David Miller, Why Do It When You Can Overdo It?

 

116. David Miller

SPEAKERS

Will Bachman, David Miller

 

Will Bachman  00:02

Will Hello and welcome to the 92 report conversations with members of the Harvard and Radcliffe class of 1992 I’m your host. Will Bachman, and I’m delighted to be here today with David Miller. David, welcome to the show.

 

David Miller  00:16

Good morning. Will

 

Will Bachman  00:17

So David, tell me about your journey since graduating from Harvard.

 

David Miller  00:24

Well, it’s it’s 30 years. And one can tell the story in many different ways. One could tell it at the the purely geographical level of a journey, the professional level, family level. Let me try a few different quick passes through, and you can dive in which one is rich fodder. So purely at the geographical level, I went from Harvard to Brown University. I spent two years in Providence before moving back to the Boston area, where I rattled around Cambridge, Somerville, Arlington for about five years, and then headed out to California. We spent two years at in Santa Monica, California, near UCLA, where my wife was pursuing her fellowship. And then we returned back to the Boston area, where we’ve been for the better part of 20 years. Again, rattling around cambered Somerville a little bit before buying a house in Chestnut Hill Newton, where we’ve been for about 20 years in and among that mix, there was a three month stint in Paris, France, which will come back in in this story. So that’s the geographic level at the

 

Will Bachman  01:31

I feel like this is one of these picture books for kids, where you you get, like, one, one sort of transparent layer at a time. So we, we’re gonna, we’re gonna keep the geographic layer in mind as you go to the next layer. I love this. At

 

01:47

the family level, I met my wife, Ruth herzman Miller at the time. Ruth hertzman, October of freshman year at Harvard. We dated on and off for most of the four years at Harvard, and it was off when we when we graduated, but we got back together a couple years later, and were married in 1996 So Ruth hartsman Miller and I have been married ever since we’re going on. Gosh, can I do arithmetic? It will be 29 years come May. I think you point on yours come May.

 

Will Bachman  02:23

How did you meet in October freshman year? It was at

 

David Miller 02:28

the Gilbert and Sullivan production of The Mikado. I was in the chorus, and she was a rehearsal accompanist, and you actually interviewed her on the show a couple years ago, so listeners can go and check that program for all the details, but keeping it at a quick high level. Family wise, we have sort of done the the stereotypical bourgeois middle class thing of get married, launch our professions. Have children. We now have three daughters. We took turns who was raising the children. So I spent some time as a full time dad with each of the kids. They are now 2523 and 17, my three daughters and at various points in school. And all three of them are in school in one form or another. So that’s the family level. Of course, it also is intergenerational, with all four of my grandparents being alive when I graduated, one by one, as they aged out, they passed away, and a couple years ago, my mother suffered a stroke. And so the last two years have been about taking care of my parents and moving them from their long time home to a nursing facility dealing with setting up new lives and closing down old chapters. So that’s one thread of the journey. Two more left professionally, and then what I do not for money so professionally, I spent I majored in mathematics at Harvard. I went to graduate school at Brown in pure mathematics. And after two years, I needed something else. I needed a pause, and so I went to industry for a bit, thinking I’d come back and do more of my math PhD. But I’ve been in industry ever since, never having gone back. So first, I worked in speech and language processing at a company named both baronick and Newman BBN Technologies in Cambridge, where largely I did speech to text transcription work and information retrieval work. These days, you pull anything out of your pocket, pull a foot out of your pocket and talk to it. It it knows what you said, but back in 1994 that was still work in progress, so I worked on developing that building information retrieval, spent 10 years at BVN in a couple of segments in between, went out to UCLA and learned bioinformatics, starting in 1990 99 and so 90 92,001 I was at a bioinformatics laboratory at the molecular Molecular Biology Institute, I think, with the name of it, studying with Christopher Lee, who was the professor there. We swapped. I taught his grad students how to program, and he taught me bioinformatics. And when I came back to Boston in 2001 I started working at a large pharmaceutical company, a Ventus pharmaceuticals, now Sanofi Aventis, applying what I learned, doing what’s called lead generation informatics. I stayed at a Ventus for a few years, took some more time off out of the workforce, and then in 2008 I joined Google, and I have been at Google, working on the search engine primarily for the last 16 years. So the things I’ve worked on at Google include the Google Books project, the people also ask search feature, which I was the first tech lead of created, launched and stewarded for several years, and then I switched to some early stage technology projects which have eventually become what is now launched as AI overviews. We launched this about a year ago. It’s one of the search offerings for the overall umbrella of the generative artificial intelligence Gemini line of work and product. So I’ve been at Google for 16 years as a technology inventor, as a manager, as an independent contributor, a variety of things, all here in Boston, but it’s allowed me to go all over the world, when you talk about the journey, including spending time at the Google Paris office, Google, Zurich Tokyo, in various durations. So that’s at the career, professional level. And finally is, what do I do that doesn’t make money with my time? And there it’s gone through a few chapters. I spent five years teaching computer science in the context of the Girls Who Code Project, which I was the founder of the Newton Chapter, of Girls Who Code back in 2000 and get my dates wrong. So we’re around 2011 I’ll have to check on that. But it was in the earlier days of Girls Who Code, and we ramped up the Newton chapter to be one of the the nationally more salient ones. I ended up distributing curriculum material to several 100 of the club chapters around the country. Hosted Boston Area Girls Who Code meetups at Google. We had a number of nice events to let the different participants from across the eastern Massachusetts area know each other, and then I spent a couple years working with Microsoft TEALS technology and learning in schools, actually going into a classroom at a public high school every morning before work. That’s not quite right, several mornings a week before work, and teaching ninth and 10th grade computer science. So it was an interesting program, linking up practicing professionals with teachers and helping them, over the course of two years, get the professional formation they would need to teach computer science in their classes when they started out as some other kind of teacher, a math teacher, a history teacher, a physics teacher. So I did that for a couple years, and then pandemic hit, and that, of course, shuffled everything. And for me, what came out of the pandemic was a brand new startup. So remember that three month stay in Paris that was in that mix somewhere as part of the I did a work rotation in the Google Paris office, and I discovered French theater. So I had studied French in high school, but put it aside when I got to college. Had a slight occasion to use it when I was at the pharmaceutical company, which was French owned, but really it was the 2013 visit to Paris, where we took the whole family and spent three months there, where I said, this is wonderful. I’m going to go to Paris. I might as well get my French back up and running. So I did. And while we were there, I started going to French theater, three or four shows, maybe six or eight. And when I came back, I thought, This is amazing. I should keep studying. So I continued to study French, just language lessons once or twice a week doing various things. But with pandemic, my wife got me a subscription to a French theater magazine called La von Santia, and I must have read 50 or 70 plays, because they came every every two weeks with a new play. And I got back issues, and I went to the library and got things, and I discovered that there’s an annual Theater Festival in Avignon, France, which is the second largest Fringe Festival in the world. Edinburgh is the best known one, but then Avignon is second largest, with something like 1500 separate productions each year at this point. So when pandemic. Was in recession. My wife and I went to festival Avignon in 2022 and I adored it. Fell in love with it, and thought, why don’t we have something like this in Boston? All it would take is somebody with knowledge of French, knowledge of the Boston theater scene, organizational capacity, some spare finance to get it up and running, nuts. That’s me. I got to do this. So I came home and started a limited liability corporation to try to do it. And it took a couple of false starts and a bunch of networking, but we got our first production up on stage April of this year, 2024 and our second one is going up two weeks from tonight, so november eight, 910, and we’ve got one in the works already on the books and listed for November for, excuse me, for May of 2025 and I’m currently booking venues and kind of signing contracts with productions for our 2025 2026 season. So that’s what I’m doing with my time right now. That does not make money. Just to be clear, theater is how do you make a large fortune in the theater business? Start with a small fortune. No other way around. I can’t believe I botched the joke. How do you make a small fortune in the theater business, start with a large one. So there’s their picture book with four or five different paths through the same 30 years, and can go in any direction you like.

 

Will Bachman  11:32

I love that a math major would be doing this in layers like this. Imagine the dimensional, the grid thinking on this. This is great, wow. A lot of directions we could pursue. I know a lot of listeners would probably want me to ask you to start perhaps expanding about the AI overview piece with Google that so much, you know, all the AI stuff is so much in the news now, and and Google’s AI overview feature is, you know, so much that something that many of us have experienced and also wondered about, wow, is, is this going to sort of cannibalize Google’s other other searches? Tell us a bit about what was involved in in building that functionality that you know so many of us are seeing whenever we search for something.

 

David Miller 12:27

Yeah. So the short answer is, I can’t, which is to say, I’m not permitted under the terms of my employment. I can refer you to press@googledotcom All right. The slightly longer answer is, I recorded four or five years old by now, but I recorded a lecture that I delivered at Boston University about four years ago named something like Bert and deep learning, Bert and the natural language processing revolution. I should find you exactly the title. You can put it on the website or whatever. But there I went over there was a lecture for very gifted high school students that were at BU for the summer. And it really talks through the underlying mechanics of how you do natural language processing and what events just at that point, from like seven years ago to five years ago had really revolutionized things. It made it that now you could do all kinds of stuff that we had dreamed of 10 and 15 and 20 years earlier, but just hadn’t been possible. And so the underlying key development was to transform problems in language to problems in arithmetic. So the key notion was every word gets mapped to some embedding, some representation in a high dimensional vector space in such a way that when you do arithmetic on those vectors, it has a real world correspondence. So, for example, the vector that you use to represent deer, like the animal with four legs that runs around in the back woods. If you multiply it by two, and you look around and say, what vector is near to deer? It’s moose. Oh, okay. And if you do things like I’ll take the vector for King, and subtract the vector for man and add the vector for woman. What do I get? Well, you get some vector, and you look nearby, and you go, oh yeah, Nearby is the vector you already declared was going to be the one for Queen, right? And so doing these kinds of arithmetic, if you tried just to put one number for each word, wouldn’t help, but if you have a high dimensional vector space with 100 or 600 or 1000 dimensions, you can capture all of these semantics and do arithmetic, and at that point you are off to the races. So all of the neural network stuff works by having these embeddings for the words and then letting the neural network do more common. Complicated arithmetic than just add and subtract that I described, and that was sort of these embeddings. Was the birch revolution, which was invented by Jacob Devlin and some of his colleagues, and it was in the New York Times in 2019 and the thing I was working on at that point at Google, called Featured Snippets, we revamped it to use the embeddings and the Bert revolution. And our chief scientist Pandu ni AK, posted a nice public blog post on Bert comes to featured snippets. So those are, those are some of the things going on. But I get the sense that’s not the sort of societal impact questions that you and your listeners might be interested in. And for those, I’m afraid I’m going to have to defer you to Google press office.

 

Will Bachman  15:43

No, thank you for deflecting it very appropriately. We’ll dial it back earlier in time. You know, now we have speech to text everywhere, and it’s, it’s become, you know, for a while you had to pay for it. Now it’s so commoditized, you can do a voice memo and just on your phone, and it just does it automatically. Take us back to 1994 when you started working on that. What would what was, sort of the technology like in those days, and how are you attacking that problem of taking just this sort of audio recording and turning it into, you know, numbers and then turning that into text on a computer. What was that like in those days?

 

David Miller 16:24

Sure. Well, you’re quite plugged in to know all about this. I’m impressed. So I started at BBN Technologies in 1994 working on this, and it was really in the thick of things. And one thing that’s maybe not so well known is that the DARPA, the Defense Administration. It was a Defense Advanced Research Projects Administration. DARPA was a major player in funding this technology throughout, obviously, it had defense and surveillance applications, and it was, you know, funded by National Security Agency, folks behind that, with its own controversial set of issues and questions, but they ran a multi decade research project. Victor Zhou at MIT was the first to demonstrate that he could read spectrograms if you gave him this print out imaging of someone’s vocal tract and the variations of it over time where they were saying things, he could look at it and go, Yep, that says Twinkle, twinkle, little star. And so that was the first proof that this is enough information in the sound waves for an intelligence to flip it back. It doesn’t need to be they have ears. And so that was the beginning of all of this. And so when at Vbn, we were involved in many of these DARPA bake offs, the technology at the time that was state of the art was called hidden Markov models, where you had a temporal aspect of a changing probability distribution that each sound if you made it, or if I made it, it wouldn’t be identical, but it would fall within some probability distribution of how you and I make the Ooh vowel. But there’s also layers of how do you pronounce the word? You know, potato. You say potato, I say potato. So if we’re trying to recognize that six letter word, P, O, T, A, T, O, there’s got to be a probability distribution over pronunciations of it. And then there’s another layer of probability distributions over regardless of what the sounds are, what’s a likely statement you’re going to say? A I just purchased three at the grocery store. What goes in that blank three potatoes, okay, that’s plausible. Now, three tomatoes, also plausible. Three tornadoes. I don’t think you bought three tornadoes at the grocery store, right? And so that’s called a language model. We’re just purely on the sequence of text, the words surrounding even before you look at the audio, you have a guess of what’s more likely and what’s not. So these sort of layered probability models was all the rage at the time, and it was a kind of reasoning called Bayesian reasoning, where you use Bayes rule to say, not what’s the most likely words given the audio, but you turn it around and you say, What’s the most likely audio to come from these words? Right? And use Bayes rule, and you have a generative model where the words generate the audio through some probabilistic model. So Bayes rule we you know, I earned 10 years of salary on applying freshman year probability over and over and over again, because one after another, the problems fell to Bayes rule. Nowadays, nobody uses Bayesian modeling hidden Markov models are nowhere in the last five years of generative AI, everything switched over to deep neural nets. At the time, neural nets was a laughing stock, right? Uh, the fellow at Brown, I’m forgetting professors Navy recently died, wrote an excellent book named neural networks, and in the introduction, he said, I have declared the failure of five neural network revolutions in my career, and I was right four times. So every decade or so, neural networks tried something and failed, and so hidden Markov models was the Bayesian generative approach. Neural networks were untrainable, were unwieldy. We didn’t have the algorithms, the hardware or the data, so those never worked, intuitive. So

 

Will Bachman  20:36

the math really has changed. It wasn’t just sort of a gradual improvement, but it really was a different approach. What what changed? Was it the technology change this memory or the speed of computers got better? Or did you change the algorithm? What happened?

 

David Miller 20:52

It was all three. It was data, it was algorithms and it was hardware. So the arrival of the Internet and then the arrival of people using it and adopting it, and putting everything online, and putting the Google Books corpus online, and putting, you know, there’s now the common crawl sorry. Common clean crawled corpus c4 common clean crawled corpus is a research consortium effort that said we’re going to forward research by standardizing the data. And they essentially sucked up everything on the internet. Wow, a subset of it, not everything, but a vast amount, which was just unfathomable. Back when we were doing it, 1994 we had, I was the discharge in 1998 at our company, and I got to partition. You get a one gigabyte drive, and you get a one gigabyte drive, and you get your research project in our company, not important enough, you only get half a gigabyte drive. And so total, we had a budget like 20 gigabytes for the entire department. That’s laughable. Like this video recording is going to take more than 20 gigabytes. So the amount of data, not just that you can store, but that’s actually out there of text, the number of words in the corpus for language model training. We used to measure in millions, and now you measure it in in trillions, right? It’s just 10, six orders of magnitude bigger.

 

Will Bachman  22:20

So the availability of more recordings was that was one important step is what you’re saying. It just audio

 

David Miller 22:28

recordings for the speech, text language, for the for the language models. It’s just vastly more available on the hardware side, the compute has gotten faster, it’s gotten cheaper, and it’s become worthwhile to make specialized hardware that is tailored for this application. So they talk about a sick application, specific integrated circuits that once you know you’re doing neural nets and you’re all in on neural networks. The core of a neural network computation is a matrix multiplication. And so they have dealt Nvidia and Google with their TensorFlow units. TPU TensorFlow processing units have invested large amounts of money in making specialized, custom hardware exactly for this problem, and it accelerates things by I’m not a hardware expert. It’s somewhere, but it’s more like 100x or 1,000x it’s not 10x and so that has become available. And then the third piece is the algorithms, where the basic neural net computation was developed back in the 1950s Oliver, it’s not Oliver Sacks. I’ll come up with his name. Marvin Minsky did some of it. There’s a bunch of those pioneers. And that was good, but it wasn’t enough. And it turns out there were algorithmic advances in the 80s and algorithmic advances in the 2010s that culminated in in breakthroughs with things like Transformers and yakabuski saying, transformers is all you need. Attention is all you need. You throw a big enough hammer at it, it all cracks.

 

Will Bachman  24:12

So what was it like for you? It sounds like neural nets were not something that you were, you know, learning in school or in the early 90s, what was it? What did it take to, you know, kind of get up to speed and be knowledgeable on that topic you were learning, I guess, along with the rest of the world. But how did you go about learning that that technology? Yeah.

 

David Miller 24:39

So just to be clear, neural networks have been around since before I was born, and they were taught at Harvard and in school I just never take. Took those courses. It was largely a theoretical interest, and the cognitive science people like playing around with it. They became really practical only in the 2000 And 2000 2000 10s. And so as they were coming in and really becoming serious applicable technology, I was, at that point, a manager at Google, and I reached the point where I felt I no longer had the deep expertise that I needed or wanted to be able to make intelligent technical decisions. And so I first just carved out a bunch more of my time to learn and study this, and then eventually went to my director and said, I no longer want to be a manager. I want to go back to being an individual contributor. I found fellow managers to take on what I had at that point a team of eight or 10 reports, because I got tired of sight reading. It’s like a musician where, if you don’t actually spend a long time with a piece of music, preparing for a concert and really delving into it, in your interpretation of it, you can still play anything they put in front of you, but it’s sight reading. And I didn’t think it was fair to my team I was leading, and frankly, it wasn’t that interesting to myself just to have surface level understanding of these things. And, you know, stroke my chin and say, but have you considered making it larger? Ah, well, have you considered making it smaller? Ah, well, yes, yes. Great insight. Thank you. Yes, yes, right? This is you can do it. And being a little sarcastic, there’s two or three levels of expertise below that that I could still apply. But at the base, I did not know what the new technologies were capable of, how they worked, what was important. And so I went back and I said, Hey, I want to be an individual contributor. I studied the code. I studied the papers, I read books, I watched videos, I tried experiments, and it was a good eight or 12 months. This coincided with pandemic, like I stepped down from being a manager. December of 2019, got on an airplane, spent Christmas and New Year’s in France, toured around for a couple weeks in January of 2020. Came home, flexed my fingers, dove into the technology, and eight weeks later, we were all under confinement. So I spent most of pandemic work wise, working on these neural nets that eventually became the things like Gemini.

 

Will Bachman  27:23

That’s such an interesting move. It’s not, maybe outside of Google, at least it’s a, it’s an interesting, unusual move for someone to raise their hand and say, Hey, I actually want to go and, you know, become an independent individual contributor. Again. How is that scene within Google is that, I guess I’ve heard from the outside that individual contributors at Google, in terms of prestige, it can be, you know, it can just be the same people. But what was that like as a decision, and how did people view that?

 

David Miller 27:55

Yeah, so everyone was enormously supportive and healthily skeptical. The you know, I talked it over with my director and with my fellow managers to say what would be the pragmatics of how to do it as appropriate. My director said this will create a number of problems, enumerate them and tell me the solution, because he’s got other things to do, and that’s my job. So I did. I wrote down a long plan, five different ways we could do this, five different strategies, scenarios. And he looked and said, Yep, I agree. Well, we’ll make something like this work. What actually happened with pandemic, of course, was all over the map. My vice president Pandu Nayak came out on one of his quarterly visits, and sat me down and said, David, if you want to do this, you can do this. Absolutely. I support you. Let me tell you why it’s a bad idea. You should keep one or two you always want to have one or two people to to do stuff, because otherwise you’ll have more ideas than than execution capability, and you’ll get a little demoralized and dispirited that there’s nobody else excited about the thing you’re excited about and working on this with you. And it doesn’t take that much to manage a couple people. Come on, just keep a couple people. So I thought it over, and I decided I was not going to go that route. And instead I was going to make sure that I chose projects that were adjacent to my colleague, my peers, and so they would have eight or 10 people working on something, and I would be working on my adjacent thing that has somewhat worked out and somewhat not like I think the people who really thrive at Google as individual contributors at a senior level Find ways to have impact far beyond themselves, and I’ve had mixed success with that, and it’s almost certainly been a career decelerator since 2019 and that’s okay. That’s okay. I have many things I like to do with my time outside of work, and. And that’s how to paying work. And so for it’s to have decelerated is fine. I still get to work on interesting things with interesting people. I launch stuff. I make an impact on our users. I’m okay doing that with less impact or speed than I could have.

 

Will Bachman  30:17

Well, I think it’s a very brave decision. Takes a lot of confidence to make a move like that. So kudos to you. I want to hear a little bit more about your French theater, and what does it take to start a production from scratch in Boston in terms of finding a director and actors and space and you know, and marketing and getting finding an audience. I’d love to hear that story

 

David Miller 30:47

Absolutely. So let’s see. So first of all, this doesn’t come out of nowhere. I sort of told the story, leaving out various backgrounds. Ruth and I, we met at Harvard Agassi theater. Actually, that’s not true. We met in the Cabot house, one of the common rooms, because that’s where we were rehearsing way back in October. But then we spent a lot of time in Agassi theater together and so forth. So after college, going out to the theater was our go to date. And as young, young professionals. We got subscriptions to a couple of the Boston theaters here, and so I’ve been a subscriber at the Lyric Stage Company of Boston for 20 plus years, and then on their advisory council, and then on their board of directors and on their finance committee. So I’d seen what a professional theater in Boston looks like from the inside. It’s a medium sized theater. It’s, you know, two and a half million dollar budget, and they do 100 and some odd 150 even performances a year of six productions. So that’s like, that’s my view of serious professional theater from the other side. As I kept getting more involved in the French work. I got more involved in the French library here in Back Bay Boston, which has been around since the 1940s it was actually in large part, funded by a Harvard Business School professor, Professor Douro, who was one of the inventors of venture capitalism. And so the French library in Boston has got a lovely building in Back Bay, and I’ve been involved with them. I run the theater club there. I am on their board of trustees. And so these two are critical partners. And when you ask, what does it take to do a French theater thing in Boston. The answer is probably key ingredients include being integrated in the French community in Boston and being integrated in the theater community in Boston. So those were the first things, which I’d laid the groundwork for without intention for five or 10 years. Beyond that, we do not create the theater. So far, the production is created by French artists in France. The things I’m looking at, I’ve gone and seen in France, they have played 10s of times or hundreds of times, right? So these are successful productions, and what I do is I show up at a venue, and I see the show, and I say, this is a great show. Let me talk to the person in charge of diffusion, dissemination, touring, and I hand them my one sheet that describes who I am and why Boston’s a wonderful place to come play your French show and the basic outline of the structure of our deal, which is, we buy your show, lock, stock and barrel, you come, your actors, your sets, your script. It’s the original production just brought over to Boston. We then add subtitles, so French and English captions that we prepare and we display on the screen during the show. So this time around, our translator is a professor at MIT. Her name is Claire concession who teaches translation of theater, specifically translation of French theater at MIT. It’s amazing. Claire is such a fantastic colleague. And so finding her went from okay, I can make a translation from French to English to Oh, okay. She has translated Nobel Prize winning French playwrights, great. And her grad students will work on this if they’re interested. Wonderful. So, so that’s another element and ingredient, and at that point, it’s finding a place and gathering the staff and renting the equipment that you need and doing the publicity, all of which turns out to be harder than I would have expected. So you think you just rent a show, you just rent a theater. There are theaters. Surely they are dark. Sometimes you. Can rent one. And that’s partly true. There’s the Boston Center for the Arts that is operated and owned by the city, city of Boston, and part of their mission is to cultivate arts and have theater so that that’s rentable in different sizes. Huntington theater is rentable. There’s six or eight others. If you’re doing this in Chicago or Los Angeles, you can rent a theater two years in advance. If you do get in Boston, you’re lucky if you can rent it eight months in advance. But French artists fill up their calendar 15 months in advance, and you got to get visas for them from the United States government to come and perform professionally, and that takes six to 12 months. So there’s a calendar juggling aspect to it. Great. You’ve got the place, you’ve got the the visas, you’ve got the people coming. There’s contract lawyers and deal memos and such, and then you need somebody actually to do the doing. I don’t know how to hang a light or program a sound board. But hire culture is a service from the Massachusetts Cultural Council. You can post ads. You can hire a contract production manager. You can hire a contract lighting designer. Our arts administrator found us on Playbill, and she’s actually remote, and so she does administration, program design, publicity, Facebook, ads, logistics. So it’s really piecing all of this together. We’re in the process of becoming a 501, c3, non profit at the moment. We have a fiscal sponsorship with the lyric Stage Company, and so that’s how we are able to accept and process tax deductible donations which fund these efforts. And so sort of little by little, we’re evolving.

 

Will Bachman  36:53

So you raise several points that I guess I wouldn’t have anticipated would be issues like finding getting visas for the calendar challenges and so forth. What’s been like actually getting an audience and, you know, advertising it, promoting it, you know, finding people who want to come to a French language theater in Boston which is not full of necessarily fluent speakers, or at least that’s not the it’s not the dominant language in Boston, yeah, obviously.

 

David Miller 37:24

So there’s sort of two versions of this pitch slash answer. One is, oh, I started in the middle. Let me give you the high level thing. Boston’s an amazing place, with cultural institutions all over the place. And that doesn’t just happen. Somebody has to make it happen. And I would love to add to the Boston landscape a new, self sustaining, long running cultural institution that does French theater. And everyone says Boston is among the most European of American cities. We have strong ties to France or hospitals. You know, one of the major hospitals is on Louis Pasteur drive here in Boston. So there are those historic ties. And so that’s what I’d like to create in some version. Should it be a Boston French Theater Festival? Should there be translated French theater into English that goes on stage here. That’s an experimental Let me try various things, but this is what I’ve got. On the other end of the spectrum, like you ask, what’s involved in driving up that audience? Well, you do market research, you discover that there’s something like one to 4% of respondents to social surveys, and five or eight years ago in the Boston area, that say yes, French has been spoken in my house during my lifetime, but bigger than you think, the French Consulate, the consulate, yes, the French Consulate here in Boston has a registration of 10,000 people in the greater Boston area, probably spreading up into main that are French citizens who have registered as, Yep, here I’m living in France in Boston, and I wanted to make sure the consulate knows about me. So there is an audience Boston, aka is a expatriate network that welcomes new French families that have moved here and have the same questions about schools and insurance and housing and taxes that all the previous French immigrants have had. And so I work with Boston, aka we’re one of their sponsors. I go to their events. There’s the French American Chamber of commerce of New England, facney there, again. Their job is to help create businesses, business ties both markets and starting up companies here in Boston. And so that’s a network of professionals that I’ve tapped into for visa legal advice and for publicity finding. Some people who are interested in being on our board of directors. So all of these both demonstrate that there is a fair amount of French community here in Boston, and are the pathway to make a new pillar of French community erected.

 

Will Bachman  40:17

I love that. I think it’s so instructive on how you had these relationships and were that you weren’t intending to result in something, but you had been pursuing your passions with the French library and with the theater, and then those kind of came together to enable this. I want to ask you just share, share one story about your involvement with girls who code that you’re that you’re proud of, just you know, maybe share one instance, there’s something that comes to mind from your your work with them, that that You feel particularly proud of.

 

David Miller 40:56

Sure. So I would say that the girls who code is an after school program has a few programs. We ran the After School Club. We were at the Newton Public Library. We had two sections of about 20 girls each. So that was about 40, but we were in network with another six or eight around the Boston area. And I organized a meet up, a year end event for everybody at the at the Google office in Cambridge. And there were various demonstrations and socializing things. And then there was some ceremony and some keynote speech. And I got tapped to fill in for one of these. So I gave some discussion about why this mattered, and gave the pitch that I had given at several of these parent meetings at the end, which is, you’re in this room, you look around and you see all these young girls and all these young women who are programming, who are getting into technology, and you you feel like you belong. And if you pursue this career, you’re going to go out and you’re going to get a job, and you’re going to be in rooms and you’re going to look around, and you might just see one or two women in that room, in that workplace, at that meeting, and you might be the only one. And I want you to look around now and remember that you belong. This is where you should be. That room we are. The only one is where you should be, and it’s a two way street. That room belongs to you. You have ownership, you have influence. You’re in the right place, and you can make it what you want. And I got a little teary as I gave that pitch. I had given it to groups of 25 kids and their parents a few times, but here I was in front of 150 and Okay, Booz, fine. Afterwards, one of my Google colleagues came up to me. She had been at Google longer than I had. She’d been at Google for 15 years. She was one of the top people in the indexing, crawl the web and index it and make it that Google can do the search. And she came up to me, she said, I have never heard that statement before, David, and thank you. That made me a little teary. I need to hear that more often, and so I’m glad that I found something to say to this room of 150 teenage girls that potentially could sustain them through a 25 year career in the industry

 

Will Bachman  43:30

I love that. Let’s go back to Harvard. Tell me about any courses or professors or activities that you were involved in that continue to resonate. And you’ve already mentioned, you already mentioned, your theater experience clearly has played an ongoing role. Any courses or professors that that resonated in some way?

 

David Miller 43:57

Yeah, so it’s a good question. There’s so I spent a lot of time at Harvard teaching. I was a TA in the math department for five different semesters, and as part of that, I got to work with Deborah Hughes Hallet, who, at the time, she was running the calculus project and writing a textbook. I wasn’t involved in that. Wasn’t involved in that piece, but I was her TA for math 1z teaching calculus, and just being able to watch how much she cared about teaching and the ways she approached it, I think very much influenced the way I approach subjects. And I still don’t know where’s the boundary between teaching and mansplaining, but I try to do the teaching side of it at every opportunity, whether it’s my colleagues or my children. I try to avoid pontificating. I’m sure I fail, but this notion. That you understand the world by teaching it to somebody else. Is something that I feel is deep in my way of looking at the world. And I some people think by writing, I think by teaching

 

Will Bachman  45:16

David, you’re a very high agency person, and which is very inspiring to speak with. For listeners who want to follow up and find out more what you’re doing or keep track, reach out to you. Where would you point them? Online?

 

David Miller 45:31

Sure. So you can go to my website, French Theater projectdotcom, it’s the English way of spelling, theater, T, H, E, A, T, E, R, so www.frenchtheatreprojectdotcom you can contact me directly, david@frenchtheatreprojectdotcom if you want to get some background on the craziness that I was doing in pandemic, you can look at Mr. Miller, M, O, N, S, i, e, u, r, Monsieur millerdotcom where I wrote several 10s of 1000s of words on being a French language learner, but best is David at French Theater projectdotcom

 

Will Bachman  46:07

amazing. Okay, we will include those links in the show notes. David, I’m grateful for you joining today. This is a fabulous conversation. Well, I

 

David Miller 46:15

appreciate you having me on and good luck with the ongoing class of 92 report project. You.