be an entrepreneur, it’s probably better. You only want to do it if it’s absolutely burning inside you, in your bones, you can’t get away from it, and you feel like you have to start something. Then it’s okay.
Jeremy: In many ways, you’ve climbed the ladder of success in your career. At one point, you’ve been offered executive positions within your company very [inaudible 20:07] and you’ve turned them down. Most people will keep on rising and so forth. It’s not just courageous to do that, but it’s interesting. I’m curious what was the thought process for doing that.
Ryan: Management is not so fun. It’s a different job. You’re doing different things. It’s fun to get in there and write codes and solve problems, do something where you can immediately see the results. You run the code right after you write it and see what it does. That immediate feedback is really gratifying.
Managing, especially for engineer personnel types is fuzzy and it’s lots of dealing with people. It’s indirect and slow feedback. Even when you’re mentoring someone who reports to you and you’re doing a great job, and they’re growing and doing great work, often it takes weeks to months at least to see the results of a specific thing you helped them with. You lose a lot of that direct fun, that satisfaction of doing something and seeing it immediately happen.
Jeremy: Going back to the subject of programming, what did you learn about programming that you apply to other aspects of your life and that other people can apply as well?
Ryan: One of the biggest things, maybe unintentionally, I’ve spent time on in my career in engineering is feedback loops. Like we were talking about before, programming computers is one of those great things where you can write a code, run it, and see it run and see the results immediately. That can be a high, that sense of control can be really exciting and intoxicating. When you’re working in that way where you immediately see the results of what you’re doing – not just programming, it could be woodworking or playing music, etc. – you can get into this flow state, this mode where you do something and it works okay, but it doesn’t work quite right. You make some change and it’s now better but this other thing over here isn’t quite right. You make another change. That immediate feedback loop, where you’re getting feedback in seconds, is hugely powerful. You tend to work very ad hoc or exploratory. There’s very little risk or cost to trying some changes. If it works, then great.
Jeremy: It is something that in the world of programming, you can get that kind of instant results. You can see this play out in the business domain or life domain.
Ryan: Of course. One of the big things I have done in a number of different places and ways in my career is finding those places in the company or the team where there are long feedback loops, and shortening those feedback loops.
Jeremy: What you’re saying is that for people who are looking for opportunities, to increase the speed at which they’re getting feedback on whatever they’re doing is critical.
Ryan: I think anything you’re trying to do something or trying to make something, a pretty consistently good way to do it, the way to get the best chances of coming up with some good in the end is to iterate. You put something out there, you see what people think or you see if it works, you see if it breaks, you see how it breaks, you make a change, you try it again. The higher the number of iteration cycles, the more the better. You look at successful products, they probably had hundreds of iteration cycles where they got something out and got some feedback, made some tweaks and got it out into the same users hands to see, “Now what do you think of it, now that we’ve improved these things?”
There’s actually one thing I’ve never heard a great answer to. But for people who want to be entrepreneurs, obviously they’ve generally heard of this idea of iterating. That’s powerful. I think one thing that people think about less, is the feedback itself. You think about, “Oh, I’m going to make changes and put it out, then make more changes and put it out.” Sure, but just important as this is getting feedback from real users who are using your products to do something they care about, and then giving you honest feedback. You can make a bunch of changes over time but if you don’t have feedback, it’s rudderless. You don’t know that you’re going in the right direction.
The catch is that early on, almost always whatever you make, the first, second, or third version is not great. Often, it’s just bad. It’s not because you’re a bad entrepreneur. It’s just because it’s hard to come up with something really good at the very first try. Almost no one does. So the task at hand is to find real people who will give you honest, direct feedback and actually try to use your thing for something they care about when it’s really bad. That’s tough.
Jeremy: Quick insight on investing, you’ve worked at a number of amazing companies at their early stages. One of the things that I discuss was the idea of not investing ideas that everyone agrees are good. Those are generally the ones that are overpriced and probably too late to invest in. So instead, investing ideas that most people are bad but that you know are good.
Ryan: If everyone thinks an idea is a good idea, it’s probably been out a while and lots of people have heard of it. People may have tried it and it turned out to be a bad idea, even though it sounds like a good idea. Or maybe it is a good idea and lots of people are already doing it. This is also standard VC wisdom, is that you invest almost more into founders and people than ideas because ideas are cheap and prevalent. As an investor, if you’re looking at ideas as well as people, often the wisdom is to be contrarian. You’re looking for ideas that are clearly bad ideas. You want to believe an idea is a good idea and have a great chance for some reason, but also that lots of people don’t think it’s a good idea. Then, it’s more likely to be new. Fewer people have tried it. There’s less demand for it.
Jeremy: Ryan, what is the one thing you want to be known for when you’re looking back on your life?
Ryan: I don’t know if I care if people remember me personally by my name. I mostly don’t expect they will after a while. But I want to make the world the better place than it was before, I think like most of it. Raising a kid is one of the biggest and is very unique. I definitely think a lot about parenting. When she is out in the world, that would be the single, biggest, indirect stuff I’ve done. I joke that my job is software prevention engineer. It’s a joke but it’s only half a joke. It’s a great quote from Peter Norvig, a computer scientist who spent a while at Google, among other places. He says all code is a liability. It’s really true. Code is useful. We use it to us a bunch of stuff that we love. But also, every line of code could have a bug [inaudible 28:20] over time improving it, changing it, being able to read it and understand it. The amount of effort it takes to write a piece of code, that’s only 10% of the total lifetime cost of maintaining that code. I definitely spend a lot of time working with people at work to say, “If we can do something in less code, if we can buy it and that’s appropriate, or if we can reuse some open source, that is a net win in a bunch of ways.” So I think about in general, how can the things I build and things I write help people reuse it more, reuse the existing building blocks? You think critically about do I jump into building something new from scratch versus do I reuse something or adapt something?
Jeremy: Absolutely. How can listener learn more about you, Ryan?
Ryan: I am one of the dinosaurs who still has a website or a blog. That’s https://snarfed.org/. Color is https://www.color.com/. We’re hiring engineers. It’s a great place with really great people and it’s an amazing mission. I get up every day feeling lucky that I get to go help make lots of people a little healthier. We’re hiring a bunch of other roles too, so check out https://www.color.com/careers. Feel free to send me an email at ryan (-at-)color.com. This has been a lot of fun, Jeremy. You’re doing a great job with this podcast. I had a lot of fun.
Jeremy: It’s a pleasure as well, Ryan. Thanks again for taking the time and I really appreciate all the insight you shared today.
To recap, some takeaways we covered during this episode:
- Questioning whether what is popular is right for you
- A better understanding of how cancer works
- Why getting feedback as early and often as possible is essential.
Leave a Reply