How the New York Times is Harnessing Tech in the News Industry with Jason Sobel, CTO

(Intro Narrator at 00:00:00) Today, we're talking to Jason Sobel, CTO at the New York Times, about how they're harnessing the power of modern tech to bring you the news. Thank you to DigitalOcean for sponsoring this episode. For simple cloud and powerful AI that's built to scale, visit digitalocean.com or just click the link in the show notes. You're listening to Joel Beasley, Modern CTO.

(Joel Beasley at 00:00:28) When I saw the opportunity come up to interview CTO of New York Times, I was like, absolutely. I would love to meet this guy, Jason. Can you tell me a little bit about what it's like being the CTO at the New York Times?

(Jason Sobel at 00:00:40) Yeah, it's been incredibly fun and challenging, a learning experience. I think the organization really lives and breathes on its reporting first and foremost. The news is the most important thing we do. But these days, technology is how we deliver it.

(Jason Sobel at 00:00:58) We still have a subscription home delivery business that's doing well for us. But our digital business is our biggest business. It's growing. We know that eventually, one day, the print business will go away. We don't know when that is. We hope not for a long time. But how we deliver the news, sports, games, product recommendations, cooking, all the great stuff we do is through technology. And so I get to work with my teams to think about how we do that in the best way possible. So we think a lot about just how do we build great digital experiences to deliver all this to our customers. And then what are all the really important technical underpinnings that we have to get right so we can continue to thrive and compete in just a really tough market.

(Joel Beasley at 00:01:47) That would make sense to me that the physical would decline and the digital would do well. Yeah. I think it's a generational thing too. Right? So you've probably got twenty, thirty years of still offering the, maybe even longer than that because we still have horse farms today. Now, even though we've got cars, we still have, so there'll still be a desire for it at some point, you know?

(Jason Sobel at 00:02:11) Yeah. We hope the print business lasts a long time. I think when I was interviewing, I met our publisher, A.G. Sulzberger, and he said something that really resonated with me that a given story, more often than not, is better in a digital experience. That you can do just all these new and creative things in a website or in an app that you can't do on the printed page. But that there's nothing like the packaging of the printed paper, that it is still the best way to package up all the news we do.

(Jason Sobel at 00:02:46) And that idea of, you know, when it was all we had, we would just sort of sit at home and leaf through the paper and the sort of the serendipity that it created. The physical space it had was really valuable. And as things have sort of gotten more and more compressed into mobile, that the overall packaging is not quite as satisfying as that physical experience. And I think that really resonates. There's so much we can do in a story with technology, and our newsroom does what we call interactives, just these incredibly rich stories that really take advantage of technology in a compelling way. But there is something when you're only seeing one or two headlines on your phone that is very different than having that physical paper.

(Joel Beasley at 00:03:29) Yeah. I like the physical stuff and the digital stuff. So what I do is I Audible almost everything that I want to read.

(Jason Sobel at 00:03:35) Yeah. Yeah.

(Joel Beasley at 00:03:36) And then I don't know if you can see it. But then I buy, it's down there, I buy the physical books that are the best books. Yeah. And I don't even ever usually read them. I usually just buy them and put them on my bookshelf because when I see the spine, I remember the principles in the book. And to me, that's the cue I need.

(Jason Sobel at 00:03:55) Yeah. It is nice kind of having both. And it's funny you talk about audio. That's been a huge focus for us recently is really just trying to make all of our great content available on as many different mediums as possible. And with the way technology has been advancing, we're able to do a lot more what we call automated voice. So we still have a lot of human read stories that we produce, but we've been able to use tech to make much more of our report available now through audio means as well. So for folks who want to listen to the New York Times, you can do that now too.

(Joel Beasley at 00:04:32) I love that, Jason. We've been tracking actually here at the podcast. We started about two years ago tracking automated podcast because it's, you know, in our space. We're technology people. Boy, has it come really far in two years. It went from janky, like, I don't know. Like, this is not that. Yeah. It's a cute little idea to, oh, wow. You can't tell the difference. It's gotten really good.

(Jason Sobel at 00:04:57) It's remarkable. It is just remarkable how much the sort of this new wave of Gen AI driven tech has, I mean, multimodal is so powerful now. And, you know, we're starting to see it with video as well. But, yeah, for us, it's been a real boon in our ability to just make the report a lot more accessible to people by, you know, taking stories and sort of packaging them up in different ways that are going to get them to the biggest possible audience.

(Joel Beasley at 00:05:23) So cool. You guys have a bunch of podcasts over there at New York Times?

(Jason Sobel at 00:05:27) We do. Yeah. We have some incredible podcasts that are going really well for us. Hard Fork for the tech team is a good one, as is The Daily. We have, you know, it's interesting too how just all these mediums are blending. Right? Like, what we used to think of as a podcast, now a lot of people consume them more on, you know, YouTube or on other video platforms than they even might in just like a podcast app or something like that. So we see a ton of people watching. Ross Douthat has a great show that's going really well for us called Opinions. So, yeah, there's just been a real successful expansion of those formats for us.

(Joel Beasley at 00:06:13) Have you been on Hard Fork?

(Jason Sobel at 00:06:15) I have not.

(Joel Beasley at 00:06:16) What? The CTO hasn't been on Hard Fork?

(Jason Sobel at 00:06:19) I don't know if there would be a conflict there or not. That's an interesting question. I have taken every opportunity I can to pitch myself as a new podcast host for the New York Times. So maybe someday. I was considering a theater major along, I ended up majoring in computer science, but I always loved performing as a kid. And so my end game is still to worm my way onto stage or TV somehow if I can.

(Joel Beasley at 00:06:44) Oh, that's awesome. I love it. I'm a split person too. I go out and I tour as a comedian, and then I did engineering, full time software engineering for seventeen years. And then the podcast got popular. That became my full time thing, and I stopped doing the hands on software every day. And I had this extra time. I was like, what do I do? And I was like, let's go out and make people laugh.

(Jason Sobel at 00:07:09) I love it. You're going to have to do some of your bits for me.

(Joel Beasley at 00:07:12) Oh, and that's exactly what I want to do right now. No audience. And let's just do them and analyze it.

(Jason Sobel at 00:07:19) One on one stand up is a tough sell, I imagine.

(Joel Beasley at 00:07:22) That is, yeah. There's very few things that you need for stand up to work, but the few things is an audience...

(Jason Sobel at 00:07:27) An audience.

(Joel Beasley at 00:07:27) And a dark room and a microphone.

(Jason Sobel at 00:07:30) Yeah. Yeah. So all right. Well, that's, I'll have to put that on my list too. Give stand up a try sometime. Where are you located at? I live in the Bay Area, actually. So I have been out here since 2004. I grew up on the East Coast in Connecticut. Came out here after college, you know, where all the tech jobs were. So it was the early two thousands. And then I got the reach out from the New York Times during COVID, and everyone was more remote. And so I've been working remote. I go back to New York about once a month. But around half of our engineers also work outside the New York area. So for me, being remote has been a little more manageable.

(Joel Beasley at 00:08:14) And so what were you doing in the Bay Area before the New York Times stuff came up?

(Jason Sobel at 00:08:20) Yeah. So I came out here. Actually, I focused a lot on computer graphics in college. And so I came out here. My first job in the Bay Area was working for Industrial Light and Magic, which is the Lucasfilm special effects company up in Marin County. And I was only there for a few months before I decided to make a change and work with a bunch of friends over at a company called NetApp. And that was sort of my first real professional software job. So I did that for a few years working on, you know, file systems in these, like, you know, this was pre SaaS, and it was these huge honking servers pre cloud. You know, we put them behind companies' firewalls, and so it was a very interesting experience. Like, real old school enterprise, like, waterfall release cycles. The cost of a bug was really high. I remember I shipped a bug once that we had to patch and going through that experience was a little traumatic. Made me not want to ship bugs ever again. And then in 2007, I started thinking about trying to do something a little different. And I had a friend who'd gone to Google and some friends over at Facebook and interviewed with Facebook and was just blown away by the people there and the ambition. I hadn't used the product a ton, so I had kind of missed it in college by just a couple years. But I knew a lot of people a little younger than me who were just obsessed with the product, and I could see why when I went inside the walls and got to meet a lot of people. And, yeah, I was lucky enough to get an offer there and joined. We had about 20,000,000 monthly active users when I joined, which seemed massive to me coming from an enterprise company. And it was just, like, right place, right time, and got to really grow in my career there. That's where I ultimately switched over into engineering management. Did that for a while, and then had another startup job here. So I tried something much smaller after about six years at Facebook. Was the VP of engineering at a startup called Quantifind for a couple years. And then, right before the Times, I was at Airbnb for about five years working in a couple different roles there.

(Joel Beasley at 00:10:31) Oh, that's pretty cool. So what created this transition from Airbnb? So you went from Airbnb to New York Times. Were you just, did you feel that you had contributed all that you could contribute to that project at that time? Or how did that go?

(Jason Sobel at 00:10:46) Yeah. So, you know, I'd been there for about five years. And most of that time was working in the trust and safety space at Airbnb, which is not something I'd done before and was an incredible learning experience because they're really hard, important problems for Airbnb, just keeping the community safe. We got to learn a lot of new technology, but also just learn a lot that really wasn't about tech. I worked with a number of people who were what we called our operations team. So these were the people sort of on the front lines of keeping the community safe, interacting with customers if they needed any help or had any issues, and really got a lot out of that experience. But then we had a new CTO join. And when he came in, I switched over to work with him on our core infrastructure. So that was more what I'd sort of grown up doing at NetApp and at Facebook. And so I spent some time doing that and kind of helping get the team to a good place. And so it was the end of twenty twenty, and we were all working remotely. We were all still kind of on distance because of COVID. And, yeah, I just felt like I'd kind of done a lot of what I wanted to do there, and I wasn't really sure where to go next. So I actually took some time off and had a few months to just meet a lot of people and explore. And when the New York Times reached out, what just really grabbed me was just how connected I was to the mission. You know? I think working at Airbnb was great, and I love the product we built. But journalism has really suffered in the digital era. And as someone who really believes in the value of journalism, but also worked at Facebook and was part of the shift of advertising dollars from traditional media to digital media, I felt really connected to the idea of helping journalism thrive in this new era and being able to use technology to do it. So that really was what pulled me in.

(Joel Beasley at 00:12:45) I'm curious. Like, I don't know much about journalism. But what I would think, you mentioned you were connected to the mission and journalism was suffering. When I look at from just my software engineering standpoint, like, isn't it thriving? Like, don't we have the most amount of people able to do this? Didn't it enable? Or can you explain to me how journalism is suffering?

(Jason Sobel at 00:13:08) I'm definitely not an expert, so I'll probably get some of these details wrong. But I think it's about 70% of journalism jobs have disappeared over the last...

(Joel Beasley at 00:13:08) Oh, wow.

(Jason Sobel at 00:13:08) Twenty, twenty five years. So basically, what happened is, you know, if you were a big advertiser, you know, TV and print media were your channels. That was how you could reach the most people. And so a lot of advertising dollars were going to those places. And, you know, the New York Times was one of the biggest papers and benefiting really well from those ad dollars. As Facebook and Google came online, I mean, Google first and then Facebook and then more and more digital properties, they just hovered over a huge portion of those ad dollars. And so it was sort of this just negative feedback loop where the money would move and also people's attention would move. So there were less people consuming more traditional media, whether it was TV or printed media. And so it was just this feedback cycle of more and more budget moving over to digital mechanisms. And so most people in America don't have a local paper anymore. Like, when I grew up in Connecticut, we would get three papers. So my dad was a Wall Street Journal reader, but then we'd also get the Hartford Courant for Connecticut. And then we had a local paper, I think the Journal Inquirer, I think it was called, that was just for our local area. And these days, basically, there's no, you know, there'll be maybe small local papers in a few areas, but that's very rare. And in many states, the even the state paper has either gone out of business or is really a shell of itself and is only able to have a very small newsroom. And so there's been a real challenge in the journalism industry as a result. And there's actually been a bunch of studies that show local and statewide corruption has increased alongside the decrease in sort of local and statewide journalism because you don't have those people poking around and sort of asking the hard questions that they were asking when I was younger.

(Joel Beasley at 00:15:18) All right. You've convinced me. Okay. But I mean, no, you're right. What happened was the stuff came out and that was like sugar, you know. It was like candy. And then everybody piled onto that. And it lacks nutrition, but it tastes really good. And then if all the eyeballs are going over there, then traditional media, I'm not going to say you guys specific. I'm just going to group everyone. It's like traditional media. It's like half the people in traditional media are like, let's play catch up and let's do more sugar. And that's where the money is. And then it just becomes such a small sliver of where the money is for truth and real in-depth journalism that it kind of almost goes into this little niche category whereas it was once the main category.

(Jason Sobel at 00:15:58) Yeah, I do think in general news consumption is just at a different scale than it was previously. And people will say they get their news from different places. So a lot of people say they get their news from Google or from TikTok or other places like that. I think there's a different level of rigor that's put into a TikTok creator's news than our newsroom.

(Jason Sobel at 00:16:21) Our newsroom is pretty incredible. I know how hard they work, the standards they live by. We preach this value of independence with the idea that we just follow the truth where it leads. That's what we're trying to do. We're not trying to tell a story.

(Jason Sobel at 00:16:35) Our opinion pages are different, but our news is about, let's follow the facts, and let's tell the story that we find when we follow those facts. And I think most traditional media or news outlets would take that same approach. That's harder and harder to find, especially in an era where there's a lot of activism, and activism is a worthy cause in many cases. But activists want you to say things in their voice and through their lens and with what outcome they're trying to drive. We view our job as something fundamentally different, which is just telling the truth.

(Joel Beasley at 00:17:14) Well, I think you guys will do well because you have a brand, you know. It's definitely an old brand that people have come to trust to get their news from. And it's interesting for me as someone who doesn't have a dog in the fight to kind of watch how this is going where independent creators pop up. But then, sometimes that's always been the case where you will read the New York Times because you like a specific writer who has a specific column. Right? That's always been the case, but they originally were only available through

(Jason Sobel at 00:17:46) the New

(Joel Beasley at 00:17:46) York Times. And now some of these people are just going out independent. That's right. And so it's like, how do you—you still need that because the breeding ground for developing these individuals and their craft is at places like the New York Times. Right?

(Joel Beasley at 00:18:02) That's where they can get really good. And it's really, really, really hard as an independent creator to start with nothing.

(Jason Sobel at 00:18:09) That's

(Joel Beasley at 00:18:09) right. Most of the successful independent creators that we would apply that label to, they came out of a major news organization. That's where they got their first 10 or 100,000 followers.

(Jason Sobel at 00:18:19) Yeah. And, you know, I think that's right. And I don't want to paint digital media or the digital ecosystem as all bad. Right? I think it had this bad effect on journalism. There are a lot of great things that came with it, you know, and there's a lot of debate to be had about echo chambers and what's working and what isn't working in media. But it has given voice to a lot more people. You know? The old model where it was the newscasters on TV and in journalism, and those were the only people who could talk at scale. You know, that had challenges too.

(Jason Sobel at 00:18:51) So I think we're really just trying to figure out how do we make this transition and hopefully just have as much news and thoughtful commentary as possible. I think we don't have as much independent journalism as I think we should have, and we're excited that the Times can continue that. And hopefully, more people can learn from what we've done because that is just a necessary thing for society. Like, we really believe that you need an independent media asking questions and digging into things and telling stories that need to be heard.

(Joel Beasley at 00:19:26) No. I'm in alignment with you on that. I like—by the way, talking to you and watching your career trajectory, you really are, I think, in the right spot for you right now. Like, you really do—it comes out in everything you say. You really do care about what they're doing there at New York Times.

(Jason Sobel at 00:19:41) I do. I do. I think that's so important. You know, not everyone has that privilege, but to be able to align your career with things you really believe in and are excited about, it's pretty special when you can get that dialed in.

(Joel Beasley at 00:19:55) Real quick, before we continue, I wanted to give a special shout out to DigitalOcean for sponsoring this episode. DigitalOcean is cloud infrastructure that's simple to spin up, complete with integrated AI dev tools plus 99.99% uptime SLAs and industry leading pricing on bandwidth. We all know DigitalOcean. It's super reliable. It's been around forever. And now with DigitalOcean's Gradient platform, developers can train, fine tune, deploy, and scale AI workloads all in one place. DigitalOcean is reliable and affordable at any budget. Companies can save up to 30% on their cloud bill when they come to DigitalOcean. Use code DO25 and get $200 in free credits to get started. Simple cloud, powerful AI, built to scale. That's DigitalOcean.

(Intro Narrator at 00:20:44) Now back to the episode.

(Joel Beasley at 00:20:46) Let's talk about the first time that you went into management. Did you just get an offer to go into management? Did you turn something down? How did that go?

(Jason Sobel at 00:20:53) Yeah. It's funny. I was actually—I was at Facebook. I'd been there less than a year, and the team was just growing super quickly. So when I joined, I think we had about 60 or 70 engineers, and we were tripling or, you know, doubling every year for the first two or three years. It was just growing so fast as the product was taking off. And my manager at the time, this guy Bobby, I think he'd gone from like 10 or 15 reports to 30. And he was still managing everything that we called infrastructure, which was kind of all the back end tech keeping things up. So, managing databases, servers, caching. Right?

(Jason Sobel at 00:21:35) We were this big LAMP stack, and this was pre-cloud and sort of the open source community hadn't really taken off. I mean, obviously, we're using Linux and Apache, Memcache, MySQL, and PHP. Those were like our core things. So anyway, he had, yeah, like 30 reports, and he was feeling he needed another manager. He needed someone to help manage more people.

(Jason Sobel at 00:22:01) And I had just taken on a big project to open our second big data center. So we had all of our kind of servers on the West Coast between San Francisco and Santa Clara and just one copy of the databases. So we ran a sharded MySQL back end, and we were running out of capacity. And our plan to scale the capacity was to open up a data center in Virginia. And instead of—we weren't going to just have web and cache there. We were also going to have a second copy of all the databases. And so we had to figure out how to make that work, and that was the project I got when I joined. So I joined in April, I think. And I was told that if we hadn't figured this out by the end of the year, we would run out of capacity and Facebook would go down. So had a tight timeline to figure it out.

(Jason Sobel at 00:22:51) So I remember I was at Thanksgiving launching the data center at the Thanksgiving table, and it mostly worked. So anyway, I'd just come off that project and it spanned the whole company. Had to pull a lot of people together, work with a lot of teams in our ops group, and software engineers, and everything. And so, you know, I guess he thought because I could handle that, that I might be good as a manager. And he took me out to lunch. We got sushi, and he asked me if I was interested. And I actually said, no. I was like, I don't know what managers do. I don't want that job. So I turned him down. He had to take a couple runs at me to get me into it.

(Joel Beasley at 00:23:28) What ultimately caused you to change your mind?

(Jason Sobel at 00:23:32) You know, I figured it wasn't a one-way door.

(Joel Beasley at 00:23:35) You can say money too, by the way.

(Jason Sobel at 00:23:37) It wasn't money. No. There was no new money.

(Joel Beasley at 00:23:39) No?

(Jason Sobel at 00:23:40) That's actually one thing that I've liked about the places I've been is we haven't viewed management as a promotion. We've viewed that as a different job that has the opportunity for sort of equal impact just in different ways. So there's only one CTO at the Times. There's not an equivalent IC engineer level. But for every other level at the Times, at Facebook, at Airbnb, there was an equivalent engineer. So you could be a software engineer and have just as much opportunity for compensation and impact as a manager, which I think is actually a really good thing to do.

(Jason Sobel at 00:24:19) Yeah. He thought I would be good at it. He talked me through why he thought it was a really valuable job in just helping people do their best work. And like I said, I didn't think it would be a one-way door. So if I hated it, I could always bail out and go back to writing code. But I ended up loving it. And I think, honestly—and I was a good engineer. I've worked with a ton of engineers who are better than me. I think as a manager, I've probably gotten further than I would have as a coder.

(Joel Beasley at 00:24:50) I would agree with myself too. Like, sorry. For me, I'm thinking right now.

(Jason Sobel at 00:24:55) Yeah. Like, yeah.

(Joel Beasley at 00:24:55) I would have been a crappy—I think you would—yeah. Yeah. I looked at your GitHub before the episode, and I agree with that fully.

(Jason Sobel at 00:25:01) Yeah.

(Joel Beasley at 00:25:02) No. It's like, I felt like I completely grasped the concept. I built and deployed full applications that were—but at the same time, the people that I've gotten to work with once the project started having money and growing, I'm just like, oh, there's just like a whole other breed. You know, there's a whole other level that I'm not even—and maybe my communication gift is something that they don't have in that way. And so I can maybe make some more impact by building and growing teams. And that's ultimately—I found that the reward I get from a successful deployment, if that's the only thing you know, it's 100%. It's max greatness. But for me, if I grow a team and watch pivotal moments in their life change, whether it's having kids or personal growth or whatever it is, that is 100x a deployment for me. So I'm like, look, that is just so much more rewarding on my time invested. And I didn't control that. I just kind of experienced both, and I had to choose. You know?

(Jason Sobel at 00:26:03) Yeah. That really resonates. I think that satisfaction I get from feeling like I've helped a team or a person do something that they might not have done otherwise. It's always hard to take credit as a manager. Like, how much was it you when at the end of the day, it's other people driving the majority of the outcome? So that's always tough, but I think there's a number of things I can look at and say, I can really see how I helped get this to a place that it maybe wouldn't have gotten to without me. That's always been hugely validating. And then there's a few people who I stay in touch with from previous jobs who just found working with me useful and my opinion valuable. And so the fact that they still want to talk to me even after they don't report to me anymore, and I don't control their comp or their career or anything. That's always been really validating as well.

(Joel Beasley at 00:27:00) Somebody—I can't remember where I picked this up, a book or a guest or something, but they explained it like you're a gardener. Like, you can't force the plant to grow, but you can control certain things in the garden and get the weeds out and make sure that there's nutrients in the soil. Like, you can do certain things, but you can't force it.

(Jason Sobel at 00:27:18) That's right.

(Joel Beasley at 00:27:19) And so that's how I started looking

(Jason Sobel at 00:27:20) at it a few years ago.

(Joel Beasley at 00:27:21) I was like, alright. I just need a good garden. I can't control it, but I can control these certain inputs, and so I'll focus on those.

(Jason Sobel at 00:27:28) Yeah. I love that metaphor. I might borrow it.

(Joel Beasley at 00:27:30) Right?

(Jason Sobel at 00:27:31) Yeah.

(Joel Beasley at 00:27:31) Take it. I took it from somebody.

(Jason Sobel at 00:27:33) That's a good one. It's a good one.

(Joel Beasley at 00:27:34) I know. I know. And, okay, let's talk about—we've got to talk a little bit about Gen AI. I know it's like

(Jason Sobel at 00:27:42) Let's do it.

(Joel Beasley at 00:27:42) in every episode. But, I mean, are you guys—do you have a model running wild just making the news up over there? Like, what are you guys doing with generative AI?

(Jason Sobel at 00:27:51) Yeah. That is definitely not on the menu. So I think that's, you know, right out of the gate, that's probably the most important thing to say is all the news is written by humans, edited by humans. You know, a ton of editorial oversight goes into anything that's going to make it into The New York Times. What we are thinking about is a bunch of different things. So one, how can we use this tech to make the production of the news easier and more efficient? So, you know, there's definitely things we can build to help out our newsroom. So one example is we built some tools to go through a bunch of video of the president and try to document—so a very concrete example was, one of our engineers went through hours and hours of video of Donald Trump to try to identify changes in his patterns that might map to cognitive decline. And that's tech that would have been, you know, much harder to build two, three years ago.

(Jason Sobel at 00:28:53) And now we are able to put together—to assist our journalists to help identify segments for them to look at. So it wasn't giving them the answer. It was just saying, hey, we're going to go through hundreds of thousands of hours of video and identify segments you should watch over the years to kind of try to map out how things might be changing with the president. We also have built it into some of our products. So if you're a New York Times Cooking user, which I desperately want to be but still never make the time to cook, we now use Gen AI to help with converting from like imperial to metric units or to help scale up or down a recipe. A lot of people think that's just a math formula, but it's actually not. There is a fair amount of sort of subjectivity and creativity that goes into unit conversion or scaling up or down a recipe. Gen AI is able to help with that. I talked about how we're using it to do more what we call automated voice so we can get the news in different formats.

(Jason Sobel at 00:29:53) And then probably the thing that I'm thinking the most about is just how our software engineers are going to use this tech to be better at their jobs. That's just something every engineer, engineering leader is thinking about these days.

(Joel Beasley at 00:30:05) Have you come up with any good—do you just mandate you guys have to use Copilot? Like, how do you actually roll that out with your teams?

(Jason Sobel at 00:30:12) Yeah. We've stayed away from mandates. You know, my view is engineers are going to use tools that make them better. Right? There's—I don't think there's anyone sitting around today that's, you know, resisting a compiler. Although, I'm sure when they first came out—

(Joel Beasley at 00:30:24) You'd be surprised. There's one.

(Jason Sobel at 00:30:28) You know, I think you have to kind of push people and encourage them, and maybe we will get to some point where it has to be a mandate. But, you know, I think in general, people are going to use tools that make them better at their jobs.

(Joel Beasley at 00:30:40) And it's still—but are they? Like, do you track adoption there? How do you do that?

(Jason Sobel at 00:30:45) We are tracking adoption, and we found we're not at 100% weekly use, but we're getting close. So we're finding that—so the main tools we're using, GitHub Copilot, Cursor, and then we're just starting with a lot more Cloudcode. And Cloudcode is something I'm really excited about for the future. So we are seeing weekly usage of those tools.

(Jason Sobel at 00:31:09) We want to start trying to have more fine-grained metrics to try to understand, are these things paying off? I think that's really hard, but that's something we're thinking a lot about now. This year's approach was basically throw these tools out there and see what engineers do with them, and where does it work, where doesn't it work, and we were willing to accept some creative waste. Right? So when you're trying to figure out how to use something new, you're not going to get it perfectly right out of the gate.

(Jason Sobel at 00:31:37) And so we wanted to give our teams the time and space to figure it out for themselves. What problems was it really helpful with? What problems was it less helpful for? Now that we're at the end of the year, we've tried to synthesize a lot of the learnings, and we're hoping to put together some real conviction about how we do this next year. That's the conversation I'm in the middle of right now with my team.

(Jason Sobel at 00:32:01) Probably what we're going to lean into is this idea of how do we have more and more engineers thinking about themselves as orchestrating the work of AI agents. How do they use these things to do the job in a different way? We've seen some really impressive results with people who've done that. That's not going to be appropriate for every coding job that somebody has, but we think a bunch of them could benefit from that approach. And so that's what we're currently thinking our way through.

(Joel Beasley at 00:32:32) Well, I get to talk to a lot of different people about how they're doing this, and a lot of them are taking that approach. They're just giving it to them and saying, here's the guardrails under which you can use it, especially if you're a health care type company. Right?

(Jason Sobel at 00:32:46) Yeah.

(Joel Beasley at 00:32:47) And so they'll just say here's the availability, and then they watch kind of what people are doing with them, and then they start highlighting people who are having success. That's right. So they'll have the lunch and learns. If this individual is having success, how are they doing it? That way, if just given the tool, if I just give you a hammer, that's one thing, Jason.

(Joel Beasley at 00:33:06) But if I give you the hammer, you play with it for a few minutes, and then I show you how to expertly use the hammer, you're like, oh, okay. That's how it works.

(Jason Sobel at 00:33:15) Yeah. I think that's just the mindset we've taken as we've brought in people from some other companies who've been a little bit further ahead. We've highlighted people internally who found really good uses of this. It's just been really interesting for me talking to a lot of peers. This is something I've invested a lot of time in this year is just picking people's brains on how they're doing it. Everyone loves to have these conversations right now.

(Jason Sobel at 00:33:40) And we found that generally smaller companies have been able to get further ahead. So they bring this in from day one. They don't have a big legacy code base to deal with. And they're just taking fifteen and twenty person teams and giving them these tools and just really empowering them. And they all feel like they're moving significantly faster than they would have if they started the company five, ten years ago.

(Jason Sobel at 00:34:04) Bigger teams like ours, where you have a much bigger legacy code base, I think it has been more this give everyone the hammer, see what they do with it, figure out what's going to make sense for your organization. I'm trying to figure out how to meet those two things, because I do think these smaller organizations have some learnings that apply to us. It's different when you need to take potentially hundreds of people and change their workflow versus just your first ten or fifteen engineers. Excuse me. So that's a lot of what we're trying to navigate now.

(Joel Beasley at 00:34:36) Yeah. I did a couple interviews with a few, there was like two different startups. And maybe, Josh, if you're there, you can remind me which one. But there was this one startup that came out of a big company, but they made essentially like a copilot or like an assistive coding tool that was specifically tuned for context for large code bases. Interesting.

(Joel Beasley at 00:34:57) And yeah, because they were having this problem within this other company of like, we've got this massive code base, but like the Gen AI tools aren't taking in enough context to actually help us with them efficiently. And they built this entire infrastructure and system out to just handle that. And then they went and started selling it to the big companies. Oh, wow.

(Joel Beasley at 00:35:15) Interesting. So if Josh finds that, maybe I'll email it to you after or something like that. But when I met them, I was like, that's fascinating.

(Jason Sobel at 00:35:19) Yeah. That'd be great.

(Jason Sobel at 00:35:25) Yeah. That's cool. Yeah. One of the interesting things we're trying right now is we have this old subscription system that's been in place, I think, since the nineties was when it was written.

(Joel Beasley at 00:35:38) Oh, wow.

(Jason Sobel at 00:35:38) And so it's been picked up and moved into the cloud and all kinds of madness. But it's kind of on an old legacy database. And the code, the database itself, it's all kind of a rat's nest. It's really complicated. It's really scary to mess with it.

(Jason Sobel at 00:35:56) And so we're actually working on a Gen AI project to try to sort of lift and shift that onto a more modern Postgres sort of stack. So that's in progress.

(Joel Beasley at 00:36:06) I love the elephant.

(Jason Sobel at 00:36:08) Yeah. Yeah. So we're hoping we can pull that off over the next few months, which would be pretty cool.

(Joel Beasley at 00:36:15) I remember when I was picking out databases, the Postgres, I chose it because of its logo.

(Jason Sobel at 00:36:21) That's always a good reason. Yeah. That's why we spend so much time on code names and logos, and it's all marketing. It's important.

(Joel Beasley at 00:36:29) It's funny. It's embarrassing. It's like picking your sports team from their uniform. It's—

(Jason Sobel at 00:36:34) I know a lot of people—

(Joel Beasley at 00:36:35) do it.

(Jason Sobel at 00:36:35) Do it.

(Joel Beasley at 00:36:36) Yeah. People do it.

(Jason Sobel at 00:36:37) Yeah.

(Joel Beasley at 00:36:37) People do it. Okay. So let's talk a little bit about leadership insight and advice that you've gotten over your career. I'll give you some constraints, and then we'll see what we can pull out. So I want a piece of advice, but here's the constraints for it. Somebody told you this. You implemented it. It worked, and you've kept it for a very long time.

(Joel Beasley at 00:37:03) It's something that you swear by now. What piece of advice is that?

(Jason Sobel at 00:37:09) I don't remember exactly how this was delivered to me, but it's this idea that clarity is kindness. You know? And I think it really resonated with me as an engineer because my native wiring is to just kind of say things as they are. And I think as you get into management or leadership, you can start to kind of doubt that approach. You can, well, maybe if I hedge or I try to say it this way, maybe they'll think this other thing, or I don't want to demotivate anyone.

(Jason Sobel at 00:37:43) I don't want to, there's just so many excuses or reasons you could come up with to avoid having just a hard, honest conversation. And sometimes you should. Right? You know, sometimes you should hedge, but I think it's relatively rare. The thing that I found in my career that I've really stuck with is just say the hard thing directly.

(Jason Sobel at 00:38:06) Don't try to hide it. Don't try to cushion it. You don't need to be a jerk. Right? You can be empathetic and caring, but still be really honest with someone. And then at the end of the day, clarity and honesty are going to win out over trying to get clever or cute and say things in a certain way to try to avoid a bad outcome or soften the blow or something like that. That's really stuck with me.

(Joel Beasley at 00:38:33) I fully agree with that too, by the way. And I don't think it necessarily becomes easier, but I do believe with experience, you can reduce your time between identifying when the hard conversation needs to happen and having the hard conversation. Does that check out?

(Jason Sobel at 00:38:48) Absolutely. One of the things I say all the time is by the time you're thinking about whether or not you should have a hard conversation with someone, you're probably a month or two too late. Because we just don't want to. It's human. I don't want to have that hard conversation.

(Jason Sobel at 00:39:07) I don't want to say that tough thing. It's going to suck. And so by the time you're thinking about it, you've probably known that you should do it. And at that point, it's just a matter of like, how do I get there? And what outcome do I want? And how do I do this as directly and respectfully as possible?

(Joel Beasley at 00:39:26) Absolutely. And it's almost always like the perspective that I took, whether it's right or not. I don't know. But I have found that it's nice if you don't say it or if you try to sugarcoat it, but it's kind if you do say it directly. And also that it's useful to them.

(Joel Beasley at 00:39:46) So if I have somebody who's underperforming and we're having that conversation, we have multiple conversations and then it's time for them to go. That gives them an indicator of like how much effort they have to put in. Like if they lose the job, if the point comes through it where they are actually like exit the organization, that gives them a new piece of data. It is saying that I performed this way and this was where the limit was. And that's important for them because when they go into, because I used to suck, Jason. That's what I'm saying.

(Joel Beasley at 00:40:17) That's what I'm trying to say. I used to be not great, and then I wasn't getting good results. And I was like, this is a horrible existence. And I found out that I just have to put in a lot more effort. I just have to try a lot harder.

(Joel Beasley at 00:40:30) And then I did. And I was like, oh, I don't suck. I just wasn't living up to what the standard in the workspace was, you know, to be this level of engineer, to be this level of business person. Yeah.

(Jason Sobel at 00:40:42) Yeah. Yeah. My version of that was at NetApp. You know, I was fresh out of college and thought I was the greatest thing in the world, unlike now when I know how garbage I am. And I was just not fun to work with.

(Jason Sobel at 00:40:58) Like, I was very critical and pleased with myself and, you know, thought I had it all. And my manager at the time, he was so nice. He gave me this feedback, and I just blew up on him. I was like, what are you talking about? They're all, it's not my fault. They're idiots. Like, this and that and that. And I just, it's so funny to look back on it and, you know, think about it with the perspective I have now and what must it have been like for him. But it was absolutely the right feedback, and it's, you know, something I had to really work on my whole career is I can be prickly with people when I'm frustrated. And, you know, it started back then, and I still have some elements of that today.

(Jason Sobel at 00:41:43) And, you know, you can't work on it if you don't hear about it.

(Joel Beasley at 00:41:48) Were your parents military at all?

(Jason Sobel at 00:41:50) No. No. No. I don't know where I got it from.

(Joel Beasley at 00:41:56) Just genetically, you're, yeah. Like, my dad was military and an engineer. And so when I look back, I'm like, oh, that's why I'm so direct. There's no, we have to be direct. And then I found out that there's also different cultures on the planet.

(Jason Sobel at 00:42:11) For sure.

(Joel Beasley at 00:42:11) And different cultures have different levels of directness. And so I found I get along very well with my Israeli friends because they are incredibly direct.

(Jason Sobel at 00:42:21) They sure are. Yeah. Yep.

(Joel Beasley at 00:42:22) And when I hang out with them, I could be me. When I hang out with everyone else, I'm like, I got this layer, this—

(Jason Sobel at 00:42:28) Yeah.

(Joel Beasley at 00:42:29) Coming layer, yeah.

(Jason Sobel at 00:42:30) Of speaking, yeah. Yeah. Layer.

(Joel Beasley at 00:42:34) That's funny. The health club management layer. Yeah.

(Jason Sobel at 00:42:37) Yeah. I don't know where it came. I mean, my definitely my family, there was a fair amount of conflict when I was young. And so I think I got comfortable with that. But it's funny.

(Jason Sobel at 00:42:49) My wife, she and her brother were apparently perfect children and never got, of course, never got in trouble. And her parents confirmed this. And so now we have three kids who are not perfect children. And I find it a little easier to handle them because I was not a perfect child, and I got in trouble, and I got yelled at, whereas my wife never did.

(Jason Sobel at 00:43:12) And so she's like, what do we do with these monsters? Anyway, so, yeah, I definitely take that point on, you know, where you came from. You just have such a different wiring about how to work with people and tone and conflict. And I think that that's one of the thing that as a manager, you have to figure out really quickly is for this person in front of me, how do I say something in the way that they're going to hear it? And the way it might make sense for me or the way I have it in my head might not be what's going to make sense to them.

(Jason Sobel at 00:43:47) And so I think you kind of have to figure out each person and their language and their way of thinking so that you can, you know, you can help them get where they need to go.

(Joel Beasley at 00:43:57) I use Grok. When I want to communicate something to my wife that I am, if I feel strongly about something, I'm like, okay. That's usually an indicator. I want to communicate this. I want to get this through, but I need to get through in her language.

(Joel Beasley at 00:44:11) So I've told all my friends, like, in my personal life, dude, try using Grok or whatever, whatever AI you like. They all kind of work the same. And you can say, hey, I want to communicate this to my wife and here's what I want. And it'll give you tips on how to phrase it. So AI is fantastic. Also, I have three kids, eight, six, and three. What age ranges are yours?

(Jason Sobel at 00:44:34) Fifteen, thirteen, nine. I have two boys and then a girl. Yeah. So we're ahead of you. Yeah.

(Joel Beasley at 00:44:40) Yeah. I got one girl, girl's oldest, and then two boys.

(Jason Sobel at 00:44:43) We're the other way. Yeah. Our boys are oldest, and then our daughter's younger. It's wild seeing her with two older brothers. I think she is a little more comfortable with certain types of humor and violence. Of course. She might be otherwise.

(Joel Beasley at 00:45:02) Of course. Well, this was fantastic. Last question here for you. If you're going to hire someone, bring them on to your team specifically, not just at the New York Times. They're going to be your direct report. You're doing the interview with them. What's the one thing that you really care about that you're, every time you do an interview, you're like, I need to find out if this lines up.

(Jason Sobel at 00:45:21) Yeah. The thing I always say I'm looking for is relentless problem solvers. That's my go-to phrase. Like, I think, I don't know how specific this is to tech. It's all I've ever really worked in.

(Jason Sobel at 00:45:34) It's probably useful anywhere, but that doesn't mean you always have the answer. It doesn't always mean you know exactly what to do, but you know how to figure it out. You know how to get through or around or over whatever wall's in front of you. You know how to get help. You know what you can do on your own, what you can't.

(Jason Sobel at 00:45:53) But just that idea that we're never sitting still. We're always trying to understand what are the most important problems right now, and are we set up to solve them? So my interview is all crafted around, can I get at that core? Is this person someone who's just going to, you know, if there's a problem that I give them, they're going to knock it down, and then they're going to figure out what the next problem is and knock that one down. And that's just their mindset as they're hungry to find those opportunities and find their way through them.

(Joel Beasley at 00:46:25) It's almost like a personality trait, right? And you're trying to figure out if they have that one because you can't train it. I don't think you can train that. I'd always say I can't make someone run, but I can find someone who's running and point them in the right direction.

(Jason Sobel at 00:46:41) Yeah, I think it's really... that question of what can be learned and taught and what do you come in with? It's such an interesting one.

(Jason Sobel at 00:46:51) I do find that when you ask people about the biggest problems they've solved... one of my favorite questions is just, what's one of the biggest problems you identified? Like, no one told you to go work on it. Just over the last couple years, what's something you saw was really important for your organization or for your team, and you just got buy-in and you just made it go happen? And how did you solve that? That, to me, is always a good sign for somebody if they're identifying really important strategic things on their own and then getting people rallied around them and solving them. That to me is a good indicator of your ability to do this.

(Joel Beasley at 00:47:34) I love it. You also find this a lot in entrepreneurs.

(Jason Sobel at 00:47:39) Yeah. Man, that looks like a hard job. It's funny, people always ask me if I'll start a company someday, and...

(Joel Beasley at 00:47:44) I just always say no. It just looks really, really hard.

(Jason Sobel at 00:47:47) I mean, my job's hard enough as it is. I don't need to make it ten times harder.

(Joel Beasley at 00:47:51) No, it's... the best way I've ever heard it described—I've built and sold three companies. Oh, wow. The best way I've ever heard it described is it's staring into the abyss and eating glass.

(Joel Beasley at 00:48:01) And I was like, yeah, I'll do that two more times. Yeah, I know, I know.

(Joel Beasley at 00:48:07) I know. But it's also very rewarding.

(Jason Sobel at 00:48:10) I'm sure. I'm sure when you get it right and you've put that team together... there is something really rewarding about bringing a group together around something you're really excited about. And you can do that inside another company, but I got to imagine that's at another level when it's just you from day one.

(Jason Sobel at 00:48:29) Yeah.

(Joel Beasley at 00:48:30) It's wild. But for me, you know, it's almost like I have to because I found... I tried working other stuff, but I realized if I'm not doing something where the odds are stacked against me, that part of me isn't clicking on to get up and do stuff. So if I...

(Joel Beasley at 00:48:48) If I wake up and I stare at a problem that's so big and so difficult and it's such a long journey, to me, that's very exciting. For sure. If I wake up and it's just easy, you know, routine, I'm like, okay, you know?

(Jason Sobel at 00:49:00) Totally. No, I mean, that's a lot how I think about my career. You know, if I can kind of see on the horizon that it's gonna get easy or there's not something I'm really excited about, then that's a good time to start thinking about how do I change something up internally, or should I start thinking about something new? So at Facebook, for example, I had three or four different jobs in six years because I just wanted to experience new things.

(Joel Beasley at 00:49:27) Yeah. But your ability to recognize that within yourself—first of all, that puts you at the next level. Then you even go to the next level by being able to act on that. So it's like, first observing it, the second is to act on it, and then the third is to realize it's a pattern and keep doing it intentionally. And that that's literally—I've done a thousand interviews, and so I'm always trying to figure out where everybody's at and why some people are up here and some... what's the difference between the top 10% and the top 0.1%? And it's really... self-awareness has got to be one of the biggest ones because that creates this energy that, when the energy gets applied correctly, it just takes them right to the top.

(Jason Sobel at 00:50:10) Yeah. I gotta bring my wife in here. You can tell her how self-aware I am.

(Joel Beasley at 00:50:17) She's anything like my wife, it won't mean that much. She'll be like, tell the nerd to be quiet. Awesome. Jason, this is an absolute pleasure. I just won a new friend. Thank you for doing this, man. I really appreciate it.

(Jason Sobel at 00:50:28) Thanks for having me. This was a blast.

(Joel Beasley at 00:50:30) Thank you so much for listening. And if you found this episode useful, please share it with a friend or colleague who you think would get value from it. And if you have topics that you'd like to hear discussed on the podcast, either add me on LinkedIn or send me an email: [email protected]. Every time I get an email or LinkedIn message, it absolutely makes my day and inspires me to keep going.