Conversation with Timothy Chou
Good day, and welcome to edition 132 of our digital industry leadership series. Today I'm quite pleased to host Dr. Timothy Chou, chairman of the Alchemist Accelerator, who is considered one of the early leaders in moving enterprises to the cloud, beginning with his tenure as president of Oracle on demand. 2004 he published the landmark book the end of software which foretold the rise of enterprise SAAS and featured three startups that we know well today Salesforce, NetSuite, VMware. In 2005, he launched the first cloud class on cloud computing at Stanford University, cs309a.stanford.edu. Today, he is the chairman of the alchemist accelerator and serves on two public company boards Blackbaud and Teradata, much of his time these days is focused on the pediatric cloud project, the project to connect all 1 million healthcare machines in all of the Children's Hospital in the world. Duck show, welcome to our digital industry leadership podcast today.
Hey, thank you, Ken, I really appreciate the invitation. And I'm just amazed 130, that's a lot of podcasts.
A million healthcare machines. That's a lot of machines. And I will look forward to talking with you quite a bit about that. So I always like to start these off to understand a little about somebody's professional journey and what got them to where they're at now. I always think about it as the red thread, what would you consider to be the red thread through your professional journey?
Yeah, it's, an interesting question, I think I've always been interested in what's next, and thinking a lot about what's next. And having the opportunity to do or move to what's next. Much of my graduate work was in VLSI. I got out of school and started teaching at Stanford in 1982, the introductory computer architecture class, because most of my work was in computer architecture, computer architecture is a very fun, intellectually interesting area. But even back then, it was fairly obvious that the era of having many different architectures was coming to an end. And so my very first job was an operating system. The hardware architecture is cool and fun, but there won’t be much of a future in it. I progressed out of operating systems ended up going to Oracle to help Oracle get out the door. At that time it was the largest database software release in the world. My second tenure at Oracle was running the beginning of their cloud computing business, which we started in applications, financials, HR, all that stuff. I moved into apps and it's funny that as time has worn on, it's almost like I recycled again and I got very interested in delivering computing infrastructure as a cloud service. A lot of the work that I'm doing today is really around what does this mean at the edge? I’m always thinking about what's next because that's where all the fun is, being at the beginning!
Absolutely. I love the theme. What's next? In fact, I think we're gonna use that as the title for this session. I'm impressed, not just starting the hardware but it's a Ph.D. in electrical engineering, you said, you kind of went up the stack OS apps cloud. Really fascinated with your time at Oracle, what a great time to be in that company, given all the changes that were happening at the time, and of course, the whole move to the cloud. I think it was just a couple of years earlier that Microsoft quote-unquote, identified the internet as a real potential direction for them. I know some of the early work during that time at Oracle was focused on the network computer. I guess some famed what I call false starts around to help spin NetSuite and Salesforce. What attracted you to Oracle at the time? And what were some of your key accomplishments?
Yeah, it's kind of an interesting story. This is late 99 or mid 99, the heyday of the first bubble. In the morning, I went on an interview with a startup company to be the CEO. They were going to be funded to the tune of $50 million once they hired the CEO. It was a website targeting the teenage Korean market. I mean I'm sitting there going, Well, what do I know about? Like, hey all this stuff. I mean, it's happening. The Webvan, the eToys are all happening. Okay, maybe I don't know, anything, I should go talk to this guy. And it literally in a garage in Palo Alto. That’s the morning of that day, the afternoon of that day because I made mention that I had worked at Oracle, in the mid-90s and still knew a lot of people there. They said, Hey, you got to come and talk to Larry about this new business, we're going to spin out called Business Online. And so I said, okay, and that afternoon I go over to the towers at Redwood Shores and go up to the 12th floor and Larry's big office. I still remember I asked him, why are you interested in what I called a fat Exodus strategy.
I ended up going back to Oracle to be CEO of Business Online, a totally separate business, delivering Oracle Applications as a service. Within the first six months, I realized this had nothing to do with the servers and it had everything to do with how you manage software, and how you structure software. And so at the end of the day, I said we can spin it out, it would become what I refer to as a mini EDS, which is what's the point of that. And so rather than spin it out we began a mainstreaming process, meaning-making this part of how the software business began to run. The pattern we develop was around how you take a traditional software company and move them into the cloud. It's been copied by many traditional software companies. Of course, we never had the advantage that Salesforce and NetSuite had building software from scratch. If you build it from scratch you will engineer it differently than what you would do in the old on-premises world.
The very first lecture I did at Stanford is about software economics. If you engineer from scratch, you can get nearly a 10x improvement in cost structures, versus, the traditional model trying to move into this. Obviously, the advantage of the folks trying to do that in the traditional world is they do have customers and those customers have bought on-premises software. So if you can standardize the management of that software, you can derive incremental revenue from an existing base that you never were able to mine before. So that model was, copied by many of the traditional software companies, but obviously, the explosive growth occurred, but the companies that basically said to hell with it, we're going to build financials from the ground up or CRM or purchasing, etc.
No, but think about all the different stories that came out of that time and all of the different success and patterns that we have it so was a really influential time and as you say, right before the first meltdown, they're so interesting timing. In 2005, you launched what was the first class on cloud computing at Stanford University, which I think you just referred to a moment ago, cs 309 a.stanford.edu. Was this viewed as a dramatic shift of the computer science curriculum or was it simply an evolution from so-called ‘Thin client computing’?
I think at the time it was what are you talking about? I mean, the origin of that class is kind of curious. I taught introductory computer architecture, CS110/EE181 for 15 years, as a part-time hobby. In essence, I started out working full time at Tandem computers and then on to Oracle. I go just teach class two times a week and go back to work. One day, I had to fly to Bali to do a sales kickoff meeting, and get back in 36 hours to teach a class. So I went, this is getting a little stupid. So I decided to give myself a leave of absence. So that leave of absence lasted six years so when I left Oracle, I was hanging around their apartment, and they said, Oh come back and teach. I was like, God, that's a lot of work. And they said, Oh, teach a seminar class, I said, Well, that would be fun. You know, I've been a manager, a lot of yours, great people, great opportunity, and I put them together, I don't have to do any work. So we started this class as you might guess, the first and last lectures are nearly identical because it's the ones I do. And in between, I have guest lectures.
We've had, if you go to that you were reading the website, you'll see it's pretty much who's who, early on, it was CEOs of companies that I knew, like Subrah at WebEx or Mark at Salesforce. And then over the years, I began to realize that most people would not be offended to be asked. So it turned into a class in which the students will say, well, it's just people Tim wants to listen to. Most of the speakers come from the enterprise space but we've had Tony Shieh, rest his soul. I got interested in the whole IoT thing. We had the vice chairman of GE come we had the CEO of Agco one of the big agricultural manufacturers, Patti Poppe, the CEO of CMS Energy, one of the very few women CEOs of a fortune 500, who has just joined as CEO at PG&E. I didn't do class last year, because you know, the pandemic, etc. But the year before, we actually had Eric Yuen, from zoom came before zoom became a verb, a noun, a way of life. I think the unique thing about the class is, on the Stanford campus, there's tons of startup activity going on but in this class, pretty much every CEOs CEO of a Fortune of a public company, or a company about to be to go public.
Reed Hoffman came, right before LinkedIn went public as an example. But it's people who've had a long track record of success. And one of the things I think that the kids get out of it is, many people see the CEO of the tech company, as a hoodie-wearing engineer, etc. And this was a way of letting them see many different variations of what CEO means. And learn some of the lessons, there are hundreds of stories that came out of that class.
My other motivation was really, to kind of try to bring the rest of campus into the computing world, because you'd be on campus and people say, oh, there's computer science and they point to the Gates building and I look at him and go, Well, why are you pointing at the geology building or the mechanical engineering building or the sociology building. Part of this was very deliberately reaching out to the med school, the law school, liberal arts and try to bring them into this to have them see that the implication of technology on the world is is all around And hopefully getting them to participate in that as opposed to standing to one side.
You actually accelerated that quite a bit by joining the Alchemist Accelerator in 2013. I think one year after your fellow lecture, Ravi Belani, founded out of the work he had done at the Harvard club, tell us a bit about Alchemist and your role there.
Ravi, is the main man behind Alchemists. Alchemists, we like to think of as the premier B2B accelerator. We've not focused at all on the consumer side of the world. A lot of times, we talk about how it's nonsexy, but as people know, there's a lot of money that sits behind enterprise software. So we're very much in the mold of many accelerators, and we have a lot of applicants, we ferret through a certain set of them, we end up with maybe 20 to 30 of them that enter a class, that class, let's call it three to four months and ends in a demo day. And along the I act in the mentor role for some set of them. And there is a set of these, I met at the beginning of their journey. Edith, CEO of Launchdarkly just got her company on that big NASDAQ billboard thing as one of the promising startups to watch. I still remember her walking into Demo Day.
I’ve also been working with Asim, CEO at Yotascale which is delivering cost observability software, for cloud computing. Another one called Oomnitza which is building software to do modern asset management in the connected state. So it's been fun to see the progress of many of these young companies and, get involved and hopefully help them as that can be very rewarding.
You know, I see that you've invested in more than 400 startups over 34 exits. So hopefully, that's been a bit rewarding along the way, beyond, of course, just the satisfaction you get of working with, great people. What have you seen as some of the success factors for those companies that have gone on to ultimately win?
As you know, the team, the people matter a lot. I think another thing that matters a lot is patience, and perseverance. I know a lot of times, we're talking to these companies who are two guys and PowerPoint slides or two guys, and a little bit of software, I said the journey is long. And if you don't believe in the mission, and what you're trying to do, the world is not in your on your side, they're going to tell you what you're doing wrong and it's not going to work. I mean, just go back in time, I still remember we had customers that came in and wanted to see the computers thir applications were running. They would go up and literally shake the cage. Seems kind of odd now. What the hell are those people thinking?
I think that perseverance, patience, and drive through a lot of dark times, and there'll be way more dark times and there are good times. That's really what you know, I think separates if you have good quality people, and there are perseverance and drive, and then obviously luck is incredibly important in the conversation but you know, as many people say, you manufacture your own luck. I think those are the things that that are really ended up being important
You’ve done so many startups have these patterns really begin to merge these meta patterns in some sense. I sense between this and your this filler, if you will, CEOs you've had at Stanford, you could write some really great books, just on the meta patterns that emerged from both of these. So I know the pandemic sadly, is brought in a new living and working pattern and you're certainly seeing that in what they call the Texadoss. I think people leave in the Bay Area for Austin. What do you see as The future of startup-based innovation and what do you see as Silicon Valley's role in that?
Yeah because I've been around here a long time, I've seen the old, Silicon Valley's, it's time is coming to a close. I've never thought that that's really how to think about it. I think it's more like, as we know, tech is pervading everything, consumer side, enterprise side, etc. And so shouldn't there be multiple locations on the planet that this type of innovation is fostered and encouraged. And whether that's Texas or Shanghai or Berlin, I think everybody on the planet has a role to play in this. I have been asked how to replicate Silicon Valley in other locations. I tell them it’s not creating a campus and we're located next to a university.
The unique thing is about Silicon Valley is the concentration of tech talent is so high. And by that I mean you're out with your kids at a soccer game, and mom sitting next to you, right works at Google, or Facebook, or whatever. And you strike up a conversation and out of that comes, oh, well, we should get together and talk about this. And that very informal network is pervasive here because of the concentration.
Then I think the second thing, and I always hope that this continues is the ethic here. One has to give credit to Hewlett and Packard for developing the ethic of helping as an engineer, I want to help you, I'm not going to help you with if I'm competing with you, okay, fine, right. But outside of that, if I can help you, I will. And I think that, again, a lot of places in the world, if you say, Well, I'd like people to go well, why do you want my help? And what, what's in it for me? And I think that those two things, the density, and the ethic, have been very instrumental to how this has prospered.
I don't know if that's the same thing that is going to happen in other locations but it is something that is so much part of the water here that I don't see it going away anytime soon. It's great, the whole pandemic illustrated what we in tech have done for a while. I mean, we don't need to be right next to each other, right there is there are technological answers to this. But I do believe that the physical concentration you see here, particularly after we all get vaccinated, is extremely useful. I'm not moving to Austin anytime soon.
Well, I don't know what the bay area would be like without you there. So I think you need to stay there and continue to pay it forward speaking to the Bay Area ethic that I'm so proud of as well. Speaking of really paying it forward and giving back. I'm really intrigued by your work on the pediatric cloud project, as I said earlier connecting a million healthcare machines and all of the Children's Hospital in the world. Can you tell us a bit about the work?
Like all these stories, they kind of have your, your red thread thing. So this story starts about five years ago, I'm teaching this class on cloud computing at Stanford and we have I'll say 90% of the kids are local course, I'm talking about the past, 90% are local 10 or kind of remote online. And one of the remote students sends me a note and says, Oh I'm a little old school. To meet the professor. So, you may remember California Avenue, there's a place called Johnny's cafe. So we made it Japanese for breakfast. And sitting a guy walks in, I'm looking at him going doesn't look like a regular student.
Turns out he has an MD and MPH, an MBA, and he's Chief of Pediatric Cardiology at Children's Hospital in Orange County. So I look at him and think what do I have to say to you, I don't know about any of that stuff. And he says, well I've decided that AI and big data and cloud computing need to meet and so I've decided to come back and get a master's degree in bioinformatics, at Stanford. Later on, we're talking to the guy who heads up bioinformatics and he says, yeah, we saw his application, we went out what the heck. I'm speaking to those of you who've ever met Dr. Anthony Chang, he takes three and a half years to finish a two-year program, because number one, he has no idea how to program.
Number two, he's still on call as chief of pediatric cardiology at number three, at the ripe old age of I probably won't get this exactly right. But I'll say he's in his early 50s at this point in time, he decides to adopt an 18-month-old and a six-month-old as a lifelong bachelor.
A classic underachiever.
I always tell people if Anthony can do it, I mean, why, why are you saying you can't. So, Anthony is my portal into their world. And so what I started to learn a lot, and I got invited, he's very big into innovation and just released a book on artificial intelligence in medicine. And what I started to learn by meeting his tribe was, people who've been in this know this, but even to this day, the primary means of data sharing for an MRI scan is called a CD ROM. By the way, with a little sticky note on it with the password. The implications of this, a story that actually occurred a little over a year ago, in Southern California, a kid went undiagnosed with brain cancer for a month, because they couldn't they had to keep re-imaging him because they couldn't move data around. And you're going, are you kidding me? So on the one hand, you hear stories like that. I’m on the board at Blackbaud and as a result of that, I know the CEO and save the children. I've known her for quite some time and I described a little bit how we were thinking about and she goes, well the number one killer of kids in Africa is pneumonia. And then you start getting more involved with the whole AI in medicine thing, Matt Lungren runs the AI in medicine program at Stanford. And time and time again, they all say the same thing, which is, we can't build these AI diagnostics, because we don't have enough data. And so as a computer guy I'm sitting here going, well, this is ridiculous!
What is going on here is your guys are living in a disconnected world and they're disconnected inside the hospital, and they're clearly disconnected across hospitals. And I looked at it and it's kind of interesting when the internet-connected a million computing machines, this is roughly 1994. That's when eBay and Netscape took off. That was critical mass. It turns out there are about 2000 healthcare machines in every hospital. Healthcare machines include MRI scanners, ultrasounds, ventilators, blood analyzers, etc. Furthermore, there are about 500 children's hospitals in the world, which most people are kind of surprised to hear how low a number it is.
So that's a million healthcare machines. And so about a year ago, we launched a project and a company to connect all million healthcare machines and all the children's hospitals in the world because we all know connecting computing machines changed how we buy stuff. Wouldn’t connecting 1,000,000 healthcare machines fundamentally alter children's health care on a planet-wide basis. And it's not that hard. But from a technological perspective, we believe we've come up with a very innovative, scalable way to go do this and it's eminently doable, technologically. The other question we get asked is why pediatrics? It’s a unique market where you realize 25 of the 500 are the leaders. And if you move those 25, you'll move all 500. Stanford is one of those. We did the first deployment of what we refer to as the edge cloud server into Stanford, two months ago.
Yeah, it's crazy. On the one hand, we know it's kind of a moonshot project. But the other side of it is that it's not. We could do this in less than four years and it would fundamentally shift everything. I mean, I'll just give you a simple example. Pneumonia is the number one killer of kids in Africa. Well, it turns out, it's not there has been introductory work on how you build pneumonia diagnostics, actually, Stanford's done a lot of that. But the challenge problem is you can't feed it enough data. Well right now, we're staring at the idea that you could take 10 hospitals, fully instruments, their ultrasounds, and build 100 terabyte database in less than a year, which would completely eclipse the current imaging Imaging Data Commons project that NIH started about a year and a half ago, which is all on oncology for adults, that database is only a 1 terabyte. We see a clear path to doing this.
Well, once you've done that, I can train the algorithms to do pneumonia detection. And then the cool side, you were talking a little about edge computing earlier is, now I can take that inference engine and run it directly at the edge. So now, the challenge in Africa is not low-cost scanner technology, there are tons of In fact, Butterfly just went public, there's tons of low-cost scanner technology, the problem is there's nobody to read the scans. I was just talking to one of the leading pediatric doctors in Rwanda and this is interesting to think about, in the entire country, there is one pediatric cardiologist.
So the promise of what AI could mean on a global basis, which, by the way, I don't think there's a whole lot of difference between Rwanda and rural America, in terms of how much expertise you have out there. And the ability to move that expertise out to the point of care is like we're in earshot of doing stuff like this, which, the implications are enormous. And I think this is a real opportunity. A lot of times when I'm talking to the kids in class, I like to say I have an opportunity to warp young minds. And so I tell him, I got at the end of the day, that is the big challenges for the future, are all in how are we going to deliver healthcare, education, food, water, and these fundamental infrastructures and then how are they transformed by technology! If you want to really go do something for good, let's go work on these problems. These are the big problems out there. And yeah, I know, there's plenty of people who want to sell you more ads. But maybe if you're not so interested in that part of the problem, this part of the problem is, at this point, is still untouched, nobody is really addressing how you use computers and software to fundamentally reshape these bedrock industries that we have.
What a great cause and you look at all of the impacts this has clearly on the children, the hospitals, life as we know it, and, even going back to computer architecture, right, and the ability to do to prove so much of that out. It is really is the perfect storm of opportunity, I understand why you're so excited about that being what's next for you, and, for the industry. So thank you for this insightful interview today.
I appreciate it, Ken! I have known about your work and your team's work for quite some time. So it's really a pleasure to meet you.
I think there's a lot of mutual admiration and I know especially the work you've done at the Alchemist. There's certainly a good Venn between all of us in that. So this has been Dr. Timothy Cho Chairman of the Alchemist Accelerator and I'd say the king of what's next.