73: Is Pair Programming Dead? | Code Quality & Engineering Culture Debate

Brittany Ellich (00:00) Welcome to the Overcommitted Podcast, your weekly dose of real engineering conversations. I’m your host this week, Brittany, and I am joined by

Bethany (00:08) Hey, I’m Bethany.

Erika (00:10) And I’m Erika

Brittany Ellich (00:11) We met several years ago working on a team at GitHub and realized we were all obsessed with getting better at what we do. We started this podcast to share what we’ve learned, and we talk about anything from leveling up your career or your skills or navigating professional development, all with the goal of creating a community where engineers can learn and connect. Today on Overcommitted, we are trying out one of our new themed episodes, and this theme is called

Overrated or underrated? We are going to go through a list of different topics related to software engineering, some that are overrated, some that are underrated. We’re gonna give our take on each of them and why. yeah, then we’re gonna just get into it. Are we ready? Are you are you both ready?

Bethany (00:58) Let’s

do it.

Erika (01:00) So to this.

Brittany Ellich (01:00) Excellent. Okay, yeah, I’m excited too. I I’m excited about this one. So we’re gonna start with some like very easy ones. the first one is dark mode. Overrated or underrated?

Erika (01:15) So I’ve heard that like this is some this has some basis in being better for your eyes. I’ve always heard it as like cool people use dark mode and like like noobs use light mode. So you know, I’m gonna say so I did look into like if there’s any science behind dark mode being better for your eyes, and it doesn’t look like there actually is. so I feel like then it just comes down to like

personal preference. I feel like the idea that like experienced people or like real developers use dark mode is definitely overrated because there are times when like you’re just working on something and like having a little bit of light makes it easier to see. And I yeah so I I think I’m gonna fall on the side of overrated as kind of a differentiator between cool people and not cool people.

Yeah.

Bethany (02:08) Hmm.

I like that. I did not do as much research as you, but I do I agree with you. But I will say I think it is underrated because so while we’ve made a lot of progress with like the dark mode experiences and various applications and stuff, like even outside the tech world, there are still so many applications that just have terrible dark mode support. They just put it in

Erika (02:30) no.

Bethany (02:31) the corner to check off a box.

And I think that both light mode and dark mode should be an equivalent experience. And I think for that reason, because they’re not always a typical or a equivalent experience, that it is underrated.

Brittany Ellich (02:46) Mmm, I like that nuance. I put overrated only because I’m definitely a light mode user. I I really like light mode. I think light mode’s more reason readable and I really do think that it’s personal preference. but I agree that you know everything should support both, ideally. Yeah, that it’s yeah.

Erika (03:06) And really like

customizable UIs too.

Brittany Ellich (03:09) Yes.

Erika (03:09) I I am also a firm believer that whatever your color scheme really does affect your mood or your like general approach to anything. So yeah, I I I like the ability to customize, depending on your mood or whatever.

Bethany (03:27) I agree. If I can put cappuccino on it, I’d probably love it.

Erika (03:30) Yeah.

Brittany Ellich (03:31) Fair, yes. I I’ve been a convert to cappuccini. So much so that now Claude just automatically applies that to everything that it creates for me. Cause it’s like, I know this is what you like. Like, wow, okay. Don’t even have to try. This is great.

Erika (03:44) Amazing.

Brittany Ellich (03:49) excellent. Okay, next one. Another another softball question. Mechanical keyboards, overrated or underrated?

Erika (03:58) I I I don’t know if this is even a question. Like I think they’re like like they’re they’re they’re awesome. So like whatever the rating is, it’s probably underrated because like everyone needs one.

Bethany (04:13) I’ll say I’ll say they’re rated, like just rated in the tech world, underrated in the broader sphere.

Erika (04:19) Hmm.

Bethany (04:21) because I think like outside of tech, not many people care about their typing experience. And one of my friends, love you, Emily, if you’re listening to this, she’s getting into like mechanical keyboards with gaming and stuff. And it’s it’s been really fun to see her like get super into like switches and and things.

So it’s I think like with outside of tech, it’s like not really a thing. But inside of tech, I think people are like very into it and like I love my keyboards. I clearly have a bunch over there. but I think some people are a little too into it. So that’s my take.

Brittany Ellich (04:53) Yeah, I feel like these are properly rated. I agree. They’re rated. I feel like the there’s a lot of hype, but that hype actually matches reality. So agreed that it’s more in the development world. I feel like developers are maybe we’re like more likely to customize our work experience since we spend so much time customizing our development experience. or maybe we’re just more picky. I’m not sure. But that could be it too.

But yeah, I like it. I it took me a long time to get into mechanical keyboards, which is another thing that Bethany, I feel like brought me into while while experiencing that through the last few years of knowing you. and I there’s no way I’m going back. Like that’s so beautiful and delightful. I just changed my switches the other day and I’m like, this is such a nice way to just like

Bethany (05:41) Wait, what did you change up to? We need to know. You still should post pics. I’d hope so if you have a

Brittany Ellich (05:42) So sorry, not switches, keycaps. See, I’m not I’m a fake I’m a fake mechanical keyboard person still. I still have the same switches. Yeah, but they’re like this pastelli.

Bethany (05:53) a rainy 75. Those switches are the most gorgeous like feeling.

Brittany Ellich (05:55) Yeah. They’re nice.

Yeah. But I got these like pastel keycaps and I’m like, it just brightens my mood every day.

Bethany (06:02) You post fix.

Brittany Ellich (06:05) well. okay. Standing desk.

Orated or.

Erika (06:11) So I think they’re underrated. I have had a standing desk for years and only really started using it like the last year or so. And I will say I I was very resistant to it. but like now.

I feel so much more energy. Like I’ll use it at the point when I’m like, I feel like I’m falling asleep at my desk. And then that’s usually the point where like I really don’t want to stand. But then if that’s when you do stand, you like end up feeling more energized. So it’s like a very counterintuitive experience, but I do think it helps a lot. again, I think like the health impacts are probably overrated. Like I don’t know, people have like talked about it.

saving your heart health or whatever and like I don’t know how much of that is actually true but from my personal experience I feel like they’re underrated because only because I’ve heard other people say the exact same thing where they’ve had a standing desk and they never use it. I’m like try it please try it like yeah even if it’s like 15 minutes of standing and then you go back down like it it can make a big difference.

Bethany (07:20) Alright, maybe I’ll stand right now to to do your tips.

Brittany Ellich (07:24) Thank you.

Bethany (07:27) I personally underrate it. I know I should stand more, and yet when I’m tired I sit on my my futon and that’s behind me instead. Which is probably not great. So I I completely agree that they are likely underrated and I underuse mine.

Brittany Ellich (07:46) Yeah, Erika, how do you like remember to use it? I feel like that’s my biggest better he’s gonna go stand now. I feel like I just never

Erika (07:53) Yeah, it’s it’s

Brittany Ellich (07:54) remembered to actually stand.

Erika (07:56) I mean, do you like do you have those moments where you’re like, I feel like I’m falling asleep? Yeah. So like that’s it’s like a body cue. It’s not like some kind of like timed interval or anything. Like it’s it’s

Brittany Ellich (08:01) Yeah. And then I go sit on my couch too. Mm-hmm.

Erika (08:08) really that moment when I’m like, I yeah, I just wanna like go lay down on the couch. That’s when I

Brittany Ellich (08:13) Hmm.

Erika (08:13) stand.

Brittany Ellich (08:14) That’s smart. Yeah, I wrote down underrated because they are genuinely good for ergonomics. the the the key though is it’s not just standing is good, it’s changing positions. this is my rare hey, I used to work in occupational health and safety and can talk about this this thing. but the thing that we talked about a lot in

classes that I took on ergonomics were the best position to be in is your next position. And so like it’s really just moving and like getting movement throughout the day. if you stand all day it’s also not great for you. Okay. Same with sitting all day. So

Bethany (08:46) Alright, so my take

is I should now go and sit. Now that I’ve standed. Yes!

Brittany Ellich (08:50) Yes, you’ve done it. You stood. Yeah.

Yep. yeah, but yeah, my I just never remember to actually use it. I should. I know I should, but I get sucked into a thing and then I need a I think I need like a little timer. I’ve been debating making like a little timer app on my computer to remind me. Like, hey, it’s time to go stand. And it’s so easy to make things now. I might as well do that.

excellent. Next one up. Vim or heavily customized editors. Underrated or overrated? Uh-huh.

Bethany (09:23) you know. You know what I’m gonna say. I

think it’s underrated. I think you should customize to your heart’s content. So

Erika (09:32) Yeah, I’m I feel like I’m gonna go on the like middle of the road, like rated thing here, because I feel like the people who really enjoy this are the people who have spent the time to learn all the key bindings, like spent that time. So for me, like somebody who has not used Ben before, it is a fairly steep learning curve and I’m constantly like

What is going to be the trade-off on the other side if I like if I spend all this time to learn Vim, like what will meaningfully change about my experience? And I’m not convinced yet that like it will meaningfully change my development experience to a point where I’m gonna like sink that time in. But I feel like for the people who do do that and already have the base knowledge, they love it. So I’m not here to like, you know, poo-poo anybody’s anybody’s joy.

Bethany (10:24) I I think that like no matter which side you are on, like whether you like customizing or not, like you will always get made fun of for it. So you should just do whatever you want. So I I support you, Erika, and not using Vim.

Brittany Ellich (10:38) Marie Kondo, your experience. Make sure that you are doing whatever sparks joy. Yeah, I like that. Yeah, I feel like I was on the side of hating this for a long time. just personally. I was like, I’m not gonna take the time to learn all those keybindings or customize things. And now that is not who I am anymore. People can change over time. I feel like I have spent a bit of time getting into my ghosty.

customization, my Z shell, after we had Robbie on, you know, all of those things. I think part of that is just like getting exposed to more people in tech that are doing things differently. So I agree. Rated. The people who are passionate about it are very passionate already. and I feel like that passion has earned. Not looking at anyone on this call in particular.

I appreciate you, Bethany. amazing. And I feel like this kind of goes with that. Dot files, underrated or overrated?

Erika (11:32) I’m gonna say underrated. I it took me a while to learn what dump files even are, and now that I know what they are, they’re super useful. I like I use them for like custom instructions, I use them for all sorts of things. and yeah, that that to me is like useful customization of like yeah, like my workspace doing a thing specifically.

Yeah, but it wasn’t something like I learned until I joined in of really.

Bethany (12:02) Yeah, no, that’s real. I completely agree. I think like like we talk about in programming, functions are what you use to make sure like things are reproducible. And I think dot files are like functions for your like configuration or your how your your editor. No matter how much or how little you like things configured, I’m sure people have like their preferences and just having like dot files and even just like throwing it in a repo so you can pull it on

whatever your new machine is, is like a huge time saver. And if you ever get a new machine, you don’t have to guess what your configuration was. You can just do there. And also if you change it in a way you don’t like, you can revert it easily. So I agree. Underrated. They are so, so useful. you can get as like deep or as shallow into it as you want, which is is great.

Brittany Ellich (12:47) Yeah, feel like these were more key when I worked at GitHub and was using code spaces all the time. too, especially if you’re doing some sort of like throwaway environment or ephemeral. Sounds nicer than throwaway environment, but something temporary, than like customizing that in a consistent way is super key.

And one of the I actually just got a new computer recently and the main reason I haven’t moved is because of like the pain that comes with setting up a new environment. So yeah, I agree. Underrated. Should use them more to save that pain. Awesome. All right, so now we’re gonna move a little bit into practices and processes. so the first one, pair programming. Underrated or overrated?

Erika (13:31) I think we’ve talked about this before, and I still love pairing because at the end of the day, like software is still a mental model, and no matter how it’s implemented, hearing somebody walk you through their thought process, whether or not it’s assisted by AI, is so useful because you get all those details of what they’re thinking, why they’re thinking it. If you don’t hear it from the person who

Wrote it, like you can guess, you can be like, well, I think that this is why you would do that, but you don’t know for sure. so yeah, I think like we’re not dealing with laws of physics here most of the time. Like we’re dealing with, you know, designs that people come up with. And so I still think pair programming, caring for people, talking through designs, talking through yeah, the mental model that they have is

Super important.

Bethany (14:27) I agree, super underrated. working with people is cool. Collaborating with people is cool and I think it’s like it’s still worthwhile. Like I guess an example is I I’ve been like working on a feature for the past couple weeks and I like made a PR stack for it. It’s like I’ve tested it, it works and everything. Didn’t pair with anyone, so now I’m stuck on asking for reviews and like now somebody has to like

either review that and like get it into their mental model or I have to like explain it, meet with them and explain it. So I think had I just paired more over the past two weeks it would have been a lot easier now and others would have gotten that experience as well with like kind of understanding the stack. So for myself I underrate it and I need to rate it higher.

Brittany Ellich (15:10) So I’m gonna sound like such a jerk saying this, but I think it’s actually overrated now, in my opinion. I know, I know. That’s a that’s a thing that I feel like was

I was strongly for pairing for a really long time, but I don’t know if it’s as important.

Erika (15:28) I I am personally

not surprised that you’re saying this for the record. Like this is exactly

Brittany Ellich (15:32) Okay.

Erika (15:33) what I thought you would say.

Brittany Ellich (15:35) gosh, okay, great. Okay. I know.

Bethany (15:37) Controversy Yes

Brittany Ellich (15:40) Controversy makes for makes for good TV, right? yeah, I I think that I was like a big fan of pair programming for a long time, but I don’t see the point anymore. because I have a new pair partner and that partner is AI tooling.

And I use that all the time. And so I feel like since I’m on a team with like a bunch of senior engineers, like if I was on a team with people who were just like getting up to speed in the organization or in software engineering, then maybe I could see a point. But I cannot remember when my last like good pair session was, either at my new job or even at GitHub. because it had been a long time, I think, before since I had like last actually needed a pair programming session to like work through a thing because

AI has like smoothed all the friction that I previously worked out while pairing with somebody. So yeah, I don’t know if it’s needed in this new world.

Erika (16:32) Yeah, I think it also depends on the thing that you’re working on. And if it’s something that’s really complicated and you have highly opinionated coworkers, pairing can be the s quickest way to get to consensus and get to the end result. If it’s something that’s a little lower like threshold, if it’s a little simpler and like everyone’s kind of vibing and approving, like I also I I

I guess I agree you, like I don’t see the point. so I guess it could it could be dependent on the project and the the work that you’re doing too.

Bethany (17:08) Yeah, it’s probably like really culture dependent as well. like I mean we were just talking about like async and remote like work or cultures with with Ben and I’m sure if we’re a more async culture, it’s probably more of a blocker than anything, and helpful to not force pairing as much. But yeah, I think it’s it’s very dependent on that.

Erika (17:29) Yeah, and also like we’re like you were talking about like the the makeup of your team and maybe if you had more junior engineers you’d maybe pair more. And I I definitely think that’s that’s important to recognize too. Like yeah, it it I know that I’ve gotten a lot from pairing

in my like earlier days as an engineer, whether it’s being forced to say your own thought process out loud or hearing somebody else say their thought process, it yeah, it it’s very valuable. I think because

And like the difference there is from like a mentoring perspective is that humans think differently than AI. Like you can ask an AI to explain its thought process, but it’s not something where I’m going to then adopt that same process myself and use it in the future. But with a human, you might add it to your toolbox and say, okay, next time I come across a problem like this, I’ll use maybe try this approach.

Brittany Ellich (18:33) Yeah. Yeah, I think it’s really crucial. I feel like that’s gonna be the thing that still survives going forward when people are learning how to become software engineers because it’s just like such a good way to

learn how to think basically, by seeing how other people do it. But yeah, I don’t know that I am as passionate of a pair programming shill as I used to be, which feels weird. But here we are.

all right, another another one that my my answer on this one’s probably gonna be obvious because I literally just wrote a blog post about it. But code review by your peers as a quality gate. Any overrated or underrated? Curious your thoughts.

Bethany (19:10) Okay. As a quality gate,

is interesting. I okay. I’ve been like thinking so much about this recently. and I’m excited to hear your thoughts on on this. But I I do think code review is still important. One for like we were saying, getting a mental model. If people have to support stuff, then they should understand what what’s happening or at least have a knowledge of it. And I think like should you should still review for like correctness, whether tests like are testing what they should, because AI still kinda

goes off the rails on those sometimes. however I think like for specific qualities like I prefer like less four or less nested checks or stuff or I prefer less like defensive guards and things those should always either be in like your your in s a high instructions file to like make sure you’re like at

An AI agent isn’t like doing those things or and in your like linting rules so that you can catch those. I don’t think like in today’s age you should be like reference or catching those on on PRs, to be honest. So that’s my take. I didn’t say overrated or underrated. I I think it’s I think it’s rated. I’m just gonna say that. But in different ways.

Brittany Ellich (20:24) Fair. Erika, do you have thoughts?

Erika (20:28) I think it is overrated as a quality gate. I think I have seen people rely on code review as a way to catch mistakes. And I think it’s up to you as the author to do your due diligence and make sure that your PR is as

good as possible. like that’s not the reviewer’s job to say like, hey, this is inaccurate or hey, this is misspelled. like that’s really your job as the author. So I think I’m with Bethany where like the most valuable code reviews are ones where you’re getting some kind of like

some kind of new information, like maybe there’s a tool or a library or something that you didn’t know existed and so like the code review the code reviewer might point you to that library. Or maybe you had some assumptions that

like it was supposed to work a certain way and it you know is not fulfilling that requirement. I do think that in that case, code review is a really late stage to understand that. And really that needs to happen earlier.

Brittany Ellich (21:38) I wrote a blog post about this fairly recently. and about like loop engineering, which is just like extreme I guess extreme programming already exists, but it’s like extreme agentic programming. with respect to

Like the two main things that I think are now very overrated is continuous deployment and code review, like the process of having another person review your code. and shortly afterwards, like a week ago, Steve Yeege actually wrote a post that had the same two arguments. So I feel pretty like pretty stoked about that, I think.

he also said that he spends like $87,000 a month on tokens, and then had a follow-up post about agent rights. So I don’t think that him and I are quite on the same page about everything. but I do think that there’s something to that. I think that like right now people are still doing human code review just purely because we’ve always done it. And I think at some point we’re not gonna need it anymore. Typically, like I feel like I’ve used that in the past to be on the same page as my coworkers.

And now I feel like that’s happening much further upstream where it’s I don’t even start the work until I anything that I’m not clear about has been figured out. And then I give it all to AI and just merge it. So I think that that’s my my theory is that that’s the future.

I’m gonna skip a few of these because I feel like We’re having really good conversations. is there anyone any other ones in the practices and process area that you think would be good to talk about or?

Erika (23:06) Hmm.

Bethany (23:07) we could do like a rapid fire.

Brittany Ellich (23:09) Ooh. Okay. yeah, let’s do that. All right.

Okay. all right. So next we’re gonna do rapid fire, the rest of these processes because we’re having such great conversations about this and we all have jobs too. So first one, stand-ups, overrated or underrated.

Erika (23:26) Overrated.

Bethany (23:28) Overrated.

Brittany Ellich (23:30) Heck yeah, I agree with that. Sprint retros.

Bethany (23:33) Underrated.

Erika (23:35) We don’t really have sprints, so rated?

Bethany (23:38) I agree with

the the sprint part overrated, the retro part underrated.

Erika (23:43) Yes, retros are great. I love retros.

Brittany Ellich (23:46) Yeah, I like not having sprints. I feel like I that was so stressful working somewhere where they’re like, all right, now you have two weeks to get this done and you never get it done in time. And then the thing at the end is always like, we pulled in too many things. but yeah, I agree. Retros are underrated. estimating software.

Erika (24:07) Okay, so estimates underrated, story points overrated. Like good, solid, well founded estimates, very useful. Story points or like guesstimates, not at all useful. Waste of time. Don’t do them.

Bethany (24:24) I completely agree. Same same boat. I think like like a lot of engineers hate on estimates, but like help your product peers out. I mean, that’s what helps them, that’s what helps the business. I think like it’s super underrated in the software world. so just like overestimate. That’s that’s my biggest tip. Just overestimate your your work and you’ll be fine. but yeah, I agree. Story points are like the amount of time I’ve spent trying to figure those out, waste of time.

Brittany Ellich (24:49) They don’t relate to time, but you know in your head you are calculating the amount of time every time. Everybody is. Yeah, I agree with that. one on ones with your manager. Overrated or underrated.

Erika (25:01) Depends on the manager.

Brittany Ellich (25:01) Maybe this is an awkward one. Yeah.

Depends on if any of our managers are listening, I guess.

Erika (25:07) Mm-hmm.

Bethany (25:08) I don’t have a manager right now, so that’s fine.

Brittany Ellich (25:10) There you go. There you go.

Erika (25:12) Huh.

Brittany Ellich (25:13) I feel like I’ve never done one-on-ones with my manager well. Like I feel like I think they’re hard to understand the point other than I think the point really is just to like chat and catch up and like have a regular touch point. And sometimes they’re really helpful when like you actually need something and it’s already on the calendar, but most of the time I never know what to talk about.

Erika (25:35) Yeah, I used them as like calibration sessions, but I have a really good manager right now, so she is very honest and like knows how to guide me in the right direction. But in the past I’m with you. It’s been like, how’s your day going? What are you doing this weekend?

Brittany Ellich (25:52) Mm.

Bethany (25:53) Yeah, I agree. They’re underrated if you have like a a really solid manager and overrated if you don’t. so yeah.

Brittany Ellich (26:01) Alright. Microservices. We’re moving on to architecture and tools. Microservices. Overrated or underrated.

Erika (26:08) Mmm, so I feel like microservices had a huge heyday and like I think the default of microservices

was overrated at some point. I think we’ve come back around where we’re like at a good place with it. I think microservices still do have their place and can be very useful. But yes, the the starting point of like everything needs to be its own service is definitely overrated.

Bethany (26:39) Completely agree. I think at one point they were overrated and then people realized that networking across services costs latency and and stuff and then I think we were right-sized with or or shifted to the right opinion now. So I think it’s like in the in the happy spot, the sweet spot for for me.

Brittany Ellich (26:58) Yeah, I agree with that. I think actually taking a look at what your software needs and what your like usage patterns are is rated. It’s a good thing. and making decisions appropriately based on that is the way to go. serverless. I feel like this also had a moment similar to microservices. It was really before microservices, if anything. and I feel like it’s sort of the same like pattern thing almost where

Yeah, that’s

Bethany (27:24) I think serverless

is like marketing speak for cloud companies, so I think it’s overrated.

Erika (27:29) I think it’s great when you don’t want to manage your own compute and you have something small and simple. I definitely think it’s overrated for anything that’s more than like a lambda, you know, like yeah.

Bethany (27:43) Yeah, I don’t know if you all had to do the same trainings with like, when I was doing AWS training and stuff, the amount of like serverless posturing, like, we made our entire stack serverless and it’s like, at what cost though? Like this is horrible.

Erika (27:57) Yeah. Yeah.

Brittany Ellich (28:00) Yeah, I feel like it is almost a way to get you like deeply into a cloud platform, really. and then it’s really hard to switch. So maybe that is, yeah, maybe it is just marketing posture. I think the only place it does make sense is if you really truly have something that is like not frequently needed and like a lambda is ideal. You know, like if you have a thing that runs and not predictably a couple times a month, then yeah, then a lambda’s great for that.

But other than that.

Bethany (28:30) Totally. Just write the HTTP server. You can you can buy a load balancer off the shelf and offset that, but yeah. Just write the server.

Brittany Ellich (28:38) Not right now, not in this economy.

Feature Flags Overrated or Undrated?

Erika (28:44) Who underrated always.

Bethany (28:47) I agree. Underrated. I love feature flags. A little too much.

Erika (28:50) Mm-hmm. Yeah.

Brittany Ellich (28:52) Hard agree. I feature flag my feature flags. I like I like using feature flags for everything. testing in production is always the way to go. I like doing that.

Erika (29:00) For sure.

Brittany Ellich (29:01) Software rewrites. I feel like this is another one that has some like extra nuance in in the time of AI computing.

Bethany (29:10) Okay. I know you wrote a blog post about this one too, which we should definitely link. I’m gonna disagree with your blog post though, and I say rewrites are underrated, especially with AI. Just throw agents at it and rewrite your whole stack and design the software you want. Like I think you should rewrite it.

Erika (29:27) Hmm, I think that hmm.

I think I’m gonna fall into the middle ground where I’m not taking a stance on this one. I might be leaning more towards the overrated side, but I have seen them be very useful in some scenarios. I think this is the distinction where if it’s sort of like ergonomic or like sort of

like a like a nice to have layer on top of what you’re doing. Like don’t bother. Like hopefully your time is spent doing something else. Spent better doing something else. and if not, I would question like it what if what you’re doing is really important. Unless it’s like your own personal project and you know it makes you happy and then you do you. But I think like

Sometimes rewrites are oversold to say rewriting this will solve all of our problems, and then you rewrite it and you still have the same like networking issues, performance issues. like you don’t you don’t ever solve the like real difficult problems that are ingrained in your software, and that’s really what makes the difference. So that’s my

That’s my that’s my take as far as like yeah, I’ve I’ve seen both sides where I’ve seen the like or like let’s rewrite it in a different language, you know? And even that, like, there can be huge benefits for rewriting something in a new language for from performance standpoints, you know, modernization, like security, like there can be huge benefits, but like as long as it’s

Done with explicit intent of taking advantage of that and not like, I decided today that I like Go versus Java and so I’m gonna rewrite everything in Go.

Brittany Ellich (31:21) Yeah, I feel like all the people who a couple years ago were arguing for rewriting everybody everything in Rust are now having like such a moment. Because actually it does kind of make sense to rewrite a lot of things in Rust now. I’ve been pretty staunchly opposed to rewrites in the past, especially on this on both online and on this podcast. but I feel like that was also like pre mythos, where now actually

Maybe it does make sense to like, you know, now that we’re seeing all of these security vulnerabilities and old software that’s been around forever, maybe there’s something to it. I don’t know. yeah, so I’m gonna go with rated. and just don’t go read that old blog post, I guess, because if it they used to be expensive. I mean, they used to be really expensive to do, and now it’s basically free or it’s like token cost, which is different than time.

Bethany (32:14) Absolutely. Then X

Brittany Ellich (32:16) Yeah.

Bethany (32:16) that don’t link it. Don’t even look for it.

Brittany Ellich (32:20) I won’t.

It doesn’t exist. yeah, opinions can change over time. yeah, there’s a lot of these that I feel like opinions have changed over time. okay, moving on. We have a few more that we’ll go through and then call it vibe coding overrated or underrated.

Erika (32:37) Man.

I’m like haunted

Bethany (32:38) Okay,

Erika (32:39) by the slot PRs I’ve seen recently. And I I

Bethany (32:44) Are we considering

like vibe coding equal to a agentic engineering? I don’t like the term vibe coding, but yeah.

Brittany Ellich (32:53) Agree. I think the term vibe coding is overrated and is overdone and at some point we need to just call it coding. But I feel like there is a difference between

Erika (33:00) Mm.

Brittany Ellich (33:01) like paying attention and not. And I think that vibe

Bethany (33:05) Yes.

Brittany Ellich (33:05) coding kind of covers both of those. Agentic engineering,

Bethany (33:08) Yes, you sh

Brittany Ellich (33:09) maybe that’s the opposite, but

Bethany (33:11) Yeah. You absolutely shouldn’t vibe code your work. You should like pay attention and don’t give people slap like Erica’s saying. Like, I don’t know. You should put out in the safe amount of effort you want the your reviewers to put in, like, for understanding your code. but yeah, I think agenc engineering definitely like it it is the future at this point, honestly.

Erika (33:35) I’m also trying to think of everyone overall versus engineers and whether society as a whole has overrated vibe coding or underrated it. I feel like people have probably overrated it because there is so much more to software than writing.

HTML and CSS or whatever people think like after vibe coding an app, they’re done with their app. For anyone listening, I’m using air quotes around around several of those terms. so yeah, I think I think as a whole, vibe coding has been overrated.

Brittany Ellich (34:20) certainly been overhyped, for sure.

Erika (34:22) Yeah.

Brittany Ellich (34:23) I do think it I yeah, I think it there’s like a few different futures people have predicted. There’s the one where like everybody becomes a software engineer and starts writing their own things to solve all of their problems. and I don’t know that everybody’s like that still requires like some knowledge of software patterns or you know like I don’t know that that’s everybody’s wants to solve all their problems.

with software

Erika (34:47) Right.

Brittany Ellich (34:48) also, you know, like that maybe people do will just write things down. yeah.

Bethany (34:50) Yeah. I’ve like

always had the mentality that everyone should know how to code similar to how everyone knows how to like do math, do like do do English. That doesn’t mean everyone’s a mathematician or a a writer, but like I do think everyone should be empowered to code because I do think it’s a helpful skill to have to like

Erika (35:07) Mm-hmm.

Bethany (35:08) automate. is everyone a software engineer? I don’t think so, but like I do think that

There’s a lot of gatekeeping around this, I feel. and like putting down people who don’t historically code or typically code, and I don’t think we should be doing that.

Erika (35:21) That’s true. And I’m not trying to do that. I’m yeah, I’m only saying like I think yeah, for a learning perspective, great. but for like production ready apps, I think what’s being sold is not reality and like yeah, i what’s being sold is is vibe coding, like vibe coding a production app or

Yeah. And I I’m I’m also with you where I feel like this is software engineers saying like, yeah, you can like vibecode your way and or like software adjacent people. And most people that I know who are not software engineers are probably not gonna do that. Like I yeah, I I very much doubt that they’re gonna spend their their time making an app.

Brittany Ellich (36:10) Turns out not everybody

Erika (36:11) Outside of like

Brittany Ellich (36:11) tries to solve all their problems with software.

Erika (36:14) Yeah. I mean I

Bethany (36:15) Yeah.

Erika (36:16) I have had really smart and thoughtful friends participate in sort of like vibe coding sessions. There was one that she did that was like, you know, it was kind of like one of these like vibe coding for good.

type of things and like she has a lot of experience in public policy and like has worked in government and she created something that like ended up getting like used and adopted, which is really cool. so I think in those scenarios it’s great, but I I feel like that’s more the exception rather than the general rule. I yeah.

Bethany (36:48) Yeah,

for the record I totally agree. Like in a professional context, y you should have like professional rigor, like with what you’re

Erika (36:54) Mm.

Bethany (36:55) pushing out there. which I think a lot of people do, but some people are just like, I’m just gonna throw these out there and someone else can deal with the complexity there and I do not think that’s right. So I I feel like I agree with you in in like a lot of the scenarios.

Brittany Ellich (37:09) Agree. Yeah. Yeah. I wonder when I’m very curious when that term is gonna die. Same with thing with like, it’s just AI slop. Like one day it’s just gonna be stuff something from AI. And that’s yeah, I don’t know. last one, 10x engineer.

Erika (37:28) So I think this is a myth. I’ve met engineers who I would describe as like two or three X. Like I’ve definitely had coworkers who I’m like, you are, yeah, two times to three times more productive than the average engineer. 10x? I’ve never seen that.

Bethany (37:47) I completely agree it’s it’s overrated. we should link it, but Sean Gadeki, who we interviewed before, recently wrote a great blog post on like how being an engineer isn’t about how smart you are. It’s about can you break down problems? Can you like slow down the problem so you can process out things? And I think that’s really the skills that you need in engineering. And I think that those are just you are like

You anyone can build those skills. Like not everyone is like just born, I know all this stuff, but or like knew what books to read, but anyone can like slow down a problem and break it down into parts and I think that’s really what what makes an engineer good, not whatever 10X means.

Brittany Ellich (38:27) Yeah, I think the only like take on 10x engineering that I’ve ever really liked is the one where you like are such a multiplier for the rest of your team that like your team as a whole has more output than they would otherwise. doing 10 times the output seems unrealistic, but then again, I have been experimenting with this loop engineering thing and am producing 10x at least the PRs. Granted.

Feel like I’m breaking PRs into smaller pieces than I would otherwise. but maybe maybe there’s something to that, but I don’t think that the people who coined the term thought that this was gonna be the outcome of what we were gonna be.

Bethany (39:07) I love I love that flipping it too, like of like it’s your

output and stuff like that, because that typically means you’re taking other work from the team if everybody else isn’t like putting out the same amount of work. So I think it roughly should

Brittany Ellich (39:17) Mm.

Bethany (39:18) be equivalent like for your level of how much you’re you’re putting out. So I think that’s a good nuance there.

Brittany Ellich (39:23) Mm-hmm. Yeah. Excellent. Well, thank you for participating in this. We don’t have a fun section today because this whole thing was fun. It was delightful.

Erika (39:32) That super fun.

Brittany Ellich (39:33) yeah, very interesting conversations here. I dig it. Excellent. Well, thank you, listeners, for tuning in to the Overcommitted Podcast. If you like what you hear, please do like, subscribe, follow, or whatever it is you’re supposed to do on the podcast app of your choice. It’s very helpful for us for visibility.

share it with your friends. Check us out on Blue Sky. And yeah, join us next week. Thank you so much. Bye.