The Stack Overflow Podcast

Will developers return to hostile offices?

Episode Summary

Ben, Ryan, and Eira convene to discuss return-to-office mandates, what’s surprising about employee attrition in 2023, and how technology can preserve digital records of cultural heritage sites before they’re lost for good.

Episode Notes

As the year winds to a close, some big employers are facing lower-than-expected attrition rates—in other words, fewer people than expected are quitting. What a difference a year or two makes.

People have strong opinions on the return-to-office conversation. Read Eira’s article and let us know how you feel.

We are just beginning to explore the effects of prompting on the capabilities and performance of LLMs.

The Humane AI pin can be described as a cross between two of humanity’s most beloved technologies: Google Glass and the pager.

People are using low-cost drones, 3D printers, and private satellites to preserve irreplaceable cultural heritage sites before they are destroyed or lost to time. (Stay tuned while Eira figures out how to apply this tech as a cemetery tour guide.)

Stack Overflow user FlipperPA earned a Lifeboat badge with their answer to The 'Black' formatter - Python.

Episode Transcription

[intro music plays]

Ben Popper Imagine a workstation where your devices seem to disappear, keeping you in a state of flow for hours. Imagine a superior typing experience and a mouse crafted for comfort. Now add smart illumination, programmable hotkeys, smart software, and connection to up to three devices. Discover MX Master Series. Crafted for performance, designed for coders. Find out more on logitech.com.

BP Hello, everybody. Welcome back to the Stack Overflow Podcast: Home Team/Content Team Edition. I'm your host, Ben Popper, Director of Content here at Stack Overflow, joined by my colleagues and comrades-in-arms, Ryan Donovan and Eira May. How's it going, y'all? 

Ryan Donovan Oh, pretty good. 

Eira May Not too bad for a Monday. 

BP Not too bad for a Monday. Eira, you brought us a story here. We're not going to name any of the big companies involved, but we will discuss the trend. What is this piece about? It's about attrition. We went through the last two years. There was the pandemic, that was a lot of layoffs, then we went through maybe 18 months of crazy hiring and quiet quitting. And now where are we? 

EM Well, according to this article, this is in The Wall Street Journal, talking about how we're sort of seeing the reverse of the great resignation. One of the first pieces that I wrote for the Stack Overflow Blog when I started working here was about how the great resignation was impacting developers in the fall of 2021. And so now here, almost exactly two years later, it sounds like at big white collar employers, people are not quitting. The attrition rates are not as budgeted for, and as a result, employers are having to reduce the number of employees in order to make up for the ones that they expected to quit who didn't quit.

BP The old fashioned way. No more Zoomers and their quiet quitting. 

RD We had so many years of tech being this golden oasis where it was like the grass is always greener somewhere else, so you stay a couple of years and you jump and then you jump. And now things are coming back to a little bit of reality, I think. 

BP Yeah, snapping back to reality. If you are experiencing sort of a whiplash with the quiet quitting and now the slow, subtle drip of firing or not rehiring, no backfilling, shoot us an email and we can talk about it on the podcast; we can talk about it on the blog. We’d love to hear from you. All right, let's make a transition– I have a story I think is interesting that I want to share. It’s in the world of prompt science: “Poking the LLM to see what it will do.” And so this was a piece of research from AWS that one of the main things people are asking these Gen AI systems to do is summarize– take a look at all these documents and tell me what it means, or read me this quarterly report and give me the best of it. Increasingly, I think that's how people are going to start to get their news or almost their search functionality. They're going to say, “What was that movie about again?” or “Whatever happened to that actress? or “What's with this current event?” and the AI is going to go read a bunch of articles, or it has consumed a bunch of text and it's going to summarize it for them. And so what this study found which was so interesting is that they called it chain of density. So chain of thought prompting is something we know gets much better results. If you ask the AI, “Hey, I'm going to ask you a question, but then show me your thought process as you work through this and then deliver me the answer.” So for chain of density, it would write a summary of the events that has almost no information, and then it would rewrite that at the same level but it has a little bit more, and then rewrite that at the same number of words but has a little bit more, and just keep going like that, 30, 40, 50 times and the end result is something that both empirically and subjectively by the judgment of human readers was far better than the original and the ones the summary is written by actual people. So I like this field where it's just kind of poking and prodding at a black hole where we don't really know what's going to make it better. It's like a big blank space on the map and you can just kind of explore.

RD It's all practice, no theory yet. 

BP Yeah, exactly. 

RD I think it's interesting. I think that a lot of the summary use cases we've seen have been like, “Take a book and then summarize it down to a chapter, and then down to a page, and then to a paragraph,” and this is interesting and it's building it back up.

EM It actually makes me think of, in reverse, a writing technique that someone suggested I use in grad school, which is when you're writing about a topic that's really complex and unwieldy and you're not really sure how to get a handle on it, start by explaining it in just the plainest most simple terms that you can imagine, as if you were explaining it to a 10-year-old, and then build the levels of complexity from there. But if you start with that fundamental basic truth– truth is a little loaded– but that basic narrative, you can kind of have that coherence all the way through. I don't know if it works the same way for an AI. I guess no one really knows. 

BP You have to come up with a fancy. No one really knows. We only know the results of this is that, “Oh, it seemed to work here.” Why? We'll find out in the future, or maybe we won't.

EM Right.

RD You have to have an SEO-friendly name for it. 

BP Yeah, chain of density is a good name. It sounds like a band, a hardcore metal band– “Chain of Density.” Ryan, what have you brought to the table for us today? 

RD We've all heard of the Humane AI pin, the little Star Trek transponder thing. The article I have is from Ars Technica that was saying this is a pretty bizarre cross between Google Glass and a pager. It's not really something that they understand is useful. They don't do apps and that's basically been make-or-break for a lot of people. 

EM I’ve got to say Google Glass and a pager…

RD Not a great combo. 

BP Two technologies that haven't really been at their height of popularity recently. I sympathize with their mission of less screens. 

RD It seems cool. I like that it's not always listening, that's great. But I think in our company chat, somebody pointed out that this is going to be stolen a lot. It's just on your shirt by a magnet. It's going to be easy to snatch. 

EM I'm going to lose it immediately. Forget about stealing it. 

BP The thing here is, I sympathize with the mission of less screen time and I'm going through this with my kids reading a lot of science about how bad screen time and social media has been for the teenagers who came 5-10 years before. And so we got them a flip phone, where you get to start to have some of the functionality. You can go to the park and I can connect to you. Your friends can text you, yada, yada, yada. But when I was reading this, I was just thinking, “I can do all of this same stuff with my smartwatch. I wear it on my wrist and I can talk to it, and I can take calls, and I'm sure if I needed to do the other things like translate. Why do I need it?” They were talking like somehow they've reinvented the wheel. I guess it has the little hologram that it projects in your hand, but literally almost all of these capabilities, if you don't want to have a phone with you all the time, you could just have in your smartwatch, which is a very elegant and mass-produced device. 

EM Right, it does seem like a bit of a solved problem from that perspective. 

BP This is like a smartwatch without the band. You wear it on your shirt. 

EM A smartwatch that's easier to lose is just what I'm hearing. 

BP That's their tagline. 

EM I hope they can do better than that. 

BP So Ryan, we were chatting a little bit earlier in the episode about kind of the contentious return to office. There's a lot of discourse about it in the tech world, you see it all the time on Hacker News and even news articles about which company will or won't enforce these rules and to what degree. But you shared a piece that I thought was interesting in that it's not so much about whether or not people should return to office, but what's tragic about being in an office. It's funny to even talk about because maybe most people from now on will not be in an office. I don't know, maybe the generation that follows us will not know. But there are things that are great about an office and there are also lots of terrible things, office politics and office foibles, and go watch Office Space and take it all in, it's a rich tableau. What is this piece specifically talking about and is it coming from the perspective of developers? I think it is. 

RD I think so. It's talking about how remote isn't always the best and that there are benefits to being in the office, but just forcing people into the office in the sort of half measure where you're in the office, but you're still async is actually the worst way to work. A lot of people, when everybody went remote, found the benefits of async working. I think we've talked about it. We found the benefits of that too, because you can get your things done when you want. But in an office, it's the worst, especially those open plan offices. Man, those were the worst things I ever worked in. I came out of my days there so mad. 

BP Just other people infringing on your space, or their noise, or their what? 

RD Yeah, it was just noise. I never wished for a cubicle more than when I was in an open plan office. 

BP I do remember at the Stack Overflow office in New York, marketing folks sat in an open plan office with some rev ops and some other people. Developers each got their own little cubicle. They were special and they needed their space and their privacy and their quiet. 

RD We were in the dungeon over there. 

BP We were down in the dungeon, exactly. But this article also points out that in an async world and then in an office, there's a lot of chance to feel left out or to wonder about whether or not you're being included or whether you're stuck waiting on something and that person is literally right in the same office with you. So it’s interesting to think about those points of friction.

RD When you have an office in sync, they talk about the extreme programming practices as a way to kind of start and get everybody together and make sure everybody is working directly together. But remote and async is basically open source. And there are companies that have talked about inner source, which is the open source methodologies applied to working in a closed source environment. 

BP I was just going to say that the thing I liked about this was that they had these quadrant charts. When it's time to whiteboard, you could do it virtually, but whiteboarding, brainstorming, ideating, let's do it synchronously. And then, now we've each got things we need to deliver, let's do it asynchronously, and we've got tickets. And maybe an office or an in-person environment is useful once a week, two times a month, once a quarter. Thinking about it that way is more useful than just that it has to be either remote or office and there is no in between. 

RD The interesting one for me besides these was the remote and sync, which I don't think anybody is really doing very well. I think for the folks that do it, it feels like monitoring. It's always on Zoom, it's a lot of directly working together on virtual whiteboards and such. Eira, have you worked in an office in-person? Do you have preferences? 

EM I have a strong preference for working remotely. I think, like you, I was sort of scarred by working as a member of a marketing team in an open floor plan office, both in-house at tech companies and in agency environments. And even with the noise canceling headphones, I just find myself really easily distracted. And I was just thinking about how back in 2013, I had a job where we never collaborated really in-person at all. It was just everybody sort of async in their open floor plan office. And so after a few months I was like, “Hey, can I work from home sometime?” And my boss just acted like I had asked for something just beyond the pale, just totally ludicrous, and now I know the same company is entirely remote. It's just a completely different landscape. I think a lot of people have come over to seeing the benefits of having a little more flexibility and autonomy over their schedules, and it's a hard ask to just bring everybody back without any level of meeting halfway or anything. 

BP What is it you think people who are demanding a full return to the office are searching for? I do think that there's a certain kind of manager and leader who feels like, “I'm going to stay until 8 PM and see who stays. I'm going to see who is showing up early and I'm going to push those people and elevate those people.” And it's more difficult when you're remote to know who's doing that, and also, what were those people doing till 8 PM? Were they delivering projects or were they just silently, begrudgingly, sticking it out at the office so that they could be the last one to leave? I remember that happening a lot and just being like, “I'm going home to my kids. I don't know, this isn't for me. You win.” 

EM The performativity of staying late. 

RD I feel like those are the kinds of managers that measure productivity by lines of code. I feel like that's negative management, where you're sort of playing defense on your reports. They're a cost drain, they're always trying to take advantage of you so you have to be there to ride them and make sure that they're in the office all the time doing stuff. It's a miserable way to work. 

BP All right, one last great little tech highlight. This comes from my days working at The Verge as a drone reporter and then working at DJI. Folks are using low-cost drones, 3D printers, and private satellites to preserve historical artifacts. So sadly, in the world there are conflicts. People drop bombs. It's not chill. We don't like it here at Stack Overflow. And sometimes it also has a big effect on a society's history, its monuments, its landmarks. And what's interesting now is that these sort of, not hobbyists, but volunteers with very low tech and low cost tools can go out and create these stunningly high-res preservations of these temples or statues or whatever they may be, flying the drones around them using the satellites to map the flightpath that they'll take from the air, and then 3D printing both parts to do this, and maybe also the objects themselves, some simulacrum of what they were trying to keep as part of their history. 

RD I think it's a great use of the technology. I'm happy to see good, low-cost 3D modeling tools. I remember 10-15 years ago somebody thought they could do it with the Xbox Kinect and model objects with that. And it did a little bit, but using the generative AI that we have now, combined with the high-res and all the advancements since 15 years, it'd be great to have an image of this beautiful ancient temple that was blown up for whatever reason. 

EM Yeah, and something that you could share with people in different geographic areas and sort of use as an educational tool. I could see it being fascinating. Even if we take things getting blown up out of the equation, it's a lot more economical perhaps to do a tour of some architectural features in another country in a virtual sense. There might be applications there that would be fascinating, to me at least. 

RD And it helps preserve the sites too. 

BP Right. Yeah, maybe people can pay for the virtual tour of the graveyard, Eira.

EM There you go. Thank you.

BP If y'all don't know this listening, there is a graveyard tour guide. And then my online donation for the virtual tour goes to support the real life preservation. 

RD There you go. 

EM That's right, that's right. That's how I'm going to monetize it finally. 

BP Exactly. 

RD I think the actual result of this will be GTA Machu Picchu or whatever. 

EM Yeah, exactly. 

BP Yeah, I’ve got to have this in Roblox. 

RD Virtual gun fights at the Parthenon. 

EM I thought you were going to say the O.K. Corral, but you really mixed it up there.

RD That's right. 

BP Everything I know about ancient history I learned from Civilization

EM There's no lies in Civ. 

BP The Granary, the most important.

[music plays]

BP All right, everybody. It is that time of the show. Let's shout out somebody on Stack Overflow who came and shared some knowledge and helped some folks get unstuck. A Lifeboat Badge was awarded to FlipperPA for helping somebody who was having an issue with the Black formatter module in Visual Studio Code. It was doing the worst thing you could do– replacing double quotes over single quotes. Unacceptable styling, but you can use skip-string-normalization to fix this. And if you want to know more about it, FlipperPA has the answer and has helped over 12,000 people, so we appreciate it, Flipper, and congrats on your Lifeboat Badge. As always, y'all, thanks for listening. I am Ben Popper, Director of Content here at Stack Overflow. Find me on X @BenPopper. Hit us up with questions or suggestions for the program: podcast@stackoverflow.com. And if you like what you hear, leave us a rating and a review, because it helps. 

RD I'm Ryan Donovan. I edit the blog here at Stack Overflow. You can find it at stackoverflow.blog. And you can find me upon the X @RThorDonovan. 

EM And my name is Eira May, and I am also on the Content Team/Editorial Team at Stack Overflow. And you can find me on most platforms @EiraMaybe.

[outro music plays]