70: Open and Async Work | Remote Engineering Culture with Ben Balter

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

Erika (00:08) Hey I’m Erika

Bethany (00:10) We met while working on a team at GitHub and quickly realized we were all obsessed with getting better at what we do. So we decided to start this podcast to share what we’ve learned. We’ll be talking about everything from leveling up your technical skills to navigating your professional development, all with the goal of creating a community where engineers can learn and connect.

Bethany (00:29) Today on Overcommitted, we’re joined by Ben Balter. Ben has spent more than a decade at GitHub, starting as its first government evangelist, bringing open source to the public sector, and most recently as director of hubber enablement, helping thousands of hubbers do their best remote work. Before GitHub, he was a presidential innovation fellow and part of the White House’s first agile development team. He’s distilled his experience into open and async.

His new playbook for building software on remote and distributed teams. Welcome Ben!

Ben Balter (01:02) Thanks for having me.

Bethany (01:03) so excited to chat with you. we’ll start off w like we start off every episode. what is one thing you’re currently building or obsessed with learning right now?

Ben Balter (01:13) Right now I’ve been obsessed with but learning, but actually for the past two and a half, three years or so, is how to write a book

not using a word processor. When I first sat down to write a book, you know, the default is you fire up Microsoft Word or LibreOffice or whatever. And it’s just was such a step backwards compared to what a developer experience is on a day, day-to-day basis, right? We spend so much time fine-tuning our craft. and so I said, what if we can take that purposefulness and bring it to prose, to writing a book? And so I am guilty of very, very much over-engineering the build pipeline for the book.

I figured it’s a book about software development best practices. so it’s built with software development best practices. Everything’s in Markdown. I have CI running on every push with thousands of tests. the print and the digital version are just built as web pages. it’s just CSS using Tailwind Tailwind CSS.

I figured, you know, I have I have decades of experience building web pages and to build something that exists in the physical world has been a really, really exciting learning adventure for me. And then the other side of that you said obsessed. I’ve been recently obsessed with the interplay between AI and writing. Right? In in software development, if you used you know, autocomplete and ghost text to write a the the the the types the type

Type signature for a function call or something, no one would think twice about that, right? That’s just something that you have to do, and it’s considered efficient. But when you do that same thing with prose writing, because writing is a form of expression, there’s a lot that goes into play there in terms of, you know, are you outsourcing your voice? Are you outsourcing your authenticity? and so I drew the line, I used AI heavily for editing, but not for writing. And so I had, I mentioned I have a bunch of tests running on CI. I had tests to make sure that the chapters were balanced.

That each chapter had call-outs for both ICs and managers, that paragraphs were the right lengths. I even ran I at one point I was convinced from reading the book so many times that I kept repeating myself. And so I had AI generate a n-gram analysis in order to determine if I use the same phrases multiple times across chapters. And so really using, you know.

probabilistic AI to build deterministic tools to aid the human in their own expression. So applying the best of software development to writing pros has been what I’ve been hacking on these days. I have a blog post in the works. I hope to open source all that of course and share with the world how I did that.

Bethany (03:45) I’m so interested to read more about that. I think that’s been such a hot issue with AI in general is when does like creativity end and how does that tap into it? Is it still yours if you use AI at all? So I am really excited to to hear more about your process and I think we all are are l looking to learn on the best practices as AI just changes so rapidly.

Ben Balter (04:14) Yeah, one of my favorite aha moments was I I I hate cliches, I’m allergic to cliches in my writing. and so I wanted to make sure that there were a CI job that made sure that there were no cliches anywhere in my writing. And I started coming up with clichés myself that I found myself writing. I said, wait a minute, you know who’s really good at coming up clichés? AI. and so I just had AI generate a list of its own cliches that it normally falls back to, and then built a CI around that. And so finding ways

to augment your voice rather than to replace it.

Bethany (04:47) Definitely. That makes a lot of sense. so you mentioned you talk a lot about open source in your book. You talk a ton about open source and how people work in in the open and and developing this through watching that process. I was shocked when I was reading the end, I think it was your author bio that

I didn’t even realize you you had a law degree and business degree and you didn’t necessarily come from C S. So I’m really curious, how did you find your way into open source and learning about this this field?

Ben Balter (05:23) Yeah, so I was sitting in the back row of law school. I was not paying attention to whatever the lesson of the day was. and I was really falling in love with open source. specifically, I was learning PHP, I was learning WordPress, I at the time I became a core contributor and did a Google Summer of Code project, found myself building websites for campus organizations, and the plan was still always to become a lawyer, like the traditional white shoe law firm path.

And I entered a really tough job market, not too dissimilar from the job market today. And a buddy of mine, who was not in tech at all, I had no plan on get on getting into tech, was like, hey, you do computer stuff on the side, like, why don’t you apply for a job doing computer stuff? And it just literally never occurred to me that I could turn my passion into a profession. and so the next day I applied for the job, a job at the FCC, the Federal Communications Commission, in there with

They

called at the time, this is how old I am, new media department, which was what the website was considered at the time. And just fell into this path of using government procurement in order to modernize government software. And I kept finding myself that a lot of the playbooks I was looking to was just saying, hey, how did the open source community solve this problem and how can we adapt that to government? And so originally it was using open data.

you know, publishing out the government’s data on GitHub as the Trojan horse to really pull governments into open source. and

what what really came down to me is that the the code was never the point, right? It was always about solving problems. Right? The code is the easy part. AI makes that even easier today. and so whether it’s legal code or computer code, it’s just a means to solving otherwise unsolved problems for collaboration. And so kind of fell into it by accident, but it’s been a really great great experience since and I wouldn’t do it any other way.

Bethany (07:33) I think that backgrounds are so important to contributing to software and the software field. Do you ever find you’re like

legal business brain showing up in your i in how you think about software or even how you think about open and async work.

Ben Balter (07:51) Yeah, so the first is is the fun flex that I get to do.

where I used to be in tr the trust and safety space, which overlaps with legal a lot and compliance and regulatory. And there’s a lot of FUD, a lot of fear, uncertainty, and doubt, a lot of lore about what you can and can’t do. and it’s a lot of fun to be the only lawyer in the meeting, and to be like, are you a lawyer? Because I am and I think we can do this. and in order to to get code across the line. but more but more seriously, constantly I think about it. for me and I actually work

Worried about this in the book that I over-indexed on this. Law is about writing. Law is writing first, it’s async, it’s precedence-based, right? Like one court makes a decision and then another court references it. That’s that decision record. It’s cited, it’s reviewable for years. That’s just git history as a legal institution. I think they’re they’re they’re really one and the same. And so thinking about how to apply that idea of proof, of jurisprudence, of having a a

canonical record that you can point to and applying that to obviously much less important decisions in terms of like ADR’s architectural decision records or why a pull request landed the way it did, I think they’re they’re one and the same.

Erika (09:08) Yeah, it’s interesting because even within the software development world, I have personally come up against people who either like want a more closed culture or like won’t put their thoughts into writing. and yeah, I’m kind of curious how you you approach that as somebody maybe wanting to affect a cultural change. like when is it appropriate to

Advocate for it and then how do you advocate for it as somebody who might want to introduce some of these more open practices?

Ben Balter (09:46) Yeah, great question. I see transparency has has two flavors or two cultures. the first is transparency as a liability. That if I write something down, you might hold that against me, you might judge me for that. You know, going back to the legal question, you see that a lot of like lawyers not wanting to put something in writing because it could show up in court. And there’s kind of the the court of corporate opinion, if you want to call it that, in terms of I wrote this thing down and now someone’s gonna think that I’m wrong or they’re gonna they’ll use it again.

Me in some way. The flip side of that is transparency as an asset, where you’re putting yourself out there, you’re building a name for yourself, you’re building a reputation for yourself, you’re helping others to opt in to join you and to collaborate with you, making sure that they’re not repeating themselves. and you’re working loudly so that others can see what you’re doing, and if there is

that potential liability, that potential conflict, you’re not catching a stakeholder by surprise at the end. They’re going to find out eventually. And so rather than doing a giant flip the switch kind of magic trick reveal, to be working along the way transparently so others can see what you’re doing, more often than not, they’re just gonna ignore you. They’re busy. They have their own things on their minds. but for that one time in ten when someone sees it and says, hey, I saw you’re working on X, have you considered why? Or I don’t think we can do that, catching those problems

earlier on where they’re easier to correct is the power of transparency and it’s just a matter of leading through example, showing the easier way to do things. If you can make it easier to work transparently, other people can see that. They can see the benefit of that. And you’re not going around with a top-down thou shalt work transparently fiat mandate, but just showing people the better way to do things and the benefit of doing things just through the way you work. And honestly just making them jealous, right?

Making them say like, hey, how are you able to do that? How’d you get there? and being able to point to the URL or the whatever and then they get that aha moment.

Erika (11:51) Yeah, I’ve definitely had both of those experiences where I’ve benefited from learning from engineers that I’ve worked with who are really good at this, of documenting their thought process and and critically documenting what is fully decided versus still kind of in draft too. Like sometimes it helps to get around that feeling of like I don’t want to put this down in writing to sort of like say it up front of like this is still a guess, like I

still not fully sure that this is true or like I’m still investigating, but like as of this point, this is what I know to be true, and this is kind of what I’ve what I’ve found. so yeah I’ve I’ve definitely benefited from learning from engineers who do that really well. and then also kind of what you said of of asking people to kind of

Do the things that you’re you’re hoping for too of hey, could I get that in writing or can I get an issue for that? Could I get a link for that? And just kind of like helping people build the habit through sort of a request and suggestion path. I haven’t as of yet had too much pushback on that, which is which is lucky. yeah.

Ben Balter (13:09) Yeah,

I I go back to early in GitHub’s history, before we had managers, GitHub’s engineering culture grew out of open source culture, right? And it it ran like an open source project before inner source was really a an established concept. And what I mean by that is I had an idea my first week, going back to government, it was about getting GeoJSON data, mapping data from government agencies onto GitHub. I bumped into someone at headquarters during my onboarding, we came up with the idea.

There wasn’t a you know product review session, there wasn’t a roadmap meeting. if you had an idea in early GitHub, you had to go around and you had to pitch it. You had to build a coalition. If you need a designer, you needed to convince a designer that it was something worth doing. If you needed a back-end engineer, whatever it was, a marketer. and so that idea of kind of building a coalition around ideas has really stuck with me even as GitHub grew, and we did have those more formal ceremonies around green lighting projects.

And to your point of hedging when something is early and setting expectations that it’s not a finished product yet. One of one habit that I picked up from a different leader at GitHub was when you’re in the early ideation stage, oftentimes that takes the place in a Google Doc or a O365 file. I would have a standard template for a header that I would put at the top of every file that I’d share around, my name, the visibility.

And the status of the document, and I’d say 99% of the time it was always drafting, regardless of what the actual status was. Because that just lowers the defense mechanisms of people that are reading it. They’re like, I’m getting it on the ground floor. I can shape this thing. And I would say, and don’t if any former colleagues are listening to this, sorry, it would stay at the drafting stage right until it was ready to be published. Just because that is the that was a hack to build a coalition. And kind of like getting people in an

alpha or a beta, they kind of have that identity of being a early adopter.

Bethany (15:16) That’s so cool. I think that’s a it I mean, part of that is also just building the safety to be wrong and to put out something that might not be finished. It just makes people more comfortable to also put out something that’s not finished or to give ideas that are half baked and say, maybe this is a thing, maybe it’s not, and contribute to those stocks. So I really love that that was so ingrained in the culture early on and it is something noticeable I I see even today with being able to

collaborate or feel free to just not be right all the time. even if it’s if it’s early on in in a project’s history. So that’s really cool.

Ben Balter (15:55) Yeah, l

later in my career I I really started to tune in and notice psychological safety and how important that was for a team. I was on one team that was the healthiest team I’ve ever been on, that had was very the ma engineering manager was great about creating psychological safety. and I tried to emulate that as I became a people manager later in my career. especially as I had teams that had people that didn’t look like me. for those of you that are listening in on podcast, I’m a white dude. and so that, you know.

It’s a lot easier for me to find psychological safety in tech. But as a manager, putting myself out there and being wrong, like there were times where I was kind of a little bit wrong, and I would purposely make sure the team knew that I was very wrong or that like I made a mistake so that they could see that it’s okay to make mistakes, it’s okay to be wrong, and be a little bit goofy about it too, right? Like let’s have fun with it. Like I used to joke, every time I came up with a bad idea, I’d say something like, you know, well, I have to have bad ideas first before I could have good ideas, and just

including that culture where you’re not judging each other but you’re yes ending and you’re supporting each other.

Bethany (17:02) That makes a lot of sense. I mean, speaking of this psychological safety and kind of going back to open source as a as a framework, is there anything about open source that just doesn’t really translate to a company format? or do did you find that it pretty much maps one-to-one with how corporations run?

Ben Balter (17:28) Yeah, I would say that successful companies take the philosophy of open source and the tools of open source, but it’s not a copy and paste, right? Open source doesn’t have strict deadlines, it doesn’t have, you know, conferences that you need to ship by, it doesn’t have investor reporting and c and quiet periods, there’s no payroll, there’s no HR,

And it’s largely consensus driven, right? There’s benevolent dictators, but it’s largely consensus driven. That’s not how companies work. Companies need to be able to make money. I’m reminded of the day that GitHub’s product org was introduced. So I mentioned that for a while we didn’t really have managers, everyone just kind of pitched their own thing and everyone just did their own thing. That was great for a while, but at one point we realized that we were building like seven or eight different startups. We had our own version of sales.

Force, we had our own version of Zendesk, we had our own version of like you name the startup, we built it. chat, we had our own chat app. and so the day that the

product org was introduced as a a a tastemaking function. We literally went through and and for those folks that are at GitHub you can find it, it’s I think issue number one in the product repo, and we listed out every single project that was in flight. And it was like 75 projects and we were pretty small even back then. and we used emoji, because of course we use emoji and either put like a ship it icon, a a snowflake icon, or a speedboat I forgot what the vocabulary was, but literally went through in the open

showing what

what we wanted what you know what we’re picking up to do what we are leaving behind and open source isn’t always good about doing that. I like to say that removing a feature is a feature and you don’t often see that in open source, right? You often see like everyone like if you want to build it, great, let’s just tack that onto the side. And there’s not that taste making sense. And I think where we see that in corporations, in business, is DRIs, directly responsible individuals, or as in OG getup culture we call them perps.

primary responsible person and I think that that DRI that can make decisions, that can be that tastemaker, they’re gonna beat a committee with a mandate any day because they have that that cross-cutting reach.

I think, yeah, and then the last part I’d just say is that you know there are limits to openness within a company, right? You have legal, you have HR, you have security, you have those private spaces that you don’t necessarily have those concerns with in an an open source project. So I would say you take from open source that async should be the default, but it’s not a dogma, right? Import the workflows that work, but don’t necessarily import the government’s model wholesale.

Erika (20:12) Yeah, it makes me think going back to communications of whatever you say only matters if the right person reads it. Like, you know, and I think one big challenge working for like a medium to large size corporation is finding that path of how to escalate or how to find the person who is your audience.

And I I haven’t worked much in open source, but I imagine like size of the group is a lot of what matters and sort of in the smaller groups it’s easier to propose something because you know there’s only so many people who are reading basically everything, but as you get larger people can’t pay attention to everything, so you have to find out how to

get your message to the right people.

Ben Balter (21:09) Yeah, in the book I talk about Dunbar’s number, which is a psychological idea that you can only remember about or have close connections with about 150 people. I think the inflection point for GitHub when we introduced managers, or technically leaders, they weren’t managers, was about 250. And I think that played out there of just not being able to keep those connections in your head anymore, and not being able to keep those one-on-one relationships. the one pro tip there that I used extensively when I entered a new code.

base

or a new part of the product was leaning into git and git history and git blame and issues and pull requests and it’s a lot

a lot more effective to be able to say the person that last changed this line of code that put this piece of copy on the website was this person. Find them in Slack, find what Slack channel they’re in, ask there, then ask in the like general engineering channel, hey, does anyone know who owns the copy on this button on this page? And so I use that a lot and a lot of digital splunking of going through from the blame to the issue to the or the blame to the PR to the issue to understand why something came to be, see who’s commenting on that issue.

and then you know kind of kind of stalking your colleagues almost through their past work so that you can you can outsource some of that remembering those connections or remembering who’s responsible for what to systems that naturally capture and expose that process just by how you work.

Bethany (22:37) I definitely use that a lot. And I feel like that was tips that came with onboarding to GitHub was that way of working, that way of just being able to break down those barriers and ask folks different questions. And I still remember so many anecdotes that you gave from your talk to like onboarding to our onboarding group. And so like it the we were talking about earlier emoji to convey

what your tone is. I used to never use emojis, now I use them constantly because it it really does help. like having a URL for everything and being able to get that to demonstrate what your what your work is. and so a lot of those are contribute to what you were talking about with like working loudly. And so I feel like there’s nuance between working loudly and maybe

bragging or self promotion or anything. And I’m curious, like, in your viewpoint, what is what is the difference between those or how do you aim towards the working loudly side if that is a there is a difference?

Ben Balter (23:46) Yeah, for me the difference is the direction of the benefit. if you are self-promoting, you are optimizing for yourself to be seen. You’re saying, Hey, look how great I am, look at this cool thing I did, I’m so awesome, please promote me. If you are working loudly,

You’re optimizing for the receiver. You’re optimizing for the work being receivable and findable. So you’re you’re narrating as you go in a durable, discoverable medium, right? you know, at GitHub we had had the practice of writing up discussion posts after a big ship about what you learned. And I I think for the most part, people saw those as good learning opportunities of sharing information. And it’s not because and this goes back even before when we had team, which was our internal internet, we had the same practice.

Us there. It wasn’t of, hey, look at this cool thing I did, I’m so awesome. But here are the mistakes I made, and here are the lessons that I learned along the way, so that you don’t have to make those same mistakes that I did. The idea is that others can can build on it, they can catch problems early, they’re not resolving the same problem over and over again. open sourcing the idea.

I toiled through this that you shouldn’t have to. Not I toiled through this, look how great I am.

Bethany (25:05) That makes sense. And I mean it does working loudly does have the same self promo benefits in the way for getting your career development and getting promotions. And there definitely isn’t anything wrong about that, but it there is definitely that nuance between, am I talking only am I self focused or am I about the team? So I I really enjoy hearing about those those nuances.

Ben Balter (25:31) Yeah, well mean

once once everything’s captured, it’s really easy to turn that into a brag doc or to turn that into your annual review. for years when review season came around, I would go to GitHub.com slash polls or github.com slash issues and just do a search for everything that I touched in that review period. I’d go through and click through it and build my brag doc off of that. As AI has emerged, I started automating that. hovers, you can find Reflect GPT, hopefully that’s still working someplace. but the idea behind that is take those receipts.

That you’ve been generating over the course of the review period and have AI just make you look amazing. And I think that’s really the power of, you know, I talk about how working transparently wasn’t intended to support AI, but it’s a welcome side effect that once you have all that context captured in a machine readable format, you can put it out in different outputs.

Bethany (26:26) Yeah, huge fan of Reflect GPT. That was what did my reflection this year and it was amazing because it’s it’s hard to talk about yourself. I

Ben Balter (26:36) I love if if

you’re if you’re an a AI nerd, take a look at the prompt. It’s like make me look amazing or something like that. I had I had fun with it.

Bethany (26:44) That’s amazing. I’ll have to take a look because it’s it it did a great job without sounding overly bragging, but it linked to everything. I was like, man, this is definitely happening next period. So big fan,

Ben Balter (26:58) And

Bethany (26:58) big fan.

Ben Balter (26:59) I I don’t know if you’re around for Thomas, the former CEO of GitHub, had a very funny line right aro right as AI was rolling out around reflecting period, about all we’re doing is you’re writing bullet points, having AI put it into prose, and then the manager on the other side is taking that prose and putting it back into bullet points. So why don’t we just send the bullet points across the line to each other and take the AI out of the equation, which I thought was was really funny.

Bethany (27:25) That is funny. I know we your your book often talked about that async first op is more equitable and levels the playing field for for different folks that might have other situations going on. and like folks that don’t live in the same country and are are distributed. and

You see in the news every day different return to work or return to office mandates, companies installing bossware and surveillance software. So I’m just curious, like is there any merit to that? Does async is that truly just it’s good for everybody and these companies are maybe afraid of that, or is is there are genuine situations where this might not be a workable solution?

Ben Balter (28:22) Yeah, I was joking with a friend yesterday, that that timing is a skill.

And coming out with a book about remote work just as everyone’s returning to office is just perfect timing. yeah, so I I I for me remote work is inclusive work. if you’re introverted, if you’re a non-native speaker and you work better in writing, if you’re a caregiver, if you’re in a different time zone, neurodiverse, or just like if you need a second to think before responding, I think that’s for me is why I have really fallen in love with remote work, been working remotely for well over a year.

decade now.

when it fails people is when it’s a veneer. When you say you’re remote, but you’re really having those conversations in hallways or conference room or the zoom after the zoom where the decisions are really made. That that thread, that issue is just theater and maybe just capturing after the fact. So someone that’s not in the office is not on equal footing. you mentioned bossware and surveillance, and I laughed a little bit. I know as as we’re exchanging emails, you shared an article about

Microsoft Teams and Green Dot surveillance and being able to like tell what Wi Fi hotspot.

The individual was on so they’d know like where in the office they were. The reason I left is because one of my favorite features when we when GitHub moved into HQ 2.0, the current one on Kali P. Kelly Street, we actually had exactly that. On every floor there were giant monitors, and it told you who was in the office, what floor they were on, what room they were in, and we even had a heat map so you can see where people were congregating. and it was one of my favorite things because I

As

a visiting hub, or was like, I really want to go talk to X. I have no idea where their desk is. I have no idea where they’re hanging out. I have no idea where the coffee, the water cooler conversations are happening. and so and this was like we had, if I remember correctly, we had motion sensors in the lights. Like it was very, very over-engineered. but the whole point of that was not for managers to be able to see where you were and make sure you’re in the office. It was to help remote folks create those opportunities for serendipity.

Right. The difference wasn’t monitoring, it was trust. And I think that that trust, when you see that, is in in remote work is moving from input, meaning you know, time at my desk, presence, keystrokes, the little green dot in Slack, which is more presence theater than productivity, to thinking about the output of my work, right? Issues, PRs, decision, meeting OKRs, objectives and key results, goals.

when I as a people manager I said explicitly to my team, We have explicit goals. If you can hit your goal in in thirty hours a week, great.

Right? Like the you’re not being paid to sit at your desk. If you’re not being productive, you should go for a walk, you should put together that IKEA furniture, whatever that’s been on your mind. I regularly do laundry or dishes during the day. that it’s it’s not about sitting at your desk. You know, it’s if you think about the evolution from the assembly line, where if you were building a widget, you had to be at the assembly line to build that widget. As a knowledge worker, your best idea can come in the shower, it can come at three AM. I get some of my best ideas while walking the dog.

And so the idea is to manage against not bossware or presence, but trust that people are doing the work that they’re supposed to do and measure them against the outcomes that you agree upon.

Erika (31:59) Yeah, I I say this as somebody who does love remote work and working remotely, but I do find that there are things that can’t be replicated more on that personal connection side. I still can’t stand the sort of like the Zoom social meetings and you know everyone getting on a call and having fun. You know, I it’s like it’s just totally different. I I

I don’t know if I’ll ever get to the point where I really enjoy those kinds of like group, group call experiences, unless it’s like a brainstorming meeting or something business related. yeah, and I I do feel like having a personal connection is is important for that psychological safety, because there’s so much that

That can get obscured in a digital world. Like, yes, you know, remote first work can be more inclusive, but there can be a lot that gets lost with a screen in front of you, whether that’s body language or you know, what whatever it might be. like there are things that we kind of pick up on as humans in those in-person interactions that

just don’t translate over over a digital landscape. yeah, so so it’s tough. I I think kind of finding that balance, but I guess to your point, the perception side of it does help a lot and giving everyone the benefit of the doubt, you know, like when somebody doesn’t respond the way that you might really expect or want them to in an async culture, taking a second

Thinking like, is this personal? Am I taking this personally? like, was it meant that way? You know, how did this person mean it? And and even if you don’t know them or you don’t have that personal connection, starting from a place of kind of like forgiveness first before jumping into that sort of like accusatory standpoint, I find I find can be helpful.

Ben Balter (34:18) I’m right there with you. I am anti mandatory fun time. I am anti cringe icebreakers. I you know, like sometimes it has its place. During the pandemic there were lots of Jackbox games and Among Us and like that was that was the right thing for the time. I I remember early in my career I would see I’m I’m not a t

say not a social person. I’m not an extrovert. I I would see very extroverted people that would fill up their calendar with just social calls and I judged them.

low key of like you’re wasting company time just like you know just making small talk and and like why do you actually get some work done you know like the rest of us are working and at one point I realized that I had that backwards that in a remote environment that that work is work making those connections that you want to affirmatively make those social bonds before you need them so that interactions are not transactional. It’s not like hey can you approve this PR and like that’s the only interaction you ever have.

Because that’s you know, remote work can very be very lonely and very isolating. but getting to know your coworkers as humans, right? In the office, you’ll see them leaving early for t-ball practice, or you’ll see a picture on their desk, or you’ll see that they have fur on their black shirt, so you know that they’re a pet person, you don’t get that necessarily unless you know a four-legged or two-something, two-legged something walks into frame. You don’t get that connection. And so going out of your way to make regular social calls and find ways to connect.

with

people outside of work, right? So at GitHub we had social channels, we had LiveMoss for Taco Bell, Paloton, Dogcom, CatOps, and you know, as I would say in onboarding,

If your boss sees you e posting a picture of your dog during the workday, you’re not gonna get in trouble. In fact, they’ll probably like you emoji react to that or something like that. That it that is part of your job and finding those those connections. one co-cower, I forgot how we got into this, we bonded over the movie and musical Mean Girls. And so we’d have like Wednesday wears pink days and we’d joke about that. another coworker, still mad at them for this, they introduced me to the video game Bellatro.

which is very, very, very addictive. I was actually just playing it last night. And so we bond over that. And that had nothing to do with work. But then and this came up multiple times when a work thing did come up, it’s like go back to the Mean Girls one, hey bestie, you know, like we already had that connection where we can we can have a serious conversation with that sense of trust and giving each other the benefit of the doubt and knowing that we’re both coming from a good place.

Erika (36:59) Yeah, yeah, I’m always I’m always thinking about how my communication comes across to other people too and and I I would love to have like as as I’m talking like as we’re I’m reading through the book and we’re talking about this like I’ve been thinking about how to have some feedback channels of you know is my is my communication

understandable, especially when it’s like AI assisted. I’ve I’ve gotten feedback from people saying this is really verbose, like this is, you know, kind of unreadable. And and yeah, I mean that’s a that’s kind of the the most feedback I’ve gotten recently about my communication style. But yeah, I have people on my team who are not native English speakers and I I don’t think I’d I’d ever even thought of asking them how

how clear my communication is to them until kind of thinking through all this of yeah, like does it does it make sense to somebody who’s not me?

Ben Balter (38:11) Yeah, I would take that in in in two directions. The first is if you think about how AI works, is it is just the mean of communication, right? It is just the average of everyone’s communication. It probabilistically determines like this is what is most common for someone to say after this word. And so

If you outsourced your your voice, like you you you sound like everyone, but you also sound like no one. And so in in async work, your writing is your persona, is your presence, is the primary way that your coworkers get to know you. And so yeah, you can tell, especially from like corporate press releases, executives, like when something is just corporate jargon AI slop versus when someone actually put thought into something. And then the second

Direction that I would take that. And so spend a lot of time, you know, manufacturing, not manufacturing, curating your voice. I have a gist that I keep that has very specific guidance for my own voice. Like I actually like M-dashes. I’ve been using them for before they were uncool. I like using technical metaphors. I you know like getting right to the point, and so I keep this gist with me so that anytime I’m in an AI context and I need to punch something up, I can say, hey, grab this gist, this is my voice, put this in my voice so people know it’s uniquely.

me. And then the other side is longtime hover John Magic has a great blog post about not outsourcing your thinking. His blog post was in the was in the area of snippets, right, of weekly reporting. And I fell down this same path of, you know, going back to once everything’s in GitHub, you can throw AI on it and get different outputs. And I did that for a few weeks for my team’s weekly reporting where I would just throw AI at it and give my boss here’s here’s everything that happened this week.

and that was doing a disservice to myself and to my team because the whole point of of going through the weekly review process is not to create the artifact that you pass up, but to actually do the thinking and to do the reflecting. And so if you outsource that thinking, you’re outsourcing yourself and you can’t show up for your team in the same way of versus when I say, okay, I’m just gonna have a deterministic script pull all the activity for the week and spin it out as a markdown file.

verbatim and then I go through and curate that, I see start seeing those connections of hey, I need to jump in and help this person, or hey, they’re falling off track, or hey, they’re doing a great job. They need to shout out. And so thinking about what you outsource, which, you know, the boring stuff, like going back to that type signature, like what’s the writing version of a type signature that you might outsource in terms of taking something that’s written at one level and writing it for a different level or a different format. or looking at you know checking pros, checking

tense checking, you know, making sure that there aren’t spelling mistakes. but definitely you don’t want to j to be delegating judgment, voice, taste, decisions, those still need to be remaining uniquely, uniquely human. I actually noticed that it was interesting, I noticed that a few times. so I used AI very heavily, as I mentioned, not for the writing, but for all the stuff that comes around it. So like I built my marketing site for the book, and there were a number of times where I noticed that the again going back to the the

AI gets you to the mean, it does what everyone else does. And the playbook for publishing a book is very writer-centric, right? It’s I need you to write a review, I need you to read this. And you know, the value where humans can can really add is not doing what everyone else does, right? That produces boring outputs, but thinking, like, why what what actually are we trying to solve here, and how can we come up with a novel way to solve that? And so making things in this case more reader-centric was going against the normal.

Going against the AI, but was a sense of taste and judgment that AI couldn’t bring to the equation.

Bethany (42:06) That’s I mean, you’ve talked a lot about your your process during this conversation and I I really enjoy especially hearing about the the writing process, like where you use AI, where you don’t. I’m curious, like maybe we can take a step back and would love to know what you consider the async first stack to put it in like programmer terms, like the stack that you think is best or

most conducive for async work. and is there anything that breaks if like not everybody is fully invested in that?

Ben Balter (42:43) Yeah, so I think chat in in Slack or Microsoft Teams is great for ephemeral communication, right? Or or coordination, right? It’s that incident, it’s the you know, we’re we’re shipping something, we need to get this out the door. but I also think that in many places chat is where information goes to die, right? In in the book I use the metaphor of a camera roll versus like an Instagram post, right? That chat just has everything. It’s a stream of conscience, there’s no expectation.

Of polished, there are typos, information gets outdated, that’s the camera role, right? Versus like that Instagram post that like you curate, the influencer gets the right angle and the lighting, and it’s designed for the consumer, right? And it’s the difference between: am I writing this for me to get it out there, or am I writing this for the future receiver or long tail of recipients to be able to read it? So chat is for coordination, short-term, near-term. The next level of that stack would be issues or discussions.

I think that’s where the actual work, the actual decisions, the actual knowledge is captured. Those are long-lived. And there’s a expectation that you distill your thoughts into writing, into something that is long-lived, and that it’s updated over time, right? If something’s no longer up to date, you post a comment on it, you update the issue, whatever it might be. So that’s where the durable knowledge happens. and then meetings. I don’t think meetings are a dirty word in the async world. I think they get a bad rap because

A lot of corporate culture defaults to meetings. It’s hey, let’s just jump on a quick sink to to kick off this idea. And like the the meeting being the starting point of any effort versus the way I think about it as an escalation point.

you escalate either because of the complexity of the situation, if it’s, you know, an ongoing incident, you know, joining joining the war room bridge to to talk through in real time, and also the urgency of the situation, right? If if prod is down, you probably don’t want to wait for an issue and like that perfect writing. You want to get the information across as quickly as possible. You ask what breaks. for me, what breaks is decisions.

Right? You see this a lot if teams work exclusively in chat. It’s really hard to define what was decided. I like to say that a decision without a URL, going back to the law analogy, a decision without a URL gets relitigated forever. Right? You keep going back to what did we decide here? Why did we decide that? And you keep you catch this churn. And I think the value of async, even though it might feel slower in the moment, you get better decisions faster.

Faster because you take time up front. You shift left in your decision making to get rid of all that ambiguity, to chase down all those edge cases, to get s an our durable artifact so that you’re not going back and revisiting decisions again and again because you thought you’re on the same page, but because of that informality of chat, there was there was some wiggle room there.

Bethany (45:49) Man, who among us has not tried to search Slack against the cutoff for for context on a crumb of context on decisions that were made. So that is very real. I think that durability is so important and very very missing a lot of times and it’s so team by team dependent. So I think you you hit the nail on the head of that decisions

needing that documentation and that durability makes so much sense.

Ben Balter (46:20) and

a quick pro tip there, if you have the Microsoft app

that’s Microsoft, the GitHub app installed in either Slack or Microsoft Teams, you can right-click on any message and it’ll automatically open up an issue. And I don’t it’s there’s like an open and issue button. and then I don’t know the current state of things, but if Copilot is in Slack or Teams, you can also ask it directly from the thread to open up a pull request to capture what’s there in documentation. the number of times when I was responsible for GitHub’s intranet, the hub, and

And someone would come in with a feature request, and in that thread, I’d just be like, hey co-pilot, can you just go do this? was a fun, fun flex in terms of being user-centric, but also pulling things out of Slack and capturing it and you know discussing things in in demos, not memos, of like, hey, is this what you were talking about? Because like this feature is now in preview, you can take a look at this, let me know if this is what you wanted. So, pro tip: try to use tools if you are using like chat is great, you should use chat, but use tools to

capture that context outside of chat when the time’s right.

Bethany (47:27) that’s so cool. I had no idea about the right clicking. I knew you could tag it and stuff, so I am definitely going to use that heavily now. Yes. Yeah.

Ben Balter (47:34) Yeah, and and GitHub has specific tooling called Xref that will also

cross reference across, but the native GitHub app it’ll pop up a little modal and I’ll like have the contacts and you can like issue, title, body, everything, and it’ll open the issue for you.

Bethany (47:49) XRF is my best friend unless I’m in a private channel. and then it’s like, you can’t do this from a private channel, silly. And I’m like, So I I love that you can just right click and and create an issue from

Ben Balter (48:02) Yeah.

Bethany (48:02) that. Amazing. so I know we’ve talked a lot about AI and like writing, outsourcing, thinking. there’s a lot of commentary around slop that is thrown around nowadays, that

typically references the AI generation, whether it be PRs, issues, and whatnot. so w in this world where people can kind of be loud for free or like outsource their thinking, which is kind of what you were hitting on, where can managers and engineers kind of find that signal through through the noise? Or is that even a problem? Is that even something we have to worry about?

Ben Balter (48:45) Yeah, you y you mentioned work loudly, which is a a phrase that I I use frequently, and it might be the wrong phrase these days, because it’s not about volume, right? It’s not about how loudly you can speak. It’s the whole point of async is that the loudest or fastest to unmute isn’t the one that gets heard. AI can generate artifacts, AI can generate more content than humans can consume, but it can’t generate your judgment. And so, you know, going back to the idea of of weekly reporting, the the

Signal just moves up a layer in the stack. So rather than being this activity happened, it’s about curation, it’s about taste, it’s about understanding what matters and putting it in context that the AI

doesn’t have the ability to do so because it doesn’t either it physically doesn’t have that context or it doesn’t have the the the the knowledge. The the and so I think trust the artifact still survives in a world in which AI is creating artifacts because you know they’re checkable in a way that vibes never were. It’s not like I’m at my desk, you still have that artifact. but for managers it’s more about judging the outcome, not just like you know, like w

We the industry went through that time where number of PRs was a metric and then we went through that time where number of tokens consumed were a metric. And so moving beyond those vanity metrics and looking actually at the content of the artifact to what there what actually is there, the substance of what’s there, rather than just the fact that the artifact exists. So quality over quantity.

Bethany (50:23) That makes a lot of sense. And like a lot of that I think is the perspective of the writer, which is great. do you have any suggestions for anyone who feels overwhelmed by the amount of artifacts that are coming through or how they can sift through that? is that something that they can also use AI for or do you have any tips for that?

Ben Balter (50:42) Yeah, one one issue that I kept coming back to every time we made a like communication architecture decision at GitHub was the concept of tragedy of the commons, which is a economic philosophy. The original one is like imagine a field in the middle of a village and everyone has I think if it was sheep or cows or whatever it is, right? Like any individual is incentivized to let their sheep graze on the the commons in the middle of the square, but if everyone does that, then there’s no

I don’t know what sheep eat, grass left for sheep to eat. and so you you see that same thing with individuals that optimized locally for their own communication, for their own need to be heard, right? A lot of channels, whether they be the engineering channel, the general channel, the engineering repo, whatever it might be, are a commons, right? And if the individual optimizes for their own needs, you’re gonna run out of grass. Well, this metaphor is going nowhere. Okay, we’ll back out of.

this one. but y you get the point, right? You need to think about if everyone d does this, like how how does I guess it’s a matter of of systems thinking, right? Like if everyone does this, how does that affect the system?

and so for individuals that’s just curating ruthlessly, right? Like knowing, yes, in an open environment, you have access to all this information, but what do you actually need, right? Like what is what is essential for your job versus what is just interesting? and then also what can you batch and curate, right? Like what rather than getting the fire hose of notifications as soon as they open up, like what save search can you have that you run Friday at four o’clock?

needing to to to constantly be online, right? Like I even even day-to-day very tactically, I wouldn’t keep Slack open most of the time. I wouldn’t keep issues open most of the time. I would finish a task and come to an actual stopping point or hit writer’s block and just need a break and then open up Slack and batch a bunch of responses. Open up issues and batch a bunch of responses. And so just being conscious of your own information diet that you’re not just kind of the metaphor comes full circle now. Not just great

Praising information throughout the day, but being purposeful about having meals and watching what you eat. Okay, I’m not gonna continue this metaphor anymore. The last thing I will say as a pro tip, and you might remember this from onboarding, you talked about the fire hose of information. If you’re making the transition to an open or asynchronous organization, you will be hit with a fire hose of links. I used to joke that URLs are Hubbers’ love language. They should it’s how they show you that they care. And so as we

You are onboarding or adapting to an open async environment, you’re gonna feel overwhelmed at some point. You’re gonna think that the company made a mistake, that you made a mistake, you might get imposter syndrome. and the advice I always give there is don’t listen to anyone that didn’t say that happened to them. That happens to everyone as they transfer to or transition to an async environment. If they say it didn’t happen to them, they’re a dirty stinking liar. Do not listen to a word they’re saying. Find someone find a different mentor, find someone else. the company didn’t make a mistake.

It’s just a natural part of evolving to that fire hose information and you will recalibrate to understand how to process that and get a sense for what you need and what you don’t.

Bethany (54:16) That makes a lot of sense. And I love the metaphor of tragedy of the commons. I was trying to rack my brain where I’ve heard that and it was from Thinking and Systems, which you were mentioning th systems thinking. So I think it is absolutely relevant in so many layers. So that’s that’s really awesome.

Awesome. Well, we are coming to the end of our main segment. and just like a personal anecdote, but this book really came at a at a great time. Like I a lot of times I get in my head about staying on Slack or that green dot and things for myself, not because it’s being tracked, but ‘cause I I’m like, I need to rank this up. But like this week my my dad had opened heart surgery and he’s been recovering in the hospital. So I’ve been back and forth between home and the hospital, like

keeping him company and stuff. And it’s been a good reminder. Like I’ve set my Slack status. I’ve been like, I am not attending meetings this week. I am going to be away. I’m just async only. And so many people have reached out and been like, hey, I hope everything’s okay. And respond at your leisure. But but like take your time. And so it’s just been a really good reminder, like a lot of these practices on how to stay visible while in the situation where I’m not necessarily in meetings or be able to hop on a call at a moment’s notice. So really appreciate that.

Ben Balter (55:33) Yeah, that’s awesome to hear. You you reminded me

of a a happy little accident that happened. so over the the lap most recent holidays, I was not going anywhere, so I was working and mostly just hacking on open source and just having fun. but I didn’t want my team to see that I was online and think that then the reverse of the green dot the reverse green dot of like them seeing my green dot so think that they need a green dot. And so I set myself to a way in Slack. So I did I even though I was on Slack, no one could tell that I was. I was in like incognito mode or whatever, ghost mode.

I forgot to turn it off for like months, literally months. And it wasn’t until like May or something when I realized I still had it on, no one said a single thing. Right? Like that it just it just is really a testament to the culture that, you know, you have I have a friend that asked to borrow my Flipper Zero to use as a mark mouse jiggler to get their green dot so they can go out and have a mental health afternoon. by the way, you can expense it, highly recommend it. That’s a lot of fun to play with. versus us where like I accidentally had my statuses off.

for months and and no one said anything. So I think that’s that’s really the testament of asynchronous culture that it can really adapt to your life if you can create that le that level of trust with your coworkers.

Bethany (56:46) 100%. Trust is often paramount there and it’s it is great to work at a place that has that trust and with people you you can lean on and stuff. So awesome. Very very cool. okay, so moving on to our fun segment. I was proud of this one. so you know MTV cribs, like where they’re showing their their house and stuff and where they live. so I

thought we could do overcommitted cribs. so I just really want to talk through like if you have your like current or ideal remote workspace, if you can describe that and then maybe if I’m fancy enough at editing I can edit in some images of of those in post. but I would just love to hear about it. I’m happy to go first to give give some time to think but when I was

creating my work from home space, my my office. I I was trying to think of what’s my like where have I felt the most productive in my life? And the answer was my the college library where I I went to college and it was like a very seventies library. Like it was clearly built in the seventies. It went underground. It doesn’t sound like a very pleasant place, but it was it was very nice and cozy. And so I was like thinking, okay, I need like wood and like

Yeah, books everywhere. So that’s how I kind of came to my my current office is lots of wood, lots of panel like wood accents and then my bookcase and stuff with a little bit of whimsy and like fun things thrown in to to to bring out my personality. But that’s that’s kind of what I I really liked. But I have a lot of tooling around my desk that I don’t necessarily use, so I wish I I had things that I need to figure that out.

that out next. Like what is my happy space with like do I need a Stream Deck? Do I need this like terminal E Ink thing right there? I don’t know, but just need to figure that part out.

Erika (58:51) My like current

My my current work setup, I really only like need two things and it’s a standy desk standing desk and a natural source of light.

that gets me through the day most of the times. But when we’re talking cribs and the most ridiculous thing you could think of that you could have, I would definitely have an on-call masseuse just like standing by. So like,

Ben Balter (59:21) Ha ha ha.

Erika (59:22) you the like pomodoro breaks, like I get like a little neck massage, like, you know, every every couple of hours. And I I think that would really yeah, that would really amp my productivity up.

Bethany (59:34) Sabe, can we firma that?

Ben Balter (59:37) You can. At least you used

to be able to.

Bethany (59:42) Erica. Forma is our our we can expense things on Forma, so

Erika (59:44) Yeah, I think I think I would blow my form of budget in like two

days if I hired a personal masseuse. So Yeah, to

Bethany (59:51) But think of the productivity.

Erika (59:53) work two days a year. There you go.

Ben Balter (59:58) Yeah, I I know if there still is. There used to be a Massus at GitHub headquarters that you could sign up for if you were visiting.

Bethany (1:00:05) Next time I visit I’m checking because that sounds amazing.

Ben Balter (1:00:09) Yeah, I’m a big fan of the standing desk. I’m a relatively recent convert to standing desks. I take every meeting standing.

and I can’t imagine doing it any other way. It just feels a lot more natural to me, especially if you’re presenting. I used to be on the speaking circuit when I was doing what we now call DevRel, and so I’m used to being on stage and walking around if you can’t if you’re watching the video, moving around a lot. and so the standing desk has been a big part of my work from home setup. if you’re looking for a project, I haven’t implemented it yet, but I found on like Kickstarter or whatever, there’s a product called the Upsy Desky. Up ski deski? It is

Like an ESP32-based IoT device that plugs into your standing desk and you can then control it using automation. It like just plugs into the Ethernet port of the standing desk. It like a man in the middle is it essentially. And so my goal is to have it sync with my calendar so that as soon as a meeting is about to start, that the standing desk just starts rising up and gets gets in place automatically. The other big one that has really changed the way that I work, and I’m using it right now, I actually have a teleprompter.

And I’m a huge fan of it. A coworker had one and I was like, wow, your your camera’s great and your you know, I thought I would use it for like a teleprompter, like sc streaming notes and presentations. I have the El Gato teleprompter and it just shows up as a third monitor.

so you can just drag screens on it. So right now I just have the recording screen on it right now. I can see all of you. I really like it because it’s a lot more natural for making eye contact. and then the other benefit there is I actually use my iPhone as my camera. I have tried so many different webcams and could not find a webcam that I liked. and so the idea of just like every time I upgrade my phone, I get a new webcam with the best camera in the world has been has been a really great quality of life improvement for me. walking path.

I don’t use it as much as I should, but I’m a big fan for any like listen-only call, just walking on the walking pad, getting your steps in. Or if you have like an all hands, I usually go to the dog park and take the dog for a walk and call all pause. That was one of my favorite work-from-home hacks. And the last one, which I think is is a flex, is wired Ethernet.

and that has helped me you know, we’ve all been there where Zoom freezes and it’s like, I think it’s you, maybe it’s me, maybe no, no, it’s definitely you, because I have wired Ethernet, it’s not my Wi-Fi. and that has just been really great for confidence and being able to trust my connection. I don’t know that I noticed the speed benefit as much, but just knowing that you know, your your your Zoom presence, your team’s presence, that’s your connection with your coworkers and just really all the the content the count the the constant between all those.

is just really investing in leveling up your persona, just as you would, you know, clothes in the office or you know, the car that you drive and or whatever else it might be, of really just being purposeful about how you show up in a remote environment.

Bethany (1:03:16) Man, I’m taking notes because I want all of those basically. I I’m so on the fence about the walking pad because I don’t even use my walking or my standing desk right now, but I don’t know, this might this might push me.

Ben Balter (1:03:31) and you mentioned

photos. I will get you photos. I just have a

Bethany (1:03:34) Yes.

Ben Balter (1:03:35) mess of wires in front of me right now. I have a very nice chaos monkey that is our four-year-old that every time I get back to my desk, things are unplugged, things are plugged into different places. so I I need to get some wires sorted out and I’ll send some photos over. But it’s a great way to know that if something breaks you have fallbacks for your video conferencing when your microphone is just randomly unplugged.

Bethany (1:03:59) That makes sense. But I mean you’re you’re growing a little cable manager in in

Ben Balter (1:04:05) Tw.

Bethany (1:04:05) practice, so they’re gonna be so good when it comes to remote work on cable management. That’s that’s a win I think. Well,

Ben Balter (1:04:12) Exactly.

Bethany (1:04:15) we will let you sit now, but beforehand, where can folks find you?

Ben Balter (1:04:21) open an and async.com. A book just came out. There are you can get it where anywhere books are sold. I write regularly at ben.balter.com where I just kind of throw thoughts into the void. Highly recommend the book over my random thoughts into the void. it’s it’s much more curated. and I’m on Blue Sky at Ben Balter. Ben.balter.com. so yeah, looking forward to connecting folks on the internet. Really, really glad to be here and to have this.

conversation and hope folks found it helpful.

Bethany (1:04:55) Awesome. Thank you so much, Ben, for joining us. This was such an awesome conversation. I’ve learned so much, and the book is truly fantastic. So, seriously, whoever’s listening, please read it. Please go out and get it. It is awesome. so thank you so much, listeners, for tuning in to Overcommitted this week. If you like what you hear, please do follow, subscribe, or do whatever it is you like to do on the podcast app of your choice. Check us out on Blue Sky and share with your friends. Until next week. Bye.