Data Science Hangout | JJ Allaire, Posit PBC | Making data science more open and collaborative
We were recently joined by J.J. Allaire, Founder and CEO of Posit PBC. What made you so interested in open source initially? It was really that the first big software projects I worked on were proprietary software, and I found it disappointing that proprietary software products kind of get very tied up with the company that sponsors them – the fate of that company, the other products that that the company might offer, or who that company gets acquired by really end up affecting the fate of the projects. I also found that while some proprietary software is either cheap or low cost, there's a certain impediment to adoption associated with the price. As a creative person, I want my work to be available as broadly as possible. I like to make the work available to as many people as possible, but also there's a dimension of the durability of the work. Is the work going to be around in 30 years or 40 years? What are the things that would make it? So that's what made open source appealing to me. What are you most excited about at Posit in the next year ahead? It's going to be related to Quarto, of course, because that's what I'm spending all my time on. I would say that the thing I'm excited about is that we recognized Quarto is really powerful and flexible, and easy to use for a certain subset of users who are very technical and very motivated, but we actually want to make Quarto available more broadly. So working on tooling that lets both technical and non technical people collaborate over documents, and also lets some less technical people participate in using Quarto. (You can sort of see some of this work in the visual editor that's in RStudio) Those are the kind of things that I'm focused on for the next year, and I'm excited to see those getting realized. In hindsight, what is one of the best decisions that you ever made for your career or your education? I think this actually has a lot of commonality with talking to a lot of people about their careers. I think when I was about in my late 30s, I had been involved in both developing software and also starting companies. I really learned about myself, that the company part of things – the management and entrepreneurship – I really did not like it all. I didn't enjoy it. So I kind of said, I'm happy to be involved in starting companies but I'm absolutely not going to do that part, and I'm going to have to find a partner or other people who are excited about doing that part. What I really want to do is focus on engineering and product development. It's very easy as a company founder to get pulled into all kinds of other things and that's what happened to me the first 2 times. Just being super clear about that, and saying I won't even do it unless I can satisfy this condition. I think that's pretty broadly applicable in that a lot of us accumulate a lot of responsibility in our careers. Some of it is necessary and important for a given role or company, but then eventually being clear about, what do I really like to do and really want to do? What do I feel like I should do or what is put upon me? I know a lot of people now who are in their 40s who have actually managed dozens of people and are like, yeah, I don't really want to do that. I don't ever want to have anyone work for me anymore. So I'm going to be really clear about that. I'm going to walk in, and people are going to try to give me people and try to make me a manager, and I'm not going to do it. I think it applies pretty broadly, but just knowing yourself and setting pretty rigid constraints about what you're willing to do in the workplace and not. Everybody's different, and both can be really rewarding. I know a lot of people who find management very rewarding, but I know a lot of people who find it really alienating. ► Subscribe to Our Channel Here: https://bit.ly/2TzgcOu Follow Us Here: Website: https://www.posit.co LinkedIn: https://www.linkedin.com/company/posit-software Twitter: https://twitter.com/posit_pbc To join future data science hangouts, add to your calendar here: pos.it/dsh (All are welcome! We'd love to see you!)
image: thumbnail.jpg
Transcript#
This transcript was generated automatically and may contain errors.
Hi, friends. Welcome back to the last Data Science Hangout of 2022. I hope you all are having a great week. If you're joining us for the first time today, thanks for coming to Hangout. I'm Rachel. It's so nice to meet you. I see a bunch of people in the waiting room, so let me just press the Admit All button one second here. But this is an open space for the whole data science community to connect and chat about data science leadership, questions you're facing, and getting to learn about what's going on in the world of data across different industries.
So together, we're all dedicated to creating a welcoming environment for everybody, and we love to hear from everyone, no matter your level of experience or your area of work as well. So there's always three ways that you can ask questions here in this space and also provide your perspective. You can jump in by raising your hand on Zoom. You can put questions into the Zoom chat, and feel free to put a little star next to it if you want me to read it if you're in a coffee shop or your dog's barking or something like that. And then third, we also have a Slido link where you can ask questions anonymously.
And I bet Hannah is sharing that in the chat here in just a second. If this is your first Hangout today, I thought it'd be nice to see maybe have you say hi by putting a one into the chat. And if you've been here before, put a two. For those who have been here before, many times we'd love to all help welcome you in as well. And so just to let everybody know, we do share the recordings of each session up to the pause at YouTube. So you can always go back and rewatch or find helpful resources. We do also have a LinkedIn group for the Hangout too if it helps you connect with each other.
One more thing on the topic of recordings and resources. I am hoping to take a quick pulse from the community this week on what has worked well for you, what hasn't and any suggestions you have for next year. So we'll share that link in the chat in a second too. But I so appreciate the thoughtful feedback that so many of you have provided already, not only in a survey, but throughout the year as well.
So with all of that, thank you so much for being here. Let's jump in. This is the first time my co-host is in the same building as me, not in the same room, but so happy to be joined by my co-host today, JJ Laird, CEO and founder of Posit. So I throw in the formerly known as RStudio. Thanks for joining us, JJ. I'd love to have you introduce yourself and share a bit about your background and maybe also let us know something you like to do outside of work too.
Okay. Well, very briefly, I'm obviously do lots of software engineering and product development now, but when I was younger, I studied political science and I got very interested in data analysis. But I could took a deep detour from that and got into writing software and worked on lots of different software products. And about I think it was 11 or 12 years ago, I kind of became aware of R and I was actually looking for opportunities to work on open source software. And I thought it was tremendously cool that sort of the thing that got me interested in software in the first place and kind of the first thing I wanted to do, which was data analysis, had this really, really vital open source project associated with it. So I kind of just said, let me get involved with R. And then I've been doing that at RStudio and now at Posit for, I think it's been 11 or 12 years.
So I started the company and I worked on the RStudio IDE. And then I worked quite a bit on R Markdown, a little bit on Shiny, some stuff related to TensorFlow. And then about two or three years ago, I started a new project called Quarto, which was sort of a re-imagining of R Markdown, both in terms of kind of like, if we had to do it all over again from the beginning, what would we do? But also to make R Markdown multi-language, to make it so it could work for R and Python and Julia and JavaScript and kind of not really make it available for anyone who's doing scientific communication or doing data analysis of any kind in any language. So that's what I've been working on and continuing to work on now. So, and I get my hands in lots of other different projects here at RStudio too. But so, and then, oh, what do I like to do outside of work? I like to play Magic the Gathering.
What drew you to open source?
Awesome. Thank you. So I'd love to ask you, what made you so like, interested in open source initially? Yeah. It was really that I actually worked on the first couple big software projects I worked on were proprietary software. And I found it disappointing that like proprietary software products kind of get very tied up with the company that sponsors them and sort of the fate of that company and the other products that that company might offer or who that company gets acquired by really like end up affecting the fate of the projects.
And I also found that, I mean, some proprietary software is either cheap or free or low cost, but there's a certain, you know, impediment to adoption associated with the price that as a, you know, as a creative person, I kind of want my work to be available as broadly as possible. So I kind of both, it's like make the work available to as many people as possible, but also there's like a dimension of like the durability of the work. Is the work going to be around in 30 years or 40 years? What are the things that would make it not? So that's what made open source appealing to me.
I kind of both, it's like make the work available to as many people as possible, but also there's like a dimension of like the durability of the work. Is the work going to be around in 30 years or 40 years? What are the things that would make it not? So that's what made open source appealing to me.
What's exciting at Posit in the year ahead?
I'm also curious, what are you most excited about at Posit in the next year ahead? The next year? Well, it's going to be related to Quarto, of course, because that's what I'm spending all my time on. I would say that the thing I'm excited about is we're, and you can sort of see some of this work in the visual editor that's in RStudio. But we really like, we recognize that Quarto is really powerful and flexible and actually easy to use for a certain subset of users who are very technical and very motivated, but we actually want to make Quarto available more broadly. And so working on tooling that both lets technical people collaborate with non-technical people over documents, and also lets some less technical people participate in using Quarto are the kind of things that I'm focused on for the next year. And I'm excited to see those getting realized.
Thoughts on AI tools like ChatGPT
Sam, I see you put a question to the chat. Do you want to jump in? Yeah, I just wanted to see what your thoughts were on AI tools like ChatGPT and GitHub. Yeah, I think that I'm a little concerned that companies are rushing these things out in the sort of excitement and novelty of them and have not thought through the implications of their broader use. Something like ChatGPT is not really like, the creators of it don't have full control of it. You can see that in the safeguards that they attempted to put in, but then you can circumvent the safeguards and get it to tell you things that it's not supposed to tell you.
So I think things that are happening in AI and generative AI specifically are exciting, but I feel like technology companies and technologists just sometimes feel like we should do this because we can or we should do this because it's cool and haven't thought about the broader social, political, economic implications. So I feel like certainly exciting to see us come up with AI that is that powerful, but I think we're going a little too fast. That would be my position.
Best career decisions and knowing yourself
Olivier, so you also asked a question in the chat. You want to jump in? Sure. I was wondering, because I've been doing a lot of retrospection lately. Yeah. In hindsight, what is one of the best decisions that you ever made for your career, your education? I would say the best decision that I've ever made is probably that, and this actually I think has a lot of commonality with talking to a lot of people about their careers. I think when I was about in my late 30s, I had been involved in both developing software and also starting companies, and I really learned about myself that the company part of things, management and the entrepreneurship, and I really did not like it all. I did not like it. I didn't enjoy it.
And so I kind of said, I'm happy to be involved in starting companies, but I'm absolutely not going to do that part, and I'm going to have to find a partner or other people who are excited about doing that part, and what I really want to do is focus on engineering and product development, and I was really, it's very easy as a company founder to get pulled into all kinds of other things, and that's what happened to me the first two times, and just being super clear about that and saying, I won't even do it unless I can satisfy this condition.
And then, so I think that's pretty broadly applicable in that a lot of us accumulate a lot of responsibility in our careers, and some of it's necessary and important for a given role or company, but then eventually being clear about what do I really like to do and really want to do, and what do I feel like I should do or what is put upon me. I know a lot of people now who are in their 40s who have actually managed dozens of people, and are like, yeah, I don't really want to do that anymore. I don't ever want to have anyone work for me anymore, so I'm going to be really clear about that. So I think it applies pretty broadly, but just knowing yourself and setting like pretty rigid constraints about like what you're willing to do in the workplace and not.
R versus Python
One was will Posit work to make Python as enjoyable to use as R? Well, we have a different kind of a special relationship with R in that, you know, we have the tidyverse. We've sort of defined this sort of system within the system that's totally internally, hopefully as much as possible, totally internally consistent, and it's really like really well tooled end-to-end. I don't know that we'll ever be able to have that sort of relationship to Python, because Python is, there's just so vast, and people are piecing together so many different things from so many different places, so I think we'll make contributions, like we're definitely focused right now on like doing Quarto for Python and Shiny for Python. Can we, I just don't actually think that any company is big enough to like do for Python what we sort of did for R, so I think we'll make it more pleasurable in places, but I don't know that it's going to be anything like what we've done with R.
You, I see you put a question into the chat a bit earlier. Do you want to jump in next? Sure. Hi, JJ. Actually, my question is kind of piggybacking on what you just said, so R as a programming language is almost 30 years old, and R versus Python is an eternal topic, so where do we see R in the next 10 years? So actually R is even more, it's even like over 40 years old, if you count S, which is its precursor. I think what we're trying to do with R is, and you can see like a lot of the stuff that we've done, and maybe it's not all super visible yet, but like some of the things we've done in the tidyverse in the last year have been like focused on like making error messages much better.
And so like I feel like R has the potential to cross the, like go over the transom from people who are, do not program, do not think of themselves as programmers, have not programmed to do successful, you know, kind of code-first data science. I think R is actually pretty uniquely positioned to do that, and I feel like we kind of are just getting started with that, so the people who today use, you know, they may use Excel or SPSS or, you know, Tableau or, you know, they use, they do, they're doing data analysis, but they're actually not able to ask and answer all the questions they should. I see R as being like fitting right in that spot of the best way for someone who thinks of themselves, I'm a biologist, I'm an economist, but I need to answer like difficult questions about data, so I see R getting better and better and better in that, in that vector, and that's certainly where, like where we're putting our time, at least in the Tidyverse team right now.
So we're now getting confessions into Slido, too, as someone said, so this is a confession and also a call for suggestions. I love dplyr, Shiny, and R, but I'm switching to Python simply because it's listed in a lot more jobs, and gives a better starting point for a possible career change in other areas of programming, so what would you advise for someone in my position? Well, there are a lot of ideas in R that you, you're familiar with, and then some of those are carried over to, to different libraries in Python, so you can sort of try to find the places in Python and adopt the projects that, there's definitely like some of the, I think like Plot9, and I think it's like Sluba, I think, sort of a dplyr-like library, like, so there are like Python libraries that are sort of very much inspired by their R counterparts. You can adopt those, you can get involved with those.
RStudio's impact on community culture
Alan, I see some plus ones to your question, so I'd love to have you jump in and ask that too. Yeah, sure. Hi, my question is about the impact that RStudio's had on the community over the years. I think it's been really positive in helping the community be welcoming and encouraging and open in a way that I think folks agree, and we make lots of jokes about, was not necessarily the way, like, when I was a grad student, for example. Was that a conscientious orientation? I'm curious about the story of how you came to that kind of contribution in particular, and how conscientious it was.
Yeah, it was, I was very conscientious. I definitely think, you know, a lot of us first encountered the R community in sort of a, like, less friendly mode, and I think there was, that was sort of disappointing, like, wow, it'd be great if more people felt welcome, and more people felt like they could come in and ask questions and succeed. But I would, I would, I would credit primarily Hadley for his orientation around that. He is, for those of you that know him or have heard him talk about this sort of stuff, he is extremely passionate about inclusiveness and friendliness and welcomingness. Like, he is, and he is, I mean, there's a lot of people here who believe in those things, but he really set the tone for that, and for the whole community, and for the company.
R and Python for machine learning
Paul, I see you have your hand raised. Do you want to jump in? Yeah, sure. Thank you. Thank you, guys. So, I guess my question, well, this is kind of a question, but it's also, like, something that being, you know, having discussions with my, sorry, I'm moving around the house. I have my newborn daughter. She was just born this Saturday, so I'm here, like, taking diapers and everything, whatnot. So, yeah. So, I work for the Panama Canal Authority, and the Panama Canal has recently created a data science unit. And so, basically, we're, like, 16 people there, and I'm the only one that is a heavy R user, because we're, like, divided. It's, like, me against the rest.
My question would be, how much, I mean, there's always a debate, and you always find on the internet, like, these papers and discussions, like, whether to use R or Python for machine learning, for AI, whatever. And I've seen, you know, all sorts of opinions about it. There's people that is, like, okay, Python is better when it comes to AI and machine learning, and then R is better for statistical computing, or for computational statistics purposes. Now, I'm an actuary, and, you know, I've been using R since I used to work for insurance companies in maybe 2006.
Yeah, it definitely is. I would say statistical analysis, interactive data exploration, communicate, you know, kind of creating data products that communicate things that are, like R is better than Python at all those things, I think. It is not as strong of, it's not as widely used as a language. So, it's not as general purpose. And I think in terms of, like, large data sets, I think that mostly that is handled by either, like, external systems you interface with or native code. So, like, dplyr or Datatable or DuckDB or Arrow or Spark. There's some means by which you're being, you're able to access large data. It's not really confined by the language. And I think that R's interfaces to those technologies are pretty comparable to Python.
I do think in terms of machine learning, there's a lot of machine learning packages in R. And, you know, some of the, like, most important things like XGBoost and stuff work very well in R. Python probably has a, like, we have created deep learning interfaces in R, both to TensorFlow and Torch. But Python probably does have an edge in deep learning, I would say, in that you can use those libraries and there's probably more, like, third part extensions to those things that are available. So, but, like, deep learning is one part of, one part of machine learning, one part of working with data. And it's not really predominant in some fields.
So, I think, you know, I think, like, I think what we're hoping to make easy for our users is a world where if you want to use R and it is so much easier for so many things, it's very easy for you to make that case. And so, like, things like reticulate for interoperability, things like that, you know, we're trying to make it easy to make that case. And so, I would say, like, for the things you're doing, R is clearly, clearly better.
Building a successful business on open source
Travis, I see you had some guidance there. Do you want to weigh in for Lisa? Yeah, sure. Thank you. Thank you, guys. So, I guess my question, well, this is kind of a question, but it's also, like, something that being, you know, having discussions with my, sorry, I'm moving around the house. So, I work for the Panama Canal Authority, and the Panama Canal has recently created a data science unit.
Travis, I see you asked a question in the chat a little bit ago. Do you want to jump in? Sure. Hey, JJ. A question about sort of open, you know, building a successful business on top of open source. So, a lot of us, probably most of us here, I myself am in pharma. So, no one talks to each other and we all guard our technical process and code as IP behind a closet somewhere. But there's, you know, momentum going towards how do we do more open source and how do we start to talk to each other and still grow and hire people and, you know, do the posit thing. I suspect there's probably a virtuous cycle kind of shout out that you're going to give here, which is awesome. I always love to hear more about that, but just kind of key pieces of advice for talking to CEOs, talking to business leaders, orgs, and kind of convincing at a high level that this is the direction that we should go and still be sustainable as a business.
Yeah. I think that like, I mean, it's different, like being just a technology vendor first is, it's a little different because you guys aren't a technology vendor first. I'm assuming you're talking about, you're like work in a pharma company. And so you're trying to figure out how can we create more ground for collaboration with other pharma companies and the community. I think there you might make the case that these sorts of things like enable everyone to collectively move forward. And then our core competency is like, you know, like we always think of RStudio is like the things that are paid products are things that you buy when you're trying to adopt are at scale.
So I think that's like, there's the line there. It's like sort of at scale. And, and I think, so it's not exactly the same reasoning chain, but it is some kind of a virtuous cycle thing where you, I think you need to make the case that somehow that like, like more collaboration with the community and other companies will, will create situations where your core competency will be more valuable. You know, I know that, that was one of the rationales, like for Google, when they open sourced to TensorFlow, they basically said, well, we're going to be really, really good at, at training TensorFlow models and hosting TensorFlow models. So if we get lots of people using TensorFlow, then that's good for us.
Posit's product roadmap
Michael, I see you, you had asked a question much earlier about the company roadmap. Do you want to jump in? Sure, I can jump in. Hi, JJ. I sort of feel like I'm talking to a celebrity having been such an RStudio fan. So I appreciate you hopping on the last call of the year here. I was just curious, you know, being such a big fan of the R community, what the roadmap looks like for Posit and maybe one, three, five year strategy, quick, high level.
Yep. Well, I think you, you, you know, on the open source side, you can see that we are, we're like, we're, we're trying to make Quarto and Shiny, you know, multi-language things. On the modeling side, we actually are, we have some, we have some Python oriented modeling projects, but we have, and we have tidy models. So that's, that's like, it's, it's, it's like TBD, how much, how deep we're going to go in terms of like Python modeling, but we certainly have a big commitment to doing, to, to, to tidy models.
On the commercial product side, I think that our, you know, our, our products now, I think, do a lot. I think they don't work together quite as well as we'd like them to. So I think kind of putting the T in RStudio or sorry, the T and Posit team, making those products, Connect and Workbench and Desktop and Package Manager all work together really, really well, is important to us. Integrating those things with the other things that people are doing in their enterprise with data science is, is really important to us. Providing a cloud version of those things is really important as well.
And then I think we, we actually see this as a newer thing, but like a significant role for, for RStudio academy in providing learning experiences that are really not, not kind of like, you know, not superficial and transactional, but really deep. And so we're, we've been doing, we're pretty excited about what we've done with RStudio academy so far, and we'll be doing a lot more with that.
Measuring success of open source initiatives
Eric, I see you just put a question into the chat as well. Do you want to jump in next? Yes. Hi, JJ. I apologize for my voice. I'm getting over a nasty foo thing, but while my voice works here, I was curious how you and your various teams at Posit, when you think about launching a big initiative, such as Quarto or Shiny for Python, et cetera, what are you using internally or at least what do you can reveal to really quantify if that's really being successful or not? Because a lot of times when I have ideas at the day job where I think we can transform like the way we operate or the way we analyze the results, it always comes back to a leadership team saying, well, what's the value you're giving us? And that's a hard question for me to answer sometimes.
Well, I look at those things. Like I always look at usage metrics and there's different proxies for usage metrics, like package downloads is one usage metric that you have often. And so I look at what's the trend of usage, right? Is it an accelerating trend? So I don't know in absolute terms if there's like a million downloads of the Shiny package. I don't know if that's a lot or a little actually, because I don't know if most people download it eight times or, you know, but I, so I kind of look at like momentum of projects and I see what, like what the trend line looks like over time. And again, is it accelerating as a way of sort of gauging, are we having the impact that we want? Is it growing organically? Is like word of mouth leading to more adoption?
So that's, I think the best we can really do is just look at like adoption metrics and how they grow over time. And comparative adoption metrics, you go like, what are some analogs, you know, to what we're doing? Okay, well, how do our adoption metrics look like compared to that analog? So.
Getting involved with Posit open source
Okay, I see somebody else had asked on Slido, any advice on getting involved with Posit's open source efforts, specifically in the shiny or machine learning areas? Yep. I mean, I'll tell you, we always welcome people who get involved. And, you know, in many cases, we've ended up hiring people who get really, really involved in projects. And I think, and then the first way to get involved, if you're just sort of just like, I don't, I'm not comfortable, you know, making pull requests, or really like trying to understand the code is just like reporting issues. And, you know, making like PRs for documentation and things like that. So that's like a first step that's kind of accessible to anyone.
But I think that I mean, a really good repo, and I think I think some of the tidyverse repos do this, they will actually tag their GitHub issues with good first issue as a tag. Meaning like, if you're just just got off the bus, and you don't know this project at all, you don't know the code, this could be a good first issue for you to try to try to tackle. And you can even you can even you can even like go to a repo and say, Hey, I don't I see you guys don't have any good first issue tags. Can you make some because I'd like to work on some issues. I can tell you like that the people who maintain those repos will be very excited about someone who says like, what are some issues I can work on. So I think working on issues is a great way to learn the tech, start interacting with the development team. And really, the like, the sky's the limit.
Learning to code and the RStudio IDE's future
And one was, there are lots of options for teaching coding. Have you thought about developing something for the younger generation to begin learning are like K through 12? We haven't. I think that would be awfully, awfully interesting if someone could could figure out how to do that. Well, I mean, really, like what we've kind of gone, I've seen, you know, I have kids, and I've seen them learn coding with various, the various like self paced online things. One complaint I have would be, you you can learn some tricks, and you can learn, you know how to write a loop how to set a variable how to, you know, but you don't actually learn how to do the work. Because when you really sit down and do the work, you're really frustrated, you don't know how things you have to learn how to do research, you have to learn how to read error messages, there's like all this stuff that is not covered in these sort of, like, drilling on these tactical skills.
So like what, like in with RStudio Academy, we've kind of gone the other way. It's like, let's, how can we have like a really thorough experience where people are doing real projects, and they have mentors, and it's like real work. So like, that's the kind of thing I'd want to see in high school for high schoolers, like, you're actually doing real work, you have a mentor, not you're just like, you know, like cargo culting some tutorials. So maybe some of the work that we're doing in Academy can eventually turn into like, hey, if you're if you want to teach a course in high school on data science, here's materials, and here's an approach.
Do you see Quarto helping to fill the gaps between business users and data scientists? Well, yeah, yeah. And that I think we do in the I think there's two ways. So like, the focus that we have on creating data products, and not just data products that are static, but like shiny applications, or scheduled, you know, reports that you can parameterize reports that you can schedule. That is a way to take the expertise, knowledge, and skills of data scientists and project them in a flexible way throughout the rest of the organization.
There was a question earlier that I had missed that was around VS code and Sam, Sam and Alan, thank you for helping bring it to my attention. It was with many useful new extensions, VS code is an alternative that a lot of our programmers are switching to. Where do you see the RStudio IDE in five years? Well, I do think that the VS code is a very, it's a, it's a incredibly powerful, but unopinionated text editing program, you know, programmers tool. So, I think if you're willing to put the time in to learning it and customizing it, learning how to drive it, getting all your extensions dialed, it's really powerful and really great.
It, RStudio has a different philosophy. RStudio has much more of an opinionated, these are the tools you're going to need to do data analysis presented in a unified, clear, hopefully straightforward and simple way. So, it's, it's actually kind of trying to do something different than VS code is. VS code is trying to be like a maximally powerful extensible programmers editor. So, I think it's possible to get, you know, with various extensions and we, you know, we have like our, we have our Quarto extension for VS code. We're definitely doing a lot of work with VS code for people who like, who kind of want that environment of like, I'm controlling everything and I'm composing my IDE out of all these different parts and I'm customizing and configuring. Great. But I actually think that many, many, many data scientists need something that's more straightforward and streamlined and that's kind of where RStudio fits.
Posit as a public benefit corporation
Oh, yeah. Hi. So POSIT is a B Corporation, a benefit corporation. And I'm wondering if you have anything to share, aside from public benefit of bringing this open source tool to all of us that use it for free. Anything else around environmental or, you know, that benefit corp structure? I mean, the biggest thing that the biggest differentiating characteristic of public benefit corporations is that they are, unlike traditional C Corps, which the kind of the evolution of C Corps in the United States and the case law and things have made C Corps like really pretty much, they're pretty much exclusively beholden to their shareholders as the stakeholder that matters. The biggest thing about B Corps is it changes the corporate law and corporate charter so that your stakeholders are considered more broadly, your customers, your employees, your community, the environment, your shareholders, all stakeholders, and your duty is to run the company for the benefit of all stakeholders and to balance their concerns and interests.
So that's the main thing. And then we also, the way the law works is you also establish a public benefit that you're accountable to working towards. In our case, it's open source software. We have, we actually get rated by the B Lab every three years. And if you go look at our report, there's quite a bit in there about sustainability, environmental sustainability, and that's like a bunch of the points come from that. So I would say like, and we're actually about to get re-rated, so there'll be a new report coming out next spring, or hopefully by February or March, that actually goes into more depth about that.
Failures and lessons learned
One last, one last question. Thank you, Jonathan, for the great question here. What are some failures that you have learned from in your career and how have you grown from your experience? Yeah. I'd say one failure was the second kind of company that I started. I, the first company I started went pretty well and I kind of thought that, oh, you know, the, the skill that I was applying there was like, I the skill was like I don't want to say this, like coming up with like brand new ways to work and brand new ideas is like a skill. And then that's what startup founders do. And I kind of like in my second company, like try to create a whole bunch of brand new things, like, oh, reimagine your work this way, do this totally different thing that you're doing now. And that didn't go very well.
And I kind of took from that, like, it's much better to be rooted in what you observe people are doing exactly what they're doing and try to make that better. So as an example, when we started, I started working in R, I wasn't a data scientist, I wasn't an R programmer, and we easily could have said, okay, we're going to reimagine how you work with R. We're going to build this brand new interface, forget about the console, forget about, but I said, well, there's lots of people using R just this way. And so if I can just make that better, then that's like a good foundation. So I think like overreaching in product design is something that I definitely paid for in the second company that I started. And I've tried to not do that, try to stay close to what people are doing, I'd say.
And I kind of took from that, like, it's much better to be rooted in what you observe people are doing exactly what they're doing and try to make that better.
I think I alluded to this before, but like, not knowing yourself well, and finding yourself in sort of with responsibilities and in a work situation, where you're not like being nourished the way you need to, you can end up in that spot, and it's very hard to dig yourself out of it. So yeah, that's happened.
Building a company for 100 years
Thank you. I know I said that was the last question, but I don't want to end on like failure. So I know something we talk about at Posit is creating a company that's around in 100 years. What do you think is one of the most important lessons learned or things that you're thinking about to get us to 100 years from now? Well, I think the number one thing was get the right corporate structure. So all the incentives are aligned. So that's the only thing we need to worry about. And we're not worried about maximizing some short term objective. So that's one.
Now, two is going to be leadership succession. So it's easy to get caught up in like, oh, hey, you know, there's four or five of us that are kind of leading things, and we're making it all work. And then like, yeah, but we're all going to get old, you know. And so trying to figure out like over time, who are the people that are going to succeed you and giving them opportunities to like even accelerate their learning and bring them into the space that you're in is I think where a lot of organizations fail to be sustainable. They may be like economically sustainable, they may have the right corporate incentives, but then like succession kind of does them in. So that's something where it's not like right on our doorstep, but it's something that's on my mind to think about.
Thank you. Thank you so much, JJ, for joining us today and sharing your experience. This has been really fun. And thank you all for being a part of the Data Science Hangout today. If this was your first one or for the past year and a half or however long we've been doing this, thank you for coming to spend time with us all. I hope you and your families have a wonderful holiday season and Happy New Year. This is our last Hangout for 2022, but we will be back here after a bit of a holiday break. So we'll be back on Thursday, January 12th for the first Hangout of 2023. Thank you all so much for making this the best part of the week.
