Mike Nolan is an angel investor specializing in machine learning. As an officer in the Air Force he trained fighter pilots on F-15s. Mike relates the opportunities offered by graduate work at MIT both professionally and in investing. Fun chat with a buddy from Walnut and a listener to the podcast.
Highlights:
Sal Daher Introduces Mike Nolan
Piction Health and Machine Learning
"... we don't frequently see companies that are on the bleeding edge of the academic research or the bleeding edge of what's going on and the top presentations at conferences and so forth, but that's okay..."
The Capabilities of Machine Learning In the Health Field
Mike Nolan's Background
"... You got to picture this beautiful country, the White Sands, New Mexico, close by, and so forth, and here's this chimp running away from this 10-mile-long sled track..."
"... They're highly creative people in a highly regimented situation. They are people who think in very creative ways, and so forth, and they have big personalities ... I suspect that there is a lot of room for creativity with fighter pilots..."
"... Yes, things can go wrong, whether you're a founder or a pilot..."
Advice to the Audience
ANGEL INVEST BOSTON IS SPONSORED BY:
Transcript of, “Fighter Jets and Machine Learning”
Guest: Mike Nolan
Sal Daher: I'm really proud to say that the Angel Invest Boston podcast is sponsored by Purdue University Entrepreneurship and Peter Fasse, patent attorney at Fish & Richardson. Purdue is exceptional in its support of its faculty of its top five engineering school in helping them get their technology from the lab out to the market, out to industry, out to the clinic.
Peter Fasse is also a great support to entrepreneurs. He is a patent attorney, specializing in microfluidics and has been tremendously helpful to some of the startups, which I'm involved, including a startup, came out of Purdue, Savran Technologies. I'm proud to have these two sponsors for my podcast.
Sal Daher Introduces Mike Nolan
Welcome to Angel Invest Boston, conversations with Boston's most interesting angels and founders. I am Sal Daher, an angel investor who just loves to talk to investors and to founders, and to other people who know a lot about building companies. Today we're really privileged to have with us my Walnut colleague, former fighter pilot, and machine learning person, Mike Nolan. Say hi, Mike.
Mike Nolan: Thanks for having me on, Sal. Hello, everyone.
Sal Daher: We're going to talk about Mike's angel investing, and then in the second half of the podcast, we're going to get into his exciting career, flying F-15s and training people in F-15s and things like that. In the beginning, Mike actually connected with me as a listener to the podcast, and that conversation led to his joining Walnut and becoming a very involved and very appreciative member of our angel investing group here in Boston, and we've invested in a couple of companies together. I just wanted to talk to Mike a little bit about his perspective on companies such as a current investment that he's in, Piction Health. Mike, how did you connect with Susan Conover, the co-founder of Piction Health?
Mike Nolan: Susan and I both went to the SDM program, system design and management program at MIT at slightly different times. I graduated and then we come back from industry to participate as well. During one of those times when I came back to participate, Susan gave a pitch on her idea, just the very beginnings of her idea for what became Piction Health. I was impressed with it. I also was very interested in and continue to be very interested in machine learning which she was using to go forward with it, so that was the start of Susan and my connection.
Sal Daher: Susan has been on the podcast a couple of times, but for those listeners who haven't listened to that podcast, tell us a little bit about the premise of LuminDX/Piction Health now.
Piction Health and Machine Learning
Mike Nolan: It started out as a idea to use machine learning to recognize patterns in people's skin for skin cancer and then they pivoted to other skin diseases, but it uses the ability of machine learning to recognize images and to do that at scale on the edge, applying significant capabilities in machine learning to help patients out with those cases.
Sal Daher: Tremendous. Basically, you connected much earlier than I did.
Mike Nolan: Yes.
Sal Daher: A tremendous founder. Susan is highly motivated. She explained to me that a shortage of skin specialists is such that people with problems, it takes a long time for them to get in front of a skin specialist, so it's really important to be able to weed out the really mundane things that you have with your skin, and separate that from possibly serious conditions so that they can prioritize an extremely valuable time of dermatologists.
She had a skin condition herself, a skin cancer that she battled with, so that motivated her. Initially, it's funny because they were LuminDX at the time. When I first saw them, they had the idea of having a consumer app that people would upload images to and they would train their AI on these uploaded images. Eventually, they'd become better and better at telling what conditions they were. She decided to pivot into actually going out and wrangling images that have been diagnosed already and training her AI in that. The last count that I saw, she had something like half a million images of different types of skin conditions, including some of my skin, which I contributed. [chuckles]
Mike Nolan: Oh, you've contributed some skin too. You have some skin in the game. [crosstalk]
Sal Daher: Yes, I contributed skin images and I got more skin in the game than usual, not just money. That iteration got to the point where her AI was pretty darn good at finding out what normal conditions are, if something is just a tag or mundane things, and the skin cancer aspects of it, they're leaving it to dermatologists to make determinations, is my understanding.
Mike Nolan: Yes. After that initial introduction to Susan and her idea, a year or two went by, and Michael Mark from Walnut brought up the fact that she was raising again and her company was raising again. I got involved then and spoke with her. I was impressed with where they were going and where they'd been. I'm investing small amounts. I ended up going through TBD Angels and investing there with a group. That worked out pretty well.
Sal Daher: TBD is a great vehicle, To Be Determined Angels. If you're thinking about writing smaller checks, it allows you to build the diversification of your portfolio very quickly. I've invested through TBD. I've done a few deals. It is addressing an important section of the market, which was not addressed before.
Mike Nolan: I think one of the things that she's been able to do is get a moat around data. I know you've talked to her about that, and one of the things that's made Piction Health successful.
Sal Daher: It's a first-mover advantage. If you go out to a lot of medical centers and you get their images, then you're training your machine learning on that, and you're going to be way ahead of where someone else who might be coming in de novo into the business. They're also setting up in such a way that they can be almost like a clinic locally in various localities. I think that's the latest iteration. The most striking thing about Piction Health is Susan's drive. Wouldn't you say that?
Mike Nolan: Yes.
Sal Daher: Very quiet, low-key determination. She's methodical. She has the list of things, she knows what she's going to do. If it's not working, she just-- "Okay, this is not working. Let me see what other iteration I can have with this." She just relentlessly, but very thoroughly goes at it and goes at it. She is just phenomenal.
Mike Nolan: Very good job and keeps us informed in terms of where things are and so forth.
Sal Daher: Excellent communication. By the way, if you are a founder or if you're thinking of founding a company, remember, don't keep your light under a bushel. Talk to your investors on a regular basis, and don't overdo it. Don't imagine it has to be a magnum opus, a great work. It's just 30 minutes every quarter.
By the way, I highly recommend that the founders read last quarter's report, and ensure this quarter's report somehow ties in with last quarter. I've seen situations where that doesn't happen. Take 30 minutes every quarter, schedule it, and do it methodically. It doesn't have to be a lot. Mike, would you say that you want to see what their runway is, what headway they've made, and what help they need?
Mike Nolan: Those are the main points there, and yes, you're right. Keep it short, but don't go out of sight on us, right?
Sal Daher: No, no, no. Don't do that. 80% of the time when a startup goes dark, goes quiet, it's because something horrible has happened, and they're embarrassed to call because things aren't going well, and that's exactly the time when they should be contacting the investors because the investors are on their corner and they should be reaching out for help. We're there to help.
The other 20% of the time, and [laughs ] I don't know if you've invested in this one, but there's another machine vision, a startup that is doing really well, but the founder has been less than communicative, and that went a long time. In the founder's mind, the ante had been raised. The report had to be really-- the company was doing really well. They raised VC money. A simple three-paragraph report, 30 minutes would've done it, but then, "Oh, I can't just-- after six months of silence, nine months of silence, I can't just send three paragraphs. That would be insulting. I have to--" No, send the three paragraphs. Don't feel bad.
Mike Nolan: No, that's certainly right in line with what we'd like to see.
Sal Daher: [laughs] Also, keep yourself in front of your investors to elbow aside other startups. What kind of investment interests you? I know your background in machine learning, you want to do a machine learning company, but are you restricted to machine learning, or would you venture outside of that?
"... we don't frequently see companies that are on the bleeding edge of the academic research or the bleeding edge of what's going on and the top presentations at conferences and so forth, but that's okay..."
Mike Nolan: I do plan to venture out as we go, but I'll make my first queue in machine learning. Sal, one thing as we see companies that have some level of AI component in them when they come present to us, frequently what they do more often than not, it's just a small component. We don't frequently see-- occasionally we do see, but we don't frequently see companies that are on the bleeding edge of the academic research or the bleeding edge of what's going on and the top presentations at conferences and so forth, but that's okay.
Sal Daher: To be there, you need to have a lot of resources, either backed by a university or a major corporation. You're not going to see that in startups. You're going to see maybe slight specialization of open-source resources, tailoring it to a particular set. Then the data set, gathering a data set and training the dataset with existing resources slightly tailored to the purpose, which is what Susan is doing in Piction Health.
I've seen a few more of these. I've had Gil Syswerda on and he's talked about the amount of computing that has to go on for you to really go on the cutting edge of these things is so massive. Also, the scarce talent is so expensive.
Mike Nolan: Most of them that we see are applying machine learning to recognize patterns. The main point of the companies is something about how they apply that to a particular business case. That's just fine for us to be able to recognize that and to see what they're going after. I think one thing in the companies that are on the forefront of applying AI/ML or doing it in a way that's innovative is to take a look at how what they're doing compares to what's available in the open literature.
Many times companies, whether their innovation is in AI/ML or somewhere else, are somewhat reluctant to share the details of that with the investors. That's maybe a separate topic, but given that's generally the case when we have people come in and talk to us about their AI/ML application, taking a look at how that compares with what's in the open literature, compared to papers that are being presented, or ask them what inspires them that they see in the open literature, we can relate that to what they're doing. I think that is a good way to start the conversation that helps both the founders and the investors out.
Sal Daher: I see you asking those kinds of questions. I hear you asking those kinds of questions in meetings and trying to dig a little bit into the AI/ML approach that they're taking and compare it to what's already known.
Mike Nolan: I think sometimes I'll go into those after reading maybe just a few abstracts of papers that are similar in a particular area that a company is in. I'll have the impression going into the meeting that what they have doesn't appear to be quite so special compared to what's available. Asking those kinds of questions gives a founder or CTO or whoever the opportunity to explain it, "No, actually, there is something different," that it doesn't appear quite so open and shut as you might think, given the open literature.
Sal Daher: What's the ideal company for you right now? A company that has all the right features.
Mike Nolan: All the right features and so forth? In terms of the team, certainly, that's been stressed to me as a new investor is to really focus on the team and recognize that probably the direction they're going right now is not going to be the same direction that will be successful for them in the future. That's one thing I look for. I do look for teams that have some credentials. That's not necessarily going to be the case in all successful companies, but I am somewhat picky about that at least right now. Then I do look for at least currently some AI/ML that looks creative to me, something that I might not have thought about that's some special sauce.
Sal Daher: Tremendous. No, I think that's very exciting because there's so much happening in that space. It is just so promising in so many variations. When I think about something as mundane as my WHOOP band, are you familiar with the WHOOP, this connected device band for tracking your fitness?
Mike Nolan: Yes.
Sal Daher: I started wearing it because I was trying to get fitter, but I didn't expect that the darn thing would be so good at helping me get better sleep.
Mike Nolan: Oh, really?
Sal Daher: Yes. The recommendations that it was giving was counter to sleep coaches. Sleep coaches tell you, "Oh, sure--"
Mike Nolan: Oh, really?
The Capabilities of Machine Learning In the Health Field
Sal Daher: Yes. You're a young man, you don't know about this, [laughter] but as you get older, when you wake up at two in the morning and you can't go back to sleep, the standard recommendation is like, "just go read a book and then you feel sleepy again and you go back to bed." WHOOP doesn't seem to-- they basically just tell you like, "Get to bed earlier," which is like, you're not supposed to do that to get the better-- By gosh darn, I get to bed earlier and it's broken through that, and I'm sleeping like a log.
Then I started thinking, I said, "Well, yes, of course, they probably have their machine learning looking at the sleeping patterns of hundreds of thousands of people, whereas these sleep studies, they're done with a few hundred patients." You start discovering weird stuff. With 200-patient studies, 300-patient studies, if you're someone who fits the mean, the median of the population, it's going to be very helpful, but there are subpopulations that don't get picked up.
It's not fine-grained enough, but when you have hundreds of thousands or it's going to become possible for you to do studies on millions and millions of people, like the COVID research database that my brother-in-law contributed to, 80 million patient record data lake, [chuckles] you can learn stuff that's astonishing and pick up patterns at a very fine grain because you have such large numbers. I'm very excited about the stuff that we're going to be learning.
Mike Nolan: Yes. Particularly, it is able to really focus on your particular case, and doing that tailoring is just so important.
Sal Daher: I probably have a weird situation, but my weird situation, they've seen a few thousand of them. One in a hundred, but it's a few thousand, whereas in a 300-person study, it's not statistically significant the results that they're going to get from one or two patients. Really mind-blowing, the potential of machine learning applied to health. Susan's going to be seeing that. Right now, her AI is basically like, what is it, 95% accuracy, something like that?
Mike Nolan: Yes, something like that.
Sal Daher: It's certainly better than a primary care physician and approaching dermatologists, but it's going to give doctors superpowers, they'll be able to figure out things so quickly, and I'm very excited about it.
Mike Nolan: I see that things are moving certainly really fast, especially with the large language models and the recent advances in ChatGPT and so forth. What are you seeing there, Sal? What's your perspective on that?
Sal Daher: I don't really have an informed perspective there. I haven't had the bandwidth to really dabble with it. I know it can be fun [laughs] to interact with it and so forth, but I must confess, I have not had a chance to dabble. I think it's very impressive, how quickly these things are moving. I'd hate to be someone in a Google search division right now.
[laughter]
Mike Nolan: Hot spot to be in, right?
Sal Daher: Yes. You're under fire because these open source ability for people to discover stuff out there is extremely promising.. I really don't have a developed view on that stuff.
Mike Nolan: That's not my particular area of AI/ML expertise. I do have a few thoughts in terms of what I look for or anticipate looking for as an early-stage investor. They talk about, is this going to be just the battle of the titans? Is this going to be just OpenAI and Deep Mind and maybe one or two others playing, or is there going to be room for others?
I think Sam Altman and others have pointed out that they anticipate that even though a lot of the resources necessary to do forefront R&D will be done by players with deep pockets that afford to do that, that doing the specific applications, there's likely to be a lot of room. I think, like you, that looks like some really exciting opportunities there.
Sal Daher: Well, I think that with the Tech Winter, combined with these artificial intelligence resources becoming widely available, you're going to have a lot of talented people going off and doing cool stuff. I'm hoping to interview a founder of a company called Alleviate, and he is somebody who used to be at Wayfair which is like a creation of internet, maybe 2.0. What he's doing is creating a platform. It's connected health really, connected therapy for people who have plantar fasciitis and other types of conditions that are amenable to help with therapists, but helping to coordinate care.
I think that AI is going to be making a lot of that much easier for people to do. As you begin to get a large number of patients, instead of having an individual coach and individual therapists, you're going to end up having some kind of a machine learning algorithm that really figures out what's the best. Have you tried this? Have you tried that? It'll become popularized, I can see it everywhere.
Mike Nolan: There's some really amazing stuff that's come out for sure about their ability to do chats and start to write code and so forth.
Sal Daher: The potential for coaching people on something like doing physical therapy, walking people through physical therapy, can do a lot of good because physical therapy is something that is really boring. The reason people fail at it is the very extremely mundane reasons. They just don't do it or they don't do it enough. They don't do it the right way. If it's done the right way, it works. If it's not done the right way, it doesn't.
I see thousands and thousands of ways that these things can be used. Then hundreds of thousands of very talented people, all of a sudden, trying to figure out what they're going to do next. I think in five years, we're going to look back and say, "Whoo, that was a watershed moment."
Mike Nolan: Yes, for sure. Probably, we'll see small companies that have maybe a moat around some particular data or have a way to clean the data. We see a number of cases where chatbots come up with some really amazing stuff and other times when they're very confident about spewing forth something that's completely wrong.
Sal Daher: [laughs] Utter nonsense. Nonsense spewed confidently.
Mike Nolan: Right. Being able to separate out that data and make sure that you're providing something that's high confidence and correct will be useful. Then having some special sauce algorithms being able to steer the general large language models in directions that help a particular area. I think the first inklings of that maybe we're seeing are what prompts you put in. I don't know if you've played with this a little bit or much.
Sal Daher: Not at all.
Mike Nolan: There's something to putting in a prompt that can elicit a general response. Then there are ways to put your prompts in a more specific terms, they can get something more useful out of it. That's the start, I think, a differentiation that the small companies will be able to build forth, going forward. Just as we are able to get better performance out of people by talking with them in a particular way or pointing them in a particular direction, I anticipate there will be ways to work with chatbots, or whatever you want to call them that that are smart, and some ways that are not too smart.
People potentially will build that as one of their skills and become more employable by being able to do that and potentially companies will have systems around how to do that smartly, that differentiate themselves in a particular area, about how to do that in a way that it's useful. I think those are some of the kinds of things that we might see in the future here very shortly.
Sal Daher: I wonder how this will play. The economist Friedrich Hayek had a concept of general information and particular information. General information is information that is true for large sets of things. Particular information are things that are very localized, like, "This street turns this way, but no, but there's a bump there and so forth," things that people who are there know, but that's someone far away, a planning authority or something have no clue about.
I suspect that what the big players in AI will be very good at is the general information and building algorithms, but the particular knowledge, the application to certain verticals, and the dynamics that exist in that vertical, training that AI, adding a little bit of a special sauce to it, making it a little bit specialized, that's where the value is going to come in. The big players will be able to capitalize on providing services for these, like servers, the way that Amazon provides web access, and so forth. I don't know exactly how they're going to monetize it, but I'm sure they will find a way to monetize their open-source technology.
Mike Nolan: I think, in general, we start to think about these as chatting with the monologue, the bot, if it's ChatGPT or BERT, whichever one it is. Actually, it may be possible to separate them out and have more than one. If you start-- just as we have companies that do well with a diverse workforce of people who think in different directions, it may be possible to have different instances of these bots who are trained to think in different ways. That group of collective agents, if you will, may be an opportunity to provide better value to customers in a particular market.
It'll be pretty interesting to see how a lot of these things shake out in terms of us investors looking to see what's coming in the door, and also as founders scouring the bushes to see what can specialize their capabilities relative to others.
Sal Daher: I think you're going to have a lot to do with your focus on AI/ML.
Mike Nolan: Yes, it's an exciting time for sure. Also, there's many unintended consequences of these capabilities.
Sal Daher: Is there anything else that you want to touch on this before we go into your background?
"... All kinds of problems and solutions that we haven't even started to touch on, but lots of opportunities to help people and to do good by doing good for investors..."
Mike Nolan: Maybe there's one or two more thoughts, in terms of what we may see as investors with things coming through the door. There's adverse consequences of these large language models. Certainly, trust is one of them. We have the opportunity for Deepfake kinds of things and just being able to trust what we're seeing with capabilities there. Solutions that help us solve those problems certainly is key.
Another example to think about is there are companies already that have come out with chatbots that enable you with animated faces and so forth to make a likeness of a deceased loved one. They've already shown in cases where they're saying that this has been a significant benefit to someone who just lost a loved one. A whole new world for us out there. There's adverse cases of that where they're already thinking about this may mean there'll be some people addicted to this kind of thing. All kinds of problems and solutions that we haven't even started to touch on, but lots of opportunities to help people and to do good by doing good for investors. Lots of new things coming down the pike here.
Sal Daher: Doing good by doing well. Is there a specific example that you want to point to that you think could be particularly promising?
Mike Nolan: That might be one where we have some adverse consequences of people who are now faced with-- whether it's that case there with a deceased relative or maybe it's the opportunity to make animated chatbots of a spouse, but a spouse who's 30 years younger than you are currently.
Sal Daher: [laughs] Boy, you're going to get into trouble.
Mike Nolan: Yes. How are people competing with that as we start to get more and more separated and there are more and more forces separating us from the real world to get into this other world? Lots of opportunities to go forward and solve those new problems.
Sal Daher: I saw a picture of 30 years ago, my family, me and my wife and our two daughters, and I'm looking, I said, "How did that goofy-looking guy get to marry this gorgeous woman?" I mean, this is-- I said, "Oh, he's such a goof. What a lucky devil." I think there's a lot of room for behavioral support for all kinds of things. I was talking to a company earlier today that provides basically an app that makes it really easy for people who are dealing with diabetes to control their diabetes. It's a very nice user interface.
They're basically taking information that's available out there right now and integrating it and putting it into a user interface that is really easy for human beings to understand. There's a coaching component to it. I can well see that coaching component being taken over, instead of a human being doing it, actually having some kind of a machine learning algorithm that really learns from the best coaches and allows the best coaches to become coaches of coaches to oversee situations where-- just look at the results and so forth and then intervene as it becomes like a force multiplier for coaching talent.
The potential there is huge because there's so many problems that we have to do with behavioral support, helping people do things that they know they have to do, but it's hard to actually do them. It's hard to implement them. I think humanity is going to benefit tremendously with the help of artificial intelligence.
Mike, what I thought I would do is do a little promo for the podcast, and then let's get into your background and your career, how you got to fly F-15s, and how you got into all that. First, if you're finding this conversation interesting, you can help get our podcast found by first rating us. We love five-star ratings. My mom always told me, "Go for the five-star ratings." Also, take a moment to leave a very brief written review of the podcast. You can be very candid in that written review. You could say all kinds of things you want to say, be polite, but you can be candid because it doesn't really matter. The algorithm sees that someone has taken the time to write and it upranks the episode.
What it does is it gets that conversation which provides really valuable information on building companies and making investment decisions to more people. It helps take it to more people. It's a way of upvoting us is to leave a review. Of course, follow us on the app that you use for listening to podcasts because that means it will show up week after week. If you like listening to the particular topic, you can listen, or if not, you can wait until next week. Anyway, Mike, what's the story? How did you end up lying F-15s? By the way, where did you grow up?
Mike Nolan's Background
Mike Nolan: From Central New York, Sal. A small town called Cazenovia.
Sal Daher: That's why you were at RPI, Rensselaer Polytechnic institute in Troy, New York.
Mike Nolan: That's right, exactly. After that, I was a civil engineer in the Air Force for a year and a half before they let me into pilot training.
Sal Daher: Civil engineer? Yes, I also studied civil engineering. All right.
Mike Nolan: I went to pilot training in Columbus, Mississippi. Then on the F-15, my first assignment was out at Langley Air Force Base, Virginia, where I met my wife, Kim. She was a nurse in the Air Force.
Sal Daher: This sounds like one of those World War II movies. [laughs]
Mike Nolan: Yes, right. [chuckles] Fun times there.
Sal Daher: Take a moment to explain to young people who don't know what the F-15 is. The F-15 is just an unbelievably hot airplane.
Mike Nolan: Yes, I felt certainly very fortunate to fly it. It's a two-tail, two-engine, and it comes in models that include a single seat for the air-to-air version and a two-seat for the air-to-ground version. It's a pretty cool-looking plane. It's about maybe the furthest from entrepreneurs and angel investing that you can see. We didn't see very many people peddling new company ideas at Langley Air Force Base, Virginia.
Sal Daher: [laughs] Everything has to go according to spec. Otherwise, stuff blows up.
Mike Nolan: That's right. Lots of large companies think-- the closest I think we came was something called an Eagle Eye. They made the F-15 with a great radar. There wasn't any electro-optical system in it. To visually identify whether it was a bad guy or a good guy, somebody had the idea of putting a deer rifle scope mounted in the cockpit. You'd actually carry these things out and screw them on right next to the heads-up display and peer in that to see what you're pointing at. Other than that, there weren't too many small company people hanging around the Air Force at Langley.
Sal Daher: It is just an amazingly cool aircraft. Imagine, it's been flying for 50 years and it's still in service. It's not a stealth aircraft, but it's just a hot airplane.
Mike Nolan: Following that, we went up to Holloman Air Force Base. That's got a colorful history. I studied electrical engineering there and did a little bit of work as electrical engineer out of the Holloman Test Track. I built the sensors and designed the system to measure the telemetry for the F-22 and F-23 ejection seat. Actually, this was years ago, I went down to RadioShack and bought components myself for a few dollars and popped those together and did a little engineering out there.
They've got a colorful history. The sled track is essentially a railroad that goes up to Mach 8 or so, really screaming. In their early days, they had chimpanzees that they tested. They were trying to understand impacts of humans in different cases here. They had in one case a chimp and they put him in a sled track, and they shot him down and he got to ride one. He decided he didn't really like that that particular railroad.
Sal Daher: [laughs] He had other ideas.
"... You got to picture this beautiful country, the White Sands, New Mexico, close by, and so forth, and here's this chimp running away from this 10-mile-long sled track..."
Mike Nolan: Yes, that's right. They were setting him up for trial number two, and he said, "I've had enough of this," and he went scampering out across the desert.
Sal Daher: He went AWOL.
Mike Nolan: That's right. He somehow got away from them. This is before I was involved, for sure. You got to picture this beautiful country, the White Sands, New Mexico, close by, and so forth, and here's this chimp running away from this 10-mile-long sled track.
Sal Daher: A chimp in a flight suit running away. [laughs]
Mike Nolan: Right, [chuckles] and a bunch of engineers who hop in their cars and they've got this caravan chasing the chimp out across the desert.
Sal Daher: Chimps are tremendously strong. They're phenomenally strong. They're five times stronger than human beings and can be vicious.
Mike Nolan: Disarming, yes.
Sal Daher: Handling chimps must be something else. It's just amazing. Once you're training there and then after training--
Mike Nolan: After training, I went to actually Holloman and I was an instructor as my first assignment there at Holloman. Then I went out to test pilot school at Edwards Air Force Base, which also has a colorful history. They did some tests there where they would put dummies in airplanes and they were testing how they would separate from the airplane as paratroopers.
This, again, was years ago. They did this particular test, somewhat close to base, and they dropped this dummy out. It didn't separate well, so it turned out with a chute that didn't open, so it was a streamer, and it goes screaming down to the ground. There were some officers' wives who were having a little get-together out in the backyard. It was close enough for them to clearly see this dummy, which appears like a person going screaming down.
Sal Daher: It must have spoiled their lunch. [laughs]
Mike Nolan: I think so. They take this pickup truck and go right by the house and out into the desert, and unceremoniously plop this dummy in the back of the pickup truck. They've got, of course, no idea that the officers' wives are partying or what's going on there. They had some explaining to do.
Sal Daher: It's just the dummy. It's not the dummy of a husband. It's an actual dummy.
Mike Nolan: Right. I got to do some interesting testing at Edwards though.
"... They're highly creative people in a highly regimented situation. They are people who think in very creative ways, and so forth, and they have big personalities ... I suspect that there is a lot of room for creativity with fighter pilots..."
Sal Daher: Tremendous. Now, the interesting thing is that fighter pilots, my read of this, and what do I know about fighter pilots, but what I've read a bit about this, and it is that they are free spirits. They're highly creative people in a highly regimented situation. They are people who think in very creative ways, and so forth, and they have big personalities. At the same time, they have to go through checklists, and so forth, so there's an interesting tension. I suspect that there is a lot of room for creativity with fighter pilots.
Mike Nolan: Yes, that's true, Sal. One point, you have to have the discipline to do things exactly in a particular order, exact steps you do to put on your seatbelt, and so forth. There's a right way and a wrong way. The specific radio calls that are done exactly in an order and so forth are very disciplined. Yet, there is the opportunity for certainly a fair amount of creativity. In the earlier combat days, your assignment is to go up and play with the clouds almost.
Sal Daher: [laughs] Yes, for free. You're just up there.
Mike Nolan: Right.
Sal Daher: 2.5 times the speed of sound and 50,000 feet up in the air.
Mike Nolan: Right. Going back to that chimp, it also takes a look at attitudes of fighter pilots. Arguably, the chimp was a little bit smarter than the pilots because he knew that he should stay away from those types of things. Particularly in the early days of testing, that was the case.
Sal Daher: Test pilots, it is a peculiar disposition that human beings have to test things themselves. The test pilot going up in an untried aircraft or doctors, there's a history of doctors experimenting on themselves. My daughter went to medical school at the University of Vermont where a lot of the initial work was done on X-rays, and these doctors were taking X-rays of their limbs and so forth.
Of course, they didn't know toxicity of radiation at the time. Those poor guys, they had to have their limbs amputated. It was horrific. It was just horrific. It continues to this day. I listened to a podcast of this physician out in Texas, Peter Attia, and he's always experimenting on himself; different regimens for exercise, different regimens. People can't help themselves.
Mike Nolan: The chimp may be smarter than some founders. Arguably they say that's not a rational act either.
Sal Daher: Well, it's a crazy thing, being a founder is these tremendously talented people going off on this venture that's likely to fail, and they could make more money doing something else. Yet they have to do it. Hats off to them. I have great respect for them. We wouldn't make progress without people doing those incredibly crazy things like these poor physicians exposing themselves to radiation the way they did, making lives better for the rest of humanity.
Mike Nolan: I think there's a couple of good lessons both for pilots and for founders. One of them is to be humble. I've got a couple of examples in the pilot side anyway of that. Following the Edwards assignment, I went to Wright-Patterson and came back to Holloman. I was a squadron commander of-- I guess there were about 60 people or so in my squadron. Some of them were contractors and their sole job was to fly maybe just one or two planes out of this special facility.
I was the main pilot for that group. I showed up at the place to take the aircraft off, and I was driving a standard transmission. One of the important things, when you drive in a standard transmission when you park it, is to put it in gear. I always would put it in gear. Another thing, I was always a belt and suspenders guy, so I would always put the emergency brake on as well.
This time I did neither. It was a pretty flat parking lot, so I went out to fly, I came back, I debriefed, and I strutted out to my car, and I saw that it was backed into somebody's jeep and had done a little bit of damage to their bumper. Things can go wrong.
Sal Daher: [laughs] If you were parking a car in Buenos Aires Argentina, circa 1979, almost every car was standard shift. Buenos Aires is a very flat city, so most streets are pretty flat. There are a few hills, but by and large, it's flat. People leave their cars on neutral and they park very tight, so the person who wants to get in moves the car a little bit forward and backwards, it's--
Mike Nolan: Oh, really?
Sal Daher: [laughs] Everybody's in neutral. All the cars, it's a flat street, they're all parked in neutral, and you get into the spot and then you very gently touch the bumper of the car and move it forward, or move the car behind you backwards a little bit to accommodate yourself into the parking spot.
Mike Nolan: Interesting.
Sal Daher: That's how you know the Portenos, the people in Buenos Aires, the Port of Buenos Aires, how they park.
Mike Nolan: Interesting. I had no such excuse.
Sal Daher: [laughs] Another thing they used to do down there, I don't know if they still do it, when I was living in Buenos Aires, is they drive around with their lights out at night and they only turn them on into the street crossing, when they're crossing and so forth, to save headlights. [laughs]
Mike Nolan: Oh my goodness. Wow.
Sal Daher: Everybody with their lights off and then they turn them on and then they turn them off.
Mike Nolan: Get a little bit better gas mileage too, perhaps.
Sal Daher: Well, no. It's because parts used to be all imported and so forth, everything's expensive. Mi Buenos Aires querida.
"... Yes, things can go wrong, whether you're a founder or a pilot..."
Mike Nolan: I get back from there and I asked if I can pay the guy to fix the bumper and so forth, and they said, "No. We'll take care of it," so I paid in alcohol, of course, as a fighter pilot. I go back there the next day for my next flight and they've with red cones marked off the spot behind where the commander parks because they need to be careful about protecting their cars [laughter] from this clueless guy who doesn't know how to park his car.
Sal Daher: Yes, the commanding officer is the biggest danger here. In the F-15, you did the checklist, but in the car, you didn't.
Mike Nolan: Not so sharp there. Yes, things can go wrong, whether you're a founder or a pilot. Another example of a similar type of thing that a colonel told, I think it was during his going away speech about a group that thought that it'd be cool if they put marshmallows in the speed brakes of airplanes.
Then when they came back to the base, they'd drop these fluffy marshmallows on the base and become famous. What they forgot was that they were going up to about 20,000 or 30,000 feet and coming back--
Sal Daher: They'd get frozen. [laughs]
Mike Nolan: Exactly, exactly. They were icicles, and they--
Sal Daher: Marshmallow bullets.
[laughter]
Mike Nolan: Yes, that's right. They had a lot of glass damage that they had to pay for when they got back. Be humble and be careful. [chuckles]
Sal Daher: At MIT, they have these pranks, they call them hacks. It's an East Campus. A famous hack at East Campus was stenciling names and designs on the carpet, they would cut a piece of paper and then design on it. Then they would put that pattern on a rug, and then they would dab alcohol on the carpet in the parts that were exposed in the pattern. Then they would flame it, and alcohol burns at a very low temperature, and it burns very quickly, so it's not enough for the carpet to catch fire, but it is enough for the color of the carpet the change.
These guys used to do that. These engineering, guys, these are the engineers. Then I know the story of a guy who was a math major, okay, math majors are completely impractical. [laughter] This guy was actually-- I heard he was actually and RA. He thought about doing the same thing, except he used the wrong fluid. What do you think he used?
Mike Nolan: Coke. No, I don't know. [laughs]
Sal Daher: Gasoline. [laughs]
Mike Nolan: Uh-oh.
Sal Daher: He's a mathematician. He has no clue, the difference between alcohol and gasoline. Gasoline explodes when it catches because it burns extremely quickly. Alcohol burns at a very low temperature. Gasoline is completely the opposite. Caused a fire, the loud fire trucks, and all that stuff, instead of leaving just a gentle mark on the carpet. There are hacks and there are crazy hacks. Mike, curious to know, in your experience in investing so far, I mean, what have you learned? Are there some mistakes that you made that are kind of obvious now in retrospect?
Mike Nolan: Yes. When I first started looking into this, I was told that one of the main mistakes that's frequently made is investors get in and they invest very quickly with a lot of funds. I was determined not to make that mistake. I made a whole different kind of mistake. I set the bar too high and hadn't really had many investments. I think I've got that figured out to a certain extent. We'll see if this works going forward. As I look around the Walnut group, I think, myself and Kim, if you look at a Monopoly board, we'd be down towards the Baltic Av and Mediterranean Av section of the group though.
Sal Daher: Low rent section.
[laughter]
Mike Nolan: Yes. I'm looking certainly to spend enough to cover about 30 companies or so because I do think it is important to get enough back to make things work.
Sal Daher: That diversification of the portfolio is important. Yes, definitely important.
Mike Nolan: I think mixing in some TBD options may be good as well because that will enable us to get in small investments and get the numbers up there. I think that could be useful.
Sal Daher: We're talking about TBD Angels, which is another angel group, in addition to Walnut Ventures which we both belong. TBD, there's a podcast that I did with David Chang and Yael deCapo. You can hear all about how it came together and the premise behind TBD. It's great for people who want to invest small amounts. They review, they don't have pitch meetings the way that we do. They look at video, it's asynchronous. People working full-time and want to invest smaller amounts, it's a great solution.
There is carry, but in the end of the day, if an investment's successful, believe me, you're going to be glad to pay the carry to the-- carry is small, it's a small percentage, you'd be very glad to pay the carry to the sponsoring person who brought the deal in. Anyway--
Mike Nolan: I know I had been concerned about the carry to start and had been down again, but yes, after listening to your discussion of it and others, I think that it does make sense.
Sal Daher: Given your experience with so many different areas, domains in the Air Force as an instructor, and then working in AI/ML and now as an angel investor, what kind of advice would you leave our audience with; audience of founders and angel investors and people who advise them?
Advice to the Audience
Mike Nolan: I think, certainly being able to leverage the community listening to what angels in different groups-- certainly, that's been very useful for me, and I think being able to join with a group and not trying to do it on your own would be very useful as well.
Sal Daher: Actually, Mike, before you go on to the next point, let me expand on that point, which is really valuable. People say, "Oh, using the resources of the group to evaluate founder," this is where it's really crowdsourcing or group sourcing the evaluation of the founder is really important because, in early-stage companies that we're investing in, they don't have established business models yet. There's not a lot of cash flow projections that you can analyze.
It's all about the motivation and the commitment of the founding; the founding team, how well they work together, how capable they are, and that's the kind of stuff where getting lots of eyeballs and talking to lots of people who know the founders, it can really be helpful and this is what angel groups excel in, making this judgment about the founding team, and then later on helping the founding team, but initially, this is a tremendously valuable thing. You cannot do this on your own. It's so much better when it's done in a group.
Mike Nolan: Yes, I certainly agree with that.
Sal Daher: Please continue. What was the next point that you had?
Mike Nolan: I think certainly my experience in Boston has been very good. Looked at a number of different groups, I'm currently as we've discussed in TBD, Walnut, and also in MIT Angels, but certainly, being able to leverage the thoughts and ideas from people all over Boston appears to be a good area to be in to able to do that.
Sal Daher: Lots of Angel groups in Boston. Great. Any other thoughts that you want to touch on?
Mike Nolan: I'd just like to thank the folks in each of these groups, and you in particular, Sal, for helping out and helping us go forward.
Sal Daher: I have to thank Michael Mark for putting up with my stupid questions. This is the great thing about these angel groups is that people always respond to your email, they always have something constructive to say. If they don't know something, they'll say they don't know, but it is so, so valuable to have the support.
Founders, if you've got angel investors on your cap table, use them as a resource, okay? Don't regard them as obtrusive characters that you kind of want to manage. They're a resource for you. They're people who have a stake in your success. Reach out if you need help.
Mike Nolan, F-15 pilot, angel investor, artificial intelligence, and machine learning maven, I thank you for making time to be on the Angel Invest Boston podcast.
Mike Nolan: Thank you, Sal.
Sal Daher: Tremendous. This is Angel Invest Boston. Thanks for listening. I'm Sal Daher.
I'm glad you were able to join us. Our engineer is Raul Rosa. Our theme was composed by John McKusick. Our graphic design is by Katharine Woodman-Maynard. Our host is coached by Grace Daher.