AI, Rust, and Resilience: Key Software Trends Seen by the QCon San Francisco 2024 Program Committee

Transcript

Thomas Betts: Hello and welcome to another episode of the InfoQ Podcast. Today I'm joined by members of the QCon San Francisco Programming Committee, who've been hard at work, making sure the conference coming up in November is filled with talks about the latest innovations in software. This episode is not intended to be a pitch to get you to register for QCon.

InfoQ is known for its annual trends reports, and each of those reports is on a different aspect of software development such as architecture and design, AI and ML or culture and methods.

I feel like today's conversation is going to be a similar theme of the trends happening right now, but across the full spectrum of software, and that's because the programming committee discusses those trends and that's what we use to decide what goes into a QCon conference. Now, before I go any further, let's meet the people who are on this year's committee. We'll go around quickly just introduce who you are and what's your current role. We'll start with Khawaja.

Khawaja Shams: Hi, my name is Khawaja. I'm a co-founder at Momento, and I'm on the program committee at QCon San Francisco.

Thomas Betts: And Suhail?

Suhail Patel: Hi, my name is Suhail. I am a senior staff engineer at Monzo Bank in the UK and I am also a program committee member for QCon in San Francisco.

Thomas Betts: And Hien?

Hien Luu: Hi, my name is Hien Luu, and I currently working at Zoox, leading the ML platform team as well I'm a member of the programming committee at QCon SF.

Thomas Betts: And Erin?

Erin Doyle: Hi, I'm Erin Doyle, I am a staff engineer at Lob, and I work on the platform team there. And we'll repeat my teammates’ intros here and say I'm also on the programming committee for QCon SF.

Thomas Betts: And just so I don't leave them out, Michelle Brush was also on the committee, but she's not able to join us today. And then I am a member of this group. I'm not just hosting the podcast today. Since I haven't done an introduction in a while, I'm Thomas Betts. I'm an application architect at BlackBaud, where I help build software that powers the social impact community. I'm the lead editor for Architecture and Design at InfoQ. And since everyone else has said it, a member of the program committee for QCon San Francisco.

For those that aren't familiar, QCon San Francisco is a little different than other conferences where they submit a call for presentations and the speakers say, please accept me to your conference. QCon is entirely curated. We meet together as a program committee, decide what we want the conference to be about. We pick out those themes that are relevant right now this year in software, and then we find track hosts and basically a track host is in charge of making a mini conference on one topic, find five speakers on one theme.

So that's kind of the background of how a QCon comes together. There's more about it on the website. You can read about how that is. You can also go to qconsf.com and see the entire schedule. We're not going to cover every talk today.

GenAI in Production and Advancements [02:37]

Thomas Betts: I want to highlight the big themes that we're talking about, what are the big trends. And if we're talking about current trends, the large language elephant in the room is Gen AI. QCon has two of its 12 tracks this year dedicated to AI and machine learning. Plus there are a handful of talks on other tracks that relate to the subject. Hien, I'll start with you because you're the PC champion for the two AI and ML tracks. What areas did QCon choose to focus on for those topics?

Hien Luu: Yes, thank you, Thomas. First of all, Gen AI is one of my favorite topics. I'm a very big believer in this, and that's why I championed this particular track. Our track is called Generative AI in Production and Advancements. Notice that's the two parts, the production and advancements. Over the last two years, Gen AI technology, there's a lot of interesting development. A lot of companies are trying to figure out how to adopt these Gen AI technologies to improve their products, customer experience and such. The first part of this track is to bring knowledge and experience from various companies around the world that have been bringing this technology and building product on top of this. So that's why we invited folks that have done this and they would like to share with the rest of the community in terms of best practices, learnings, pitfalls, and all that good stuff like that.

So that's why we have sessions like how to bring Gen AI for employee productivities, how to deploy these ML/LM models into the enterprise. There's lots of tips and tricks as well. And then some use cases from companies like Pinterest that really integrate into how to do search. So these are some of the sessions that we want to bring people that have been there done that as well as there's some advancements. This field is evolving rather rapidly and one of the advancements that's very exciting in the industry right now is about AI agents. So we have a session to help us to learn about what it is and how do we best integrate that into our products moving forward. So I'm very excited about this track as far as the lineup, lots of practical knowledge and experience being shared as well as looking forward what the next few years might be like. So that's for me.

Thomas Betts: Yes, I like that idea of getting those people who have this experience that have done this stuff in production that's kind of foundational to QCon. We like the idea of being the information Robin Hoods, take the knowledge from a few people who have it and share it more broadly. So that's the Gen AI Production and Advancements.

AI and ML for Software Engineers, the Foundational Insights [05:08]

Thomas Betts: What about the other one, the AI and ML for Software Engineers, the Foundational Insights? Is that more about how to get started?

Hien Luu: Yes. For that one, it's a little bit different, but ML is still bread and butter for a lot of companies in terms of these predictive models. So it's still a lot of knowledge learning best practices there for a lot of use cases out there. So similar kind of vein. We want to bring sessions where folks has been there and done that and best practices of doing recommended systems at scale, for example. And what are some of the pitfalls of bringing model to productions and as well as a little bit of overlapping, but let's also talk about how to evaluate as we integrate these LMs into these products, how to evaluate and assess them, how is that different than assessing classical machine learning models. So a very similar vein in terms of bringing these best practices from the industry, the people that have been there done that to share the wealth of the knowledge and pitfalls.

Thomas Betts: The QCon audience is fairly technical. It's mostly senior engineers and above, but even acknowledging you might have 10 or 20 years experience, but you might not have done ML stuff. So some of that foundational insights isn't acknowledging, I'm a junior engineer, I'm new to this topic, but I want to get up to speed. So having something for everybody is really interesting. And I think that's why we think it's important.

Embracing Shift-Left in Data Architecture [06:32]

Thomas Betts: One of the other tracks that's sort of AI related, it's not specifically about ML, but it's called Embracing Shift-Left in Data Architecture. And that's I think about the idea that ML and a lot of these big data models and machine learning used to be an afterthought. We used to say, here's our system data and now do some analytics off to the side. And we're pulling that more and more to the left.

It's becoming more integrated into our products. And so that track kind of talks about how we can support those use cases, how do we take traditional batch processing and get better at stream processing. How do you build more reliability, resilience and auto-scaling to handle these complex workloads because they're now core parts of the system. Anyone have something to say about that, about how all these AI and ML things are not just an add-on and we're making them core to our systems these days?

Socio-Technical Resilience [07:18]

Erin Doyle: Yes. I'll also shout out that we have the socio-technical resilience track, which also incorporates how are we integrating AI into our teams and how is that changing the landscape of how we do our work. So it's kind of impacting everything, all aspects.

Thomas Betts: Yes. Tell me more about the socio-technical resilience. I think that's interesting. I see that the title of the track and I wonder what is that? How do I respond to these changes? And QCon's always had a few tracks dedicated to people and processes. How do I get better at being a software engineer? How do I do my job better? And is that supposed to be in this age of AI? How do I adapt?

Erin Doyle: Yes, for sure. I think it's how are we looking at how AI is changing our jobs. It's changing how we approach our work in so many aspects, and I think all throughout the engineering organization, from the IC level, we need to know how can we use it to enable us to move faster, how can we do our jobs better. How are our managers going to look at the ways we're leveraging those tools. And it really pervades the entire organization. It changes landscape completely.

The Path to Principal Engineer and Beyond [08:28]

Thomas Betts: And I know we've got another track on the path to principal engineer and beyond. That seems to be like you're talking about the individual contributor using these things. Khawaja, what else is on the professional growth and maybe it's not about AI. What other things are we talking about?

Khawaja Shams: Yes, the path to principal engineering is not AI centric as a track per se, but so much of engineering is less about writing code and architecting systems, but just about how to become a better individual contributor and how to be a better team member, so that you can maximize the productivity for yourselves and for the broader company in the process. And principal engineering is hard. It's so much more than just being a great solutions builder. It's about the energy you bring to the table. So I'm pretty excited.

We have everything from how to build a personal brand and increasing your influence to really diving in, and that's by Pablo at Harness and to really diving into the biology of our brains and how to use that to increase your influence, and just throwing as Charlotte, the speaker for this from Bravely, she calls it throwing soft skills out the door, because she's really, really passionate about this particular topic by the way, like soft skills and the way they're framed today almost seem like a dark art, but she really turns that into a science and understanding that really helps you improve the way you carry yourself and to build empathy for your teammates, which then help you become a better engineer. But it's a variety of talks in this particular arena and I highly recommend people to come visit and attend those.

Thomas Betts: I think the soft skills is a call back to an earlier podcast a few months ago that I did with Jacqui Read and her book about Communication Patterns for Architects, for Engineers. She doesn't like the term soft skills. She calls them power skills, because that's what actually enables you to be better at doing these higher level things. And you're right, it isn't just about coding. You need to contribute in a different way. It's different skills and I think that's why we chose to focus on that.

Khawaja Shams: Yes, I think the soft and soft skills typically refers to the squishiness of the definition and how malleable the definition is. And I like the new takes on it for sure.

Rust in Production [10:35]

Thomas Betts: Well, let me take us to actually writing the code. We've got a couple tracks about languages. So what are the language trends? The schedule shows one track for programming languages and paradigms for the next decade. And a second track for Rust in production. Why is Rust so special that it gets its own track, Khawaja?

Khawaja Shams: Well, Rust is a trend that everybody seems to be talking about and it's still a very nascent ecosystem, but it's finally getting to a place where we're seeing massive adoption at larger enterprises for Rust. It's ready to take off. And what we wanted to do in this particular track is not just talk to you about the philosophies of Rust and why it's faster. I think a lot of us just take that for granted now, but really inspire folks with real world use cases and how people are using it in production, what their mileage has been. In fact, there's a talk by Ramya in this particular track, where she talks about, well, we were a Java team, we went and tried Rust and let me tell you how it went, what our hypothesis were, what we learned along the way, and should we have made this transition or not. So this is not a Rust evangelism track by any means. You'll see some inspiring ways where people have taken it to production and you'll also hear some objective rationale as to why you should and shouldn't take it to production in your stack.

Thomas Betts: I think anyone who says they're a software architect at some point knows the answer is, it depends. And I think using Rust in production is another one of those, it depends. Maybe it makes sense, maybe it doesn't.

Programming Languages and Paradigms for the Next Decade [12:01]

Thomas Betts: Okay, so Rust is hot right now. Maybe it shouldn't be used everywhere, but what about predicting the future? Suhail, what are some of the highlights from the programming languages and paradigms for the next decade?

Suhail Patel: Yes. In the programming languages and paradigms track, I really wanted to shift the narrative because historically QCon has always had a programming languages track, which sort of showcases the new things that are happening in programming languages. There's always something new coming out, a new configuration language and maybe a new academic exercise. And a lot of these things are really, really interesting and popular. And I want to strip that back to the world of software engineering for how most people do it on a daily basis. So effectively we've got a mixture of talks talking about programming languages and finding and identifying bugs within programming languages, and for example, statically finding bugs with typing. So I'm really, really excited for Jake's talk from Stripe, where they've introduced Sorbet into their big code base to help find bugs with typing in Ruby. And I'm also really, really interested in Paula's take on platform engineering.

Platform engineering is a little bit of a weird talk to put within this particular track, but something that I've seen in the world of software development specifically is that there's a lot of stuff that's taken care of in the world of platform engineering. So I want to see where that line is being blurred, where you'd write pretty defensive code, you're now putting those layers of defense elsewhere and you're writing pretty linear code. A common phrase that I hear is, just let it crash. If you look at Erlang, that's quite a common paradigm. Just let it crash, and the system will recover itself. So here the platform is just let it crash and another instance will spin up and take care of it. So I'm really, really interested to see how that world has shifted. We've also got a really, really interesting set of talks on Pkl, which is a new configuration language that was recently open-sourced by Apple, as we got Dan talking about that. And also Jordan talking about Clojure and how they use it for Datomic for data system at Nubank, which I'm really, really interested in.

Thomas Betts: And just for the transcript, how do you spell Pkl?

Suhail Patel: Oh, it's P-K-L. Yes, not like the food that you eat.

Thomas Betts: But pronounced that way.

Khawaja Shams: I actually read it as P-K-I, and I was like, wait, is this a security talk or?

Thomas Betts: You have to capitalize the full acronym, otherwise you won't know what it is. It does look like P-K-I, if it's lowercase.

Topics for Platform Engineering, SREs, and Everyone Else [14:15]

Thomas Betts: Well, you mentioned platform engineering and I think that's one of the things because we only have 12 tracks at QCon, and we choose here are the big themes of the year that we want to talk about. Sometimes stuff gets left out. Someone mentioned to me that we don't have a platform engineering track and we don't have a track for SREs. Why was that? Why didn't they fit in and is that okay? Are those people still going to find something interesting or we think those aren't popular themes right now?

Suhail Patel: I identify as a platform engineer is usually the talks that I sort of gravitate to myself when I'm at QCon. And what I found really interesting and something that we've talked about as a programming committee pretty much like every week, is that in each of our tracks what talks fit into which groups of people, where do people identify and how we program the schedule and how we program the calendar all the time slots and things like that. I think there is a talk, multiple talks, I think you're going to be in good competition, multiple talks per day, even as a platform engineer or if you're on platform engineer SRE archetype, that you'll be deliberating like do I go to somewhere in hardware architectures or even in the path to principal track, where you'll be looking at some large architectural problems in the frame of people problems as well. I think you're going to be spoiled for choice if you fit within this archetype that you can't this November.

Khawaja Shams: Yes, I remember we actually went through every single slot, time slot and made sure that there was something at any given point for an SRE or a platform engineer. It's a very common job family for our audience and yes, we don't have a track called that, but to Suhail's point, if you look at the hardware architectures, you should know about, I mean that's heaven for SRE or platform engineer, like CPU based fine-tuning for llama, and then there's a whole lot of systems level things that get covered in the Rust track as well and deploying it in production, using it with serverless, fine tuning observability on it. So I think there's a lot of incredible sessions for a given SRE. To me personally, it's better to not just have all the SREs in one room either, and this helps the SRE ultimately get a better experience by still having a menu of choices to pick from, but also intermingling with other parts of the community and attendees at QCon.

Erin Doyle: I have to agree. I am also a platform engineer and I am struggling with how am I going to go to all the sessions that I'm interested in. We also have the Architectural Evolution track, which is got a lot of great talking about why certain companies and teams made the shifts that they did. I think that's super relevant and of course architectures you've always wondered about. We're going to get into some nitty-gritty of things some really interesting companies have been doing. So I agree. I think we've got some good stuff.

Thomas Betts: And I think that's one thing I personally enjoy about QCon is it doesn't pigeonhole people like, oh, you're a programmer, you're going to go to this programming session, you're going to do blah, blah, blah. The conference exposes you to these other things. I have found more value in some of the people and processes tracks than I ever expected to. A lot of the stuff a few years ago was how to make remote work successful. Turns out you needed to know all that stuff now. So I like that we think of the person not just, oh, you have the job title of SRE and here's a thing for you to do. It's you are in the software development community and you should try and expose yourself to a little bit more of these things, and then you get to meet the people who aren't in the same role as you but are interested in the same topic. I think that crossover is fantastic.

Khawaja Shams: I'll just reinforce that with, I mean, it's for the same reason why we don't have a software engineering track like you said, right? And we want that intermingling.

Suhail Patel: One of the best top tips I've ever gotten is to broaden your perspective and horizon. You go to conferences to enhance your knowledge base, so to sit in a track where that is what you do on a day in day out basis, you're already exposed to a lot of the depth, then you're definitely going to learn something, most definitely going to take away something, something that might be applicable, but also spending some time on something that is completely left field for what you work on a day-to-day basis. So for example, if you work as an SRE, I think there'll be a lot of stuff in the fundamental ML tracks that Hien was talking about a little bit earlier, or even socio-technical resilience as well, the sort of mixture between technical and socio-economic factors as well. So I think there's something to learn for everyone, even just by diversifying your horizons.

Thomas Betts: Yes, I think that's the same mindset that got us whatever ops, right? Start with DevOps and now it's MLOps and DevSecOps, like you're putting all these things together and that's not a, oh, this is a new role, we should have that one person go do the DevOps thing. The idea was let's bring the developers and the ops people together because you have two different viewpoints of how to solve this problem, and so put them in the room together and you get better results. Oh, here's how I would solve it as a software engineer, but in operations I want better analytics and logging and observability. Oh, well, I can build that in and then you work together. And so that's what I like about going to these conferences. What I liked about being on the committee is talking about all these topics and trends that you all find important that weren't top of mind for me, because they're not stuff that I'm dealing with my day job, but I'm now excited.

Next-Generation Inclusive UIs [19:15]

Thomas Betts: Same thing Erin said. I don't know how I'm going to go to all these sessions, so I'm going to watch all the videos afterwards. So there's a similar mindset for what will be trending over the next few years. We just talked about the language trends over the next few years, and that's getting to next generation inclusive UIs. Erin, what do you think these inclusive UIs are going to look like and why is it an important trend for us to talk about right now at the conference?

Erin Doyle: I've done a lot of work in web accessibility in my career and it's a topic that while we certainly don't have it nailed, we've got a lot of work to do, making the web more inclusive and accessible, but it's something that people talk about. There's lots of resources out there, there's lots of courses and whatnot. What we're not hearing a lot about yet is how do we make these other UIs inclusive. And many of them, so when we're getting into augmented reality, virtual reality, mixed reality, extended reality, voice controlled UIs, you can go on and on. These are all very sensory based. So the experience is really designed to be pretty deep into specific senses. And so if someone has a limited capability, whether temporary or permanent, they're either going to have a subpar experience or they might not be able to have that experience at all.

So it's like we need to look further than the web and bring that mindset of how can we approach whatever the user interface is with a more inclusive perspective. And we can't just put this on designers really as engineers, we need to be thinking outside of the box of how can we still achieve the goal, whether it's the experience, whether it's a task somebody needs to do, whatever that is that we're building in our product, how can our users achieve that in either supported ways or completely alternative ways. And the tools and the solutions people are coming up with are really innovative. So that's why I thought this would be a great fit for the conference.

Thomas Betts: Yes, I think QCon and InfoQ really focused on those innovator and early adopter trends, and so these are things that are out there in the forefront. There wasn't an inclusive UI in virtual reality four years ago. That wasn't a thing, right? VR barely existed, AR does all those. You listed off like five of them. I barely kept track, but there's a lot of those different things we're going to start being exposed to and just the chance to think about it again, you might not even realize it's the thing you need to think about, but if you go to one of those sessions and sits in the back of your head, maybe when you work on that project a year from now, you'd be like, oh, I remember someone talking about that. We need to be thinking how do we handle these specific use cases? How do we make this more inclusive and get that baked in early on? I like that you said it's not just for the designers to solve the problem. The engineers have to be involved.

Erin Doyle: And I really want to strongly encourage anybody coming to the conference that works on any user facing interfaces at all. I think they're going to get a lot out of this track. You don't have to be working on XR, spatial computing, AR, VR, the whole list. You don't have to be doing any of that specific work to get a lot of insight out of what these talks have to share.

Getting the Most Out of a Conference [22:20]

Khawaja Shams: Two of my favorite things at QCon that I highly recommend every audience member to do, one is to go attend tracks and talks that have nothing to do with what you work on, if only to open up your horizons and to learn about the intricacies of a new domain. And the second one is interrelated is to try to meet people. So much of the power of QCon comes from the intermingling with the other audience. I've met some incredible people just sitting next to them in a talk at QCon and watching them ask questions that kind of tickled my brain and inspired me to walk up to them and have a conversation. It is an incredibly inclusive community, and one way to be a part of that is to attend tracks that have nothing to do with what you do. And the other one is to just strike conversations with others that aren't necessarily working on exactly what you work on either.

Suhail Patel: Just to add to that, I think a really good opportunity is to attend the unconferences as well, because you get a lot of like-minded individuals as unconference sessions for vast majority of these track, and they effectively invite like-minded people to have a conversation. And I think the way that they are run with a flip chart and small groups and things like that make it a really, really inclusive community, a community where you can be heard. And I've definitely had friendships sparked from that. People working on similar things and that intermingled with the speaker to attendee ratio really, really helps because you basically getting access to these practitioners that work on this and think about these things day in day out that you can ask questions to for three days.

Erin Doyle: Yes, I love that you brought up the unconference because I love those. In all honestly, I'm an introvert, so it's really hard for me to go to these conferences where there's just this massive room of people and they're all talking to each other, and how do I approach walking up and talking to someone and networking, and once you get over that little hurdle at the unconference of stepping up and speaking up and making a suggestion, suddenly you have all these people who have the same questions that you do or who have answers, and you're striking up a really relevant and fruitful conversation with, like you said, like-minded people just like you that maybe you had no idea you were surrounded by people who have the same problems and thoughts and issues that you do.

Thomas Betts: Yes, I'll just chime in. The unconferences are fantastic. I've been going to those since the first time I went to a QCon years and years ago. And for those people that aren't familiar with them, instead of wandering the hallway and having to walk up to someone, and as an introvert, that's really difficult and saying hi and approaching the conversation. It's also not, we go and sit in a room and then whatever happens happens. It is just enough structure so that it's a productive conversation for a small amount of time. It's a time box conversation. You put topics up. It's a little bit moderated just to make sure that it flows well. But as Erin said, it's for the people, again, there's a lot of introverts in software. I think we can agree that.

It makes it a lot easier to put up a sticky note on the board and say, I would like to talk about this, and you see three more sticky notes go up that are basically the same thing, and pretty soon it's like, "Oh, I'm not alone with that question. We all have something to come into this conversation with". And that's where it's really valuable is being able to share those ideas that yes, there are people up on the stage, but they're all practitioners as well. None of these are amazing experts. We're all doing the job, and the people in the audience are also practitioners doing their job. And you have something to learn from all of them too. It's not just listening to the presentations.

Khawaja Shams: Yes, the one nuance I'll make is it's a bunch of like-minded people, but they all have their unique perspectives, and they share the same problems and passion for the same topic, but it's just nice to hear unique perspectives from different people in different companies, different geographies as well.

Architectural Evolution [26:05]

Thomas Betts: So it wouldn't be a QCon without a good amount of content for architects and software architecture discussions. We briefly mentioned the architectural evolution. Who can tell me a little bit more about what's in that track? Let's go into that a bit.

Khawaja Shams: Well, I think one of the parts of it was getting back to the basics, and we see in so many places, we see the pendulum swinging from one direction to another, like monoliths to microservices, serverless back to servers, and then maybe some people went all in on microservices and they realized that there were way too many services to tangle around and that maybe they should head back into the other direction. So this particular track talks about the evolution and also areas where people are trending backwards in time, but forward in productivity, and informing their unique architectural instead of just becoming zealots about a particular architectural philosophy. So part of being an architecture, being an architect and as, Thomas, you mentioned it depends. This is a track where you come in and find out what does it depend on.

Hardware Architectures You Need to Know About [27:07]

Thomas Betts: Now, QCon is a software development conference. Suhail, can you please explain why there's a track titled Hardware Architectures you need to know about?

Suhail Patel: Oh, yes. Great question. Well, we're going to rename and repurpose QCon, right? We're going to shift it one track at a time into a hardware conference and get rid of all this silly software stuff. Now, in reality with hardware architectures, engineers need to become more and more hyper aware about what's going on in the hardware, squeezing out every ounce of performance. Typically, that's been limited to trading funds and hedge funds and things like that where a lot of the microseconds nanoseconds have matters. But nowadays, we're running these really, really large AI models. Our demands for compute are ever higher. And as much as the big cloud providers would like to tell you, servers are not infinite. If you try to get a GPU nowadays, you'll be waiting back of the line, unless you have a really friendly relationship with your account manager, your friendly cloud provider, you'll be waiting back of the line.

So being able to do more with less. And I think we've got an entire roster of really wonderful speakers who have effectively achieved that. We've got people coming in from the big CPU manufacturers, AMD, Intel, coming from companies like Geico, where they've been in the field and they have effectively optimized their applications, tuning models and tuning pieces of software and things like that, so that they run better on the architectures that we use on a day-to-day basis. So having a bit more empathy for the hardware is becoming ever important, and I think this track has a really well-rounded roster of speakers that sort of speak to that.

Thomas Betts: And I'm going to give a little plug for, Khawaja, you're also doing one of the keynotes on something about hardware. Can you give us a little teaser for what that is?

Khawaja Shams: My keynote is about the use of newer hardware types, and I think this is common now in terrestrial computing, where people are starting to become more aware of what's underneath the software that they're writing, as Suhail mentioned. My keynote covers the hardware architectures that we're willing to send to space. And traditionally for space, there was very, very specialized architecture hardware that would get sent up into space with massive vibration testing and radiation hardening and so forth. The flip side of that is that hardware is not going to be very powerful because all of your innovation has gone into radiation hardening, and the world has advanced so much with Arm and with GPU, so the keynote covers taking bold risks and sending new unique architectures up into space and the science that comes from it via NASA's latest missions.

Thomas Betts: Yes, I think it's a little bit about hardware, but also sounds like you said, it's about taking risks. I've always found keynotes at QCon to be very thought-provoking. And so that's probably one of those examples of I'm not the person building hardware that's going to Mars. That sounds cool, but that's not my job. But I'd like to learn about it, and I'd like to understand what were the challenges they had, and if I face similar challenges, how do I relate what I'm facing to that and how do I let that influence my decision?

Architectures You’ve Always Wondered About [30:12]

Thomas Betts: And then I have to talk about the architectures you've always wondered about. That's been a mainstay at QCon since the beginning 18 years ago. I'm not going to go through every talk. We've got all the biggest hits, all the companies you'd want to know about Netflix, Uber Eats, Slack, GitHub.

One that I'm really interested in though is GitHub. It's not a talk about using an LLM. It's about the networking stack, about how do you make something that's in your IDE super performant because you want that very fast response time, so it acts like it's running on your computer when it's actually running in the cloud somewhere really far away. And I think that's what's a fun idea at QCon is these things that just work and you wonder, it's like, well, I take it for granted. How does that actually work on my machine or the applications I use? And that's where QCon tries to bring in some of those talks of people that you know are working on software, they're engineers just like you, and they're solving a slightly different problem. How did they accomplish that? So maybe you'll get something out of it that you can use for your software. Maybe it's just a fun story to know the details of how it works. So I'm hoping there's something in that track for everybody. I'll be hosting it this year. So yes, I'm looking forward to the architectures you've always wondered about.

Closing Thoughts [31:17]

Khawaja Shams: It's always my favorite track of the conference, and like you said, it's been a mainstay and you always learn about such fascinating architectures and unique perspectives on companies that you know and admire.

Erin Doyle: Yes, I'd say super excited about that as well as your keynote, Khawaja, because these are really novel, complex problems in really interesting contexts that a lot of us aren't dealing with these kinds of problems day to day. And so it's really interesting to hear how they've been approached and solved and maybe what issues and hiccups they ran into and how did they finally figure it out. So definitely interesting stuff.

Thomas Betts: Any last thoughts from you, Suhail?

Suhail Patel: I'm particularly interested in the Rust in production track, and I love the architectures track, very similar to Khawaja, but the Rust track has some really, really interesting talks. For me, the track that I'm going to keep a very keen eye on is the socio-technical resilience, mostly because I think that's the one that I know least about. And that's the one where I want to broaden my horizons. So I'm going to try and spend most of my time in that particular track, maybe flip-flopping between all three. And so that means we are not spoiled for choice just on the Monday itself. And yes, we've got two days afterwards.

Thomas Betts: I'll echo all that. The number of sessions I want to go to greatly exceeds my capacity to be in multiple places at once. I've really enjoyed working with you guys for the last several months discussing all these things. It's been topics that have just been ruminating in my head, and now seeing the conference come together, seeing these speakers, I think it really is, these are the trends that are important right now. This is what people want to know about, and I hope the audience that shows up at QCon San Francisco appreciates all the things we're bringing together.

So I want to thank my guests today, Khawaja, Hien, Suhail and Erin. And if you want to know more about QCon San Francisco and to register, go to qconsf.com. And registration is now also open for QCon London, which is in April. So thanks again for listening to another episode of the InfoQ Podcast.