Subscribe
Share
Share
Embed
- GopherJS 1.20 released
- Listen to interview with Grant Nelson, Episode 53
- Results from the 2025 Go Developer Survey
- Interview with Dominic St-Pierre
Creators and Guests
What is Cup o' Go?
Stay up to date with the Go community in about 15 minutes per week
This is Cup of Go for 01/23/2026. Keep up to date with the important happenings in the Go community in about twenty minutes per week. I'm Shay Nehmad. And I'm Jonathan Hall. And we have
Jonathan Hall:a hard cutoff at seventeen minutes here. Oh, sixteen minutes now because we have an interview coming in. So we're gonna I think we'll hit our target this time.
Shay Nehmad:Let's see. Big news for this week is the survey results that we wanna dive into and try to dissect. Sort of like sight unseen, we are cursed with the curse of knowledge that we have already used Go for a while and also specifically dissected the previous surveys and previous episodes. But this week, we're really gonna try to approach it as if this is the first survey we've ever seen and try to understand the results at face value. Before we dive into the survey, though, there's one library release that we wanted to mention, Gopher JS.
Jonathan Hall:Yeah. Gopher JS just released version 1.2. Longtime listeners may recall back in 2024 in March, we had Grant Nelson on who was working on getting generics added to Go Forks. Yeah.
Grant Nelson:Sure. So I'm Grant Nelson in Montana. I've lived here almost my whole life. Went to MSU and took a long break for a while, and now I'm back working on my PhD. Okay.
Grant Nelson:I prefer to be a teacher, but I ran out of money, so I had to go get a real job.
Jonathan Hall:Go for JS compiles go to JavaScript for another browser. Yes. So about a year ago, I think, 1.18 finally came out with generic support. Since then, we're slowly getting caught up. I'm one of the co maintenders as well, although I didn't really do any work other than code review on this recent change.
Shay Nehmad:Honestly, with AI code, I've been doing more code review than coding recently. Yeah. For
Jonathan Hall:sure. But, anyway, big shout out to Grant for doing a great job on getting Gopher JS 1.2 out, And he's already making progress on 1.21. So pretty soon, we'll be up to the modern versions of Go. Nice. So Fine can finally switch back to Go
Shay Nehmad:for JS. Andy, we're calling you out specifically. So an interesting release. There are more things happening. Like, our backlog is
Jonathan Hall:Really?
Shay Nehmad:Oh my god. It's getting big. And before we dive into the only thing we wanna talk about this episode, other than the Go for JS release, which is a survey, I want to mention, right now at the top of the show, it's our three year anniversary. We're celebrating next week in a live episode in San Francisco in the Go Meetup that I'm arranging. So if you're in the Bay Area, obviously, just look for gosfinmeetup.com and tell me that you're coming.
Shay Nehmad:But also, we would like to have, your voice notes. We got some voice notes. We would like to get a little bit more. If you're uncomfortable sending a voice note and you just wanna send us, like, a text and we'll use AI voice to generate it into a note for you, that's also a 100% okay. Just let us know, you know, something you'd like to be on the air on the show.
Shay Nehmad:All right. Awesome. Survey time. Survey says? We're doing that again.
Jonathan Hall:Okay. What else can we possibly do? That's the only catchphrase I can think of related to surveys.
Shay Nehmad:Damn. Now I need to find more catchphrase related to surveys. What is the GoDeveloper survey?
Jonathan Hall:Go Developer Survey goes out every six months roughly, and it's a survey of Go developers about their experience with Go. This time, 5379 Go developers responded. Obviously, there's more Go developers than that out there, so it's a volunteer basis. But we still get some interesting insights, the Go team uses these insights to sort of prioritize, I guess, what they're working on, determine if there's features that are sorely missed or other efforts that can be made to help improve the Go ecosystem. And this year,
Shay Nehmad:or I guess this survey, there were 5,300, respondents, which is a pretty sizable amount. I don't know statistics, but I just, spoke to a friend who works at the at some big company, I don't wanna because I'm gonna say something bad about that company, but at some big company doing experiments related to user experience on a physical, device, a foldable phone, and it's like, oh, I got to the statistic team and they don't know statistics. Like, they did an experiment, checking if the phone is comfortable, and the experiment was asking their friend if the phone is comfortable. It's like n equals one. So I don't know if n equals 5,300 people is good enough.
Shay Nehmad:To me, it sounds like a pretty sizable amount. I wonder how many of these people are a go couple of go listeners as well. Oh. But who are these people? Right?
Shay Nehmad:The first interesting question before you even dive into the results is, like, who answered? And generally, it's relatively experienced professional Go program. So people who've been doing this for between three and five years and use Go for their primary job, not necessarily the only language or even the primary language that they are using, but they are using it at work, or for, open source projects as well. A little bit, for school, which I, like seeing always. And most, if not all, you know, are professional developers.
Shay Nehmad:Like their job is to develop code. And by the way, the age bracket is pretty interesting as well, between 25 and 54, with most responses between twenty five and forty four.
Jonathan Hall:Yeah, two thirds in that bracket, 25 to 44.
Shay Nehmad:Yeah. I always find it hard with like which industry your organization operates at, because every company is a technology company. Right?
Jonathan Hall:Yeah. Like,
Shay Nehmad:even my, like even small businesses now have websites and databases and whatever. So I find it hard to look at the like, which industry best describes your organization, but it seems like, you know, there's about half working in technology just as technology, and then the rest are the other, it's like a long zip zip f tale of other industries, like media gaming, retail, health care, education. That's interesting. Like, I don't know who we had on the show that's doing go for education. I guess, boot dev lane.
Shay Nehmad:I guess the company is considered education.
Jonathan Hall:Education about tech.
Shay Nehmad:Yeah. And we had there's public sector nonprofit. We had the the I forgot his name. The person who's doing the advent of code, working for, a co op or something. Yeah.
Jonathan Hall:Yeah. Yeah.
Shay Nehmad:I guess that's where they, identify. And obviously a lot of people in The US, and Europe. That's like that's the people who answered the survey. Not surprising to me. I guess if I would think about the prototypical Go engineer, they would be someone in The US or Europe about between 25 and 44 and working as a software developer.
Shay Nehmad:Like, that's not surprising at all to me.
Jonathan Hall:Yeah. So 23% of respondents are in The United States, and second on the list is elsewhere at 20%. So that tells you that it's a long tail.
Shay Nehmad:Yeah. Yeah. By the way, I am pretty sure, and I'm gonna check it right now, like live. Well, this is not live live, but live is the time of the in media res, I guess I should say. Mhmm.
Shay Nehmad:How our listener geography chart matches to the so it it matches. It matches like one to one. United States is first, Germany is second, UK is third. Oh, but our listeners, we have Canada at fourth, and Canada is only fifth in the in the chart. So there's some divergence between the people who answered the survey.
Jonathan Hall:Quite a bit higher on the survey than they are in our listenership.
Shay Nehmad:I don't wanna France, you need
Jonathan Hall:to represent and start listening to our podcast.
Shay Nehmad:But I'll tell you what. And I think can I engage in stereotype against the French people?
Jonathan Hall:No. We want them to listen. We don't want them to stereotype them.
Shay Nehmad:Generally, in the podcast, you only listen, but in the survey, you can complain. So, of course, they I'm just kidding. I'm just kidding. Okay. So these are the people who answered.
Shay Nehmad:Generally, they feel very positively about Go. So this is, in line with previous years, where more than 90% of, respondents say they feel satisfied about working with Go. What is your take on this? Is this, like, selection bias? The people who answer the survey are the people who are satisfied?
Jonathan Hall:Well, I think that's a big part of it because I've certainly met people who don't or or I I've brushed next to people on social media who don't like Go. You know? They they they hate static types or they definitely want their, I don't know, Java style inheritance or whatever. But there aren't very many of them. So, I mean, I think I I I do think that, if you could do a truly random sample of developers and ask them what they think of Go, at least the ones who've tried it, I do think the majority would say that they have a positive opinion of it.
Jonathan Hall:I don't think it would be 91%.
Shay Nehmad:I also think that, you know, people who work with like, the question is how satisfied have you been using Go during the past year? If you were to use Go for something that it's not a good fit for, like, when we evaluated Go for our current startup, we decided to go not with Go, even though I obviously really like it personally because we figured, oh, we're trying to iterate kind of fast and we're trying to model some high level concepts and maybe Go's like rigidity and whatever doesn't exactly fit. And also not exactly like the The hireability of the people we wanna hire are not necessarily Go developers. We're looking for more like AI ish people, might be more used to high level languages. So, you know, I use Go for other things in other avenues, so I answer that I'm satisfied with it.
Shay Nehmad:But I can imagine if I were to pick Go for this and it would've for, like, the back end of the current startup, I would answer I'm, like, neutral or even dissatisfied because it's not a good fit for what we're doing, Even though I think Go as a language is correct. I'm wondering how many people, how many of the 9% that didn't answer satisfied are just trying to use it for something that it's not a good fit for, and therefore they're like, it's not great. I'm trying to use it to, like, build my front end or whatever.
Jonathan Hall:Yeah. I would like to know I would be interested to know that as well.
Shay Nehmad:Yeah. Unfortunately, the the quotes they put here are just are only strictly from people who answered very satisfied Mhmm. Which is kind of frustrating. You mentioned like which concepts you brushed against that people would really like. They did ask they basically picked our stumper question from the first year, which is like, what feature would you remove from Go or whatever?
Jonathan Hall:Mhmm.
Shay Nehmad:So they asked what is the next favorite language feature that's not in Go? So people said inheritance as, like, the number one thing, which to me is weird because Go has, like, quasi inheritance.
Jonathan Hall:I don't I don't think that's exactly what this chart is saying. I I was thinking the same thing. I I think what they've done here so the paragraph before this chart is a bit confusing, but it says that they ask people what other languages they like. And then from that, they took features of those languages and ranked them. So if you like Python, then that counts towards inheritance, not because you like inheritance, but because Python that you like has inheritance.
Jonathan Hall:I don't I don't think we can read too much into this chart. But, yes, inheritance was the first thing.
Shay Nehmad:So it doesn't necessarily mean Go doesn't have it. It's just like, I like well, out of the four things here,
Jonathan Hall:Go But it also doesn't necessarily mean that people like those features specifically. They like languages that have those features.
Shay Nehmad:Yes. Well, I mean, when you compare when you look at the survey, one of the things that pops out, especially when people mention, like, what's their favorite programming language or, like, what's the next language, is that Go is not anybody's first language, basically. People start with a different language and move to Go, which is a pattern we saw in the past. And with the sentiment generally being positive, but people moving to this language from another language, I think a natural derivative of that is people like doing the Venn diagram of, oh, well, I came from Python and in Python, I liked x, y, zed, and then I moved to Go, and in Go, I like like all the overlap with Python plus the things that, Go is, bringing on, but I don't like the part that's in the Venn diagram that's, oh, left behind in Python, I might need it. I might miss it as well.
Shay Nehmad:So the main thing I think that belongs in that, like, slice of the Venn diagram here is typesafe phenoms, which basically everybody who's who's working in Go and I've talked to has been and we've talked about this at length on the show as well. It's like, man, I wish Go had, like, proper enums. Right?
Jonathan Hall:Mhmm.
Shay Nehmad:Generally, it is interesting to look at what the answer of what's your favorite programming languages, even though I don't like that phrasing. Like, I would like people to answer what other languages they're actually working at because Rust is 11%. And, I don't know, other than a few people who are building developer tooling and one guy works on trains, I don't know people who do Rust full time professionally. I know people who do some Rust a little bit to optimize specific parts of their whatever real backend that's written in Python or TypeScript or Go, but like certainly not 11% of programmers I talk to are Rust programmers. I think this is like a way over presentation.
Shay Nehmad:I don't know, maybe you have a different experience.
Jonathan Hall:No, I mean, I have a few friends who use Rust professionally, but I don't know how much they use it like, exclusively. I have a feeling they go between Rust and c plus plus or whatever else is suit for the job, you know? And also, like, favorite language isn't isn't necessary. Like, there's many ways to measure a favorite. Like, you know, maybe I really like Erlang.
Jonathan Hall:You know? Erlang is just so great. It it fits all my personal quirks, but I never use it because there's no projects written in Erlang, you know? So, you know, on the other hand, you know, maybe I think C plus plus is the worst language in the world, but it works really, really well for this task that I'm working on, so I use it every day and it and it's effective. You know, there's different ways to measure favorite.
Shay Nehmad:One other so, you know, people have their favorite languages and whatever, maybe Rust is a favorite language they don't use that much. One thing that they added to the survey is how much did they do the people who answered the survey sorta have the sentiment not towards the language, but towards the project. Mhmm. So the people building it, the leadership, the contribution style, questions and responses, etcetera. And here, the responses were they were overwhelmingly meh.
Shay Nehmad:They weren't negative. They were very, like, you know, not they they weren't, overwhelmingly positive like they are about the the Go language itself, which is surprising. Like, if the product is pretty good, you'd expect the people to say that the kitchen is pretty good as well. But basically, what they're saying is my feeling toward the kitchen is meh, but I really like the dishes. Mhmm.
Shay Nehmad:Do you agree with the, like, oh, I don't what did you answer to these?
Jonathan Hall:Like I don't remember what I answered. I do I'm pretty sure I did do the survey. I don't remember exactly how I answered it. It's it's a little bit tricky. Some of these I think I remember talking about this, actually, when we talked about the survey.
Jonathan Hall:Like, one of the questions is I trust the Go team to do what's best for developers like myself. Like, that's kind of a tricky question because, like, I I don't know how much I'm like other developers or, you know, you know, if if they the only way to answer that, like, I'm a 100% agree would be, like, if they tailored Go specifically to me, you know, and I wouldn't expect them to do that. So I would expect this answer to not be a 100% because we're trying to aim for some sort of average developer, which doesn't really exist. Right?
Shay Nehmad:I mean, my understanding of the question is I trust the Go team to do things that are generally positive, like, you know, in the latest release, generally the garbage collector is better, so generally my software is gonna be faster, which is good for developers like myself. And they're not gonna introduce a crypto scam, a la Gastown that's or suddenly turn it into a closed source, and you have to pay the enterprise a fee for the, you know, quantum, the whatever cryptography or, you know, do some go to thing that's gonna bifurcate the community. Like, they're not gonna have one of these moments, even though they definitely could. Like, this is Google. And if they wanted to try and monetize enterprise how, like, Oracle is doing it, you know what I mean, with enterprise licenses, etcetera, etcetera, they definitely could do that and and harm developers like myself.
Shay Nehmad:So the that's why I answered that I agree. I I don't think the Go team is planning to do anything like that. I don't think that's, like
Jonathan Hall:That's true.
Shay Nehmad:What they're gonna do. And I do trust their leadership because so far, I've been following it super closely in the last three years, and they've done as part of doing the show. Right? And they've only had good decisions, like, okay, we're not gonna talk about error handling anymore. Enough.
Shay Nehmad:Like, that's a great leadership decision or
Jonathan Hall:But that's a great example that goes to the to sort of my take on the question. If they were going to do what's best for developers like myself, they would make a different, you know, error handling decision than they would if it was for for you or for for for Bob or for Alice because it's such a controversial sort of point. Right? So they have to they have to settle on some sort of average, what's best for some sort of average developer rather than what's specifically best for whoever the person answering the interview or the survey is, if that makes sense.
Shay Nehmad:Yeah. I think, though, you know, you can imagine how annoying it is to be. Unfortunately, I have experience, not firsthand, but very, very close, secondhand, of, like, enterprises that are still stuck in the Python two to Python three migration, and, you know, the pains of, oh, I have to develop it once in Python two, and then I have to migrate it to Python three, and then I get another code review in Python three, and then I have to go back to Python two. Like, very frustrating. So I feel like even if you personally, a developer like yourself, feel like, you know, the question mark operator would make your code so much nicer, what's best for you actually is still to have a very stable language.
Shay Nehmad:And that's better for you than having this one specific feature that you really, really want. And another example of, you know, the my confidence, at least, to the Go leadership is if you remember the telemetry thing. Do you wanna refresh our listeners memory in case they forgot?
Jonathan Hall:Yeah. Yeah. Hope they forgot because it's been a while.
Shay Nehmad:And since then, they turned it on silently. No. I'm just kidding.
Jonathan Hall:That was one of the first things we talked about on the show for for a few weeks. Yeah. They added telemetry to the Go toolchain, and it was going to be opt out. And then based on feedback from the community, switched it to opt in. So it's there for everybody, but you have to opt into it first.
Shay Nehmad:Yeah. And they have, like, a little pop up on Versus Code when you install it for the first time. It's like, please do that. By the way, if you haven't done it, go run GoTelemetry on or whatever the command is. Go do it now.
Shay Nehmad:You're not working on the most top secret, most secret, whatever, blah, blah, blah. Just give them your telemetry. Yeah. So I don't know. I'm sort of interested in the responses to this.
Shay Nehmad:Definitely, one thing that I agree with is that the process of contributing to Go is clear to me. Most people don't feel that way. Only a quarter of people do. I often read the issue discussions and often read, like, a Garrett change list. And I don't have like I don't feel comfortable creating a pull request or whatever right now.
Shay Nehmad:I don't think I'll be able to. Without guidance from like someone like yourself, but you already did it, I don't think I'll be able to do that. And maybe it's not important because I don't have anything specific I wanna contribute to the language. And maybe actually raising the bar and making it more difficult is is good right now because otherwise people would create like slopware, pull requests, that'll just overwhelm the Go team. Maybe it's not something even to be improved.
Jonathan Hall:Maybe it's okay. Think there's two aspects to this. One is there's many ways to contribute, and writing pull requests isn't the only way. That could be improving documentation or organizing a Go meetup or, you know, there's so many ways you can contribute to the language or to the ecosystem or the community. Second, this question doesn't really answer with a more relevant question, which is I want to contribute and it is clear to me how to contribute.
Jonathan Hall:Like, if you don't if you have no desire to contribute, as you sort of alluded to, then then it doesn't matter if you know how. Right? So what's really important, I I think, is if it's if the process of contributing is clear to those who want to contribute.
Shay Nehmad:Yeah. But if you make it more accessible, then more people will want you know, if it
Dominic St-Pierre:How am I gonna get the bands to come? If you book them, they will come.
Jonathan Hall:So, yeah, I agree. And I think this is a hard question. Like, do we want more contributors or not? I don't know the answer, and I'm sure different people have different answers to that question anyway. But I I don't know.
Jonathan Hall:As a contributor myself, I do feel like the contribution process is fairly opaque and kind of difficult more so than necessary. Like like, I fixed a bug. I think it's coming out in 01/1926, the bug fix, but it took me about a year to get it reviewed. Several times, I had to ping people over over and over again to get it reviewed, and it it was a simple one line fix, but just the process of getting it reviewed and and merged was much more complicated than I think it should have been for something that was a bug fix. It wasn't even like a controversial feature.
Jonathan Hall:It was a bug fix. An objective bug fix.
Shay Nehmad:The error handler to the correct format, which is a question mark. A clear bug fix.
Dominic St-Pierre:No. I'm just kidding.
Shay Nehmad:Yeah. So shifting from the Go team to the again, back to the people who answered, what are the respondents, like, building? Generally, I I think it won't surprise you that people are building CLIs, they're building like backend services, they're building libraries and frameworks, and they're building cloud infra things, data processing things. Surprisingly, 12% are saying they're building desktop and GUI applications, which is That's cool. Really surprising to me, but I guess great
Jonathan Hall:news friends.
Shay Nehmad:Yeah. Fine. One new bracket here that's very interesting to me is that 11% are saying they're building AI models, tools, or agents. This is a big, like, category, right, that starts with actual machine learning frameworks like we had on the show Hugo, right, and ends with all these, like, orchestration frameworks that are coming out recently, like, oh, I paralyzed 10 cloud codes in my rig and my blah blah blah. You know what I mean?
Shay Nehmad:And, I mean, there's also the question of does your software use AI functionality? And it's more than, like, you know, I'm sure if we asked that like ten years ago, it would have been like 2%. Like a very low amount of people were using AI as part of their stack. Here, it's like almost a third of the respondents have AI in their software, and 20% of them are actually working integrating that feature themselves. So about one person in every software team is, like, adding a chatbot to their website right now.
Shay Nehmad:Are you happy about this, response? Just generally for the software engineering world, before we even try to understand what what the impact is on the language?
Jonathan Hall:So I I think this is a a misleading question in the same way that people
Shay Nehmad:You really don't like the questions.
Jonathan Hall:Well, I should say a misleading answer or misleading option. In the same way that answering technology company is misleading because everything is a technology company. Like, every company is doing AI work these days virtually, and if they aren't now, they will be within twelve months probably. So everybody who does CLI tools and API RPC services and libraries, all all these other categories will also be doing AI ML model work. So it it's I think we need a better display.
Jonathan Hall:We need a Venn diagram here rather than bar charts.
Shay Nehmad:Yeah. But that that'll make things, even more difficult to understand. I think it is interesting to see what LLMs and AI is used for. So the top thing is summarize things. So if you have a back end endpoint that, like, receives an email and then calls OpenAI to summarize the email and shove it in a DB in the summary column, that counts.
Shay Nehmad:And that's like the number the two top things are classifying things and summarizing things. So summarizing things was basically a non feature. It wasn't a possibility that I think for most software developers before LLMs, right? Like reading, understanding something conceptually, and then outputting a summary of it wasn't really a thing you can make your computer.
Jonathan Hall:I think the closest we had was like sentiment analysis, right?
Shay Nehmad:And obviously it's very, very interesting. Like a lot of what people do in general is look at the thing and summarize it, especially now that a lot of content, which is the third ruby key, is generating things. So you use an LLM to generate like a huge ass email, then another one to summarize it on the other end, you know, net positive just for Anthropix stock price. But people do use it for the things you would expect, summarization, classification, generation, suggesting solutions to problems, chatting with end users. I was surprised that that's 30% because every time I use an like, an AI agent answers me, I get super frustrated.
Shay Nehmad:Recently, like, I was looking for an apartment or I guess rather I should say a house. And, you know, there were a few human agents that talked to me, like real when I say agents, I mean like real estate agents. And there were a few AI agents agents that talked to me. And I can tell you for a fact that I did not answer the AI agents agents. Yeah.
Shay Nehmad:Like, you're trying to get me to sign on a so and so dollar lease a month? You better have, like, a guy that is, like, sweaty and has a Rolex and can walk me through the poverty.
Jonathan Hall:My my wife just did a job interview with an AI chatbot, you know, visually, and she said it was really weird. I don't know. Like, I wouldn't I wouldn't even accept it, but she needs a job. So, you know, I understand why she did, but
Shay Nehmad:yeah. So moving on to the things that can be improved and not just who answered and what do they feel, what do people say the
Dominic St-Pierre:What do
Shay Nehmad:biggest challenges are or the most frustrating thing when writing Go? So the
Jonathan Hall:top one is following best practices and Go idioms, which is interesting because I see a lot of non idiomatic code out there. Maybe that's good or bad or maybe it doesn't matter, but the point I wanna make is that a lot of people aren't even trying.
Shay Nehmad:And maybe that's You say non idiomatic. When you say non idiomatic, what do you mean? Can you give it
Jonathan Hall:So things that you know, they they clearly violate the the stuff that's in in this published idioms from the the Go team. You know, they they use all caps for constants or they use underscores in variable names or file names. You know, most of them are fairly cosmetic deal. You know, cosmetic things, they aren't necessarily a big deal, but it's, you know, it's clear that they aren't even trying to implement or maybe they don't use CamelCase or whatever. It doesn't matter, but they're not trying to do the things that are well established in the Go community as the, quote, Go way of doing things.
Shay Nehmad:I think that one interesting side effect of this I don't know. This is me sort of putting my my chips down, but I would be super surprised if the Go team doesn't release in the same way that, you know, Supabase released Postgres best practices and Vercel just released the React best practices agent skills. I would be like, the Go team, I think, will will miss a moment in time right now to release, oh, we saw the survey and we compiled all of our best practices, idioms that don't fit into GoFix or Go, you know, that modernizes your code automatically into Go best practices, like an agent skill that you can install on your agents. Maybe I'm living in like a Twitter, San Francisco, SF Bay Area hype bubble, agent skills are not really a thing outside my bubble. But I think that could be a good opportunity for them to like make an official agent skill thing instead of just people doing the open source dance, and then we'll end up with five different skills versions.
Shay Nehmad:Other than that, the next thing is obviously a feature I value from another programming language isn't part of Go, namely probably, you know, error handling and enums. And the third one is something that I found frustrating as well, which is finding trustworthy Go modules and packages.
Jonathan Hall:That's how you identify what's I don't even
Shay Nehmad:do that. Like, when you want to bring in a dependency, like, what's your checklist?
Jonathan Hall:Well, I usually start with Google or maybe an LLM to ask what the options are. Probably Google for that. LLMs are less good at finding unpopular options, right? They'll tell you what the general community does, not necessarily to something that came out six months ago that's even better. So I usually use Google to find, you know, what packages do the thing I'm looking for.
Jonathan Hall:Then I'll look at the GitHub or whatever. You know, it's usually GitHub to see how many stars they have, how active the repository is. Are there a lot of outstanding bugs? Do they respond does a maintainer respond to issues and pull requests and things like that? I want something that's getting the love it needs.
Shay Nehmad:Although, if, you know, if a package is like finished, it's pretty simple and it's finished.
Jonathan Hall:That's rare, but there are some
Shay Nehmad:It examples of looks like abandoned on GitHub, but honestly, lot a of good packages that I've used look like that,
Jonathan Hall:where it's pretty You know, they're fairly they have a very limited scope, like maybe creating diffs or something like that, know, something fairly limited scope that doesn't need to be updated. There are a few of those that I use, but usually there aren't a lot of competitors for those sorts of things. Because it's been solved well and it's not a complicated problem with lots of nuance that, you know, there's 13 different opinions on how to do that thing. You know, it's less of a problem, I guess, anyway, usually.
Shay Nehmad:And by the way, I don't think this is a solved problem in any ecosystem. It's not easy to find trustworthy GO modules and packages. It is maybe slightly easier to do a package like supply chain security in Go rather than in in dumpster fires like NPM or whatever. It's it's not easy. And I think it it's not an easy solution.
Shay Nehmad:It's a very social problem. Like, defining if a package is maintained and trustworthy, it's like a judgment call more than any technical score you can put on it. But definitely, official SDKs, that's that could be one thing. Like, you know, you wanna talk to a company, a company can register their official SDK, you know, because a good example is OpenAI. They have, like, 15 different Go packages until they release their Go SDK, but now they have their official Go SDK.
Shay Nehmad:It might not be the best, but I know for sure that I would like to use the official OpenAI SDK when talking to OpenAI, you know, to build my LLM summary that I just talked about from the previous section, rather than some dude's package who might be cool, but, you know, he might be traveling next month and not answering. You know what I mean? So I don't know how to I don't know how to approach that problem, but it is a problem. By the way, unlike ensuring Go code follows practices and Go idioms, I think that's pretty easy. Just improve Go fix and Go vet, and teach more people about Golang c I lint, and that's gonna make it easier.
Shay Nehmad:And a feature I value from another programming language isn't part of Go. If it's error handling, GoFish. And if it's, enums, maybe we can make it happen at some point. These ones, they feel like solved to me. Like, I know what the direction is gonna be.
Shay Nehmad:For Go modules and packages, I don't know. There's a lot of other things here that people find frustrating, like Go PLS is not responsive enough or, you know, optimizing for resource consumption or testability or, you know, model dependencies, latency, etcetera. But I think that's like enough of the challenges. One final thing that we have to mention, of course, is that Jonathan uses Linux.
Jonathan Hall:I swear, man, you mentioned that far more often than I do. I think you're jealous. I think you want to use Linux.
Shay Nehmad:I have my my Corpo Slop Mac OS with the glass effect, and I'm happy with it. However, 60% I, by the way, don't understand this because 60% said macOS and 58% said Linux, 20% said Windows, and 14% said Windows subsystem for Linux are the platforms that they use when writing Go code, which means a lot of people use both macOS and Linux. So I guess people are, like, when they develop on a dev container that's, like, Linux based, they consider it Linux even if they're running it on Docker on their Mac machine from work.
Jonathan Hall:And if they answer the Windows subsystem for Linux, do they also answer Windows and Linux?
Shay Nehmad:Oh, yeah. They count three times. Yeah. And obviously containers are, like, where people deploy their code almost exclusively. In deployment, 96% of people deploy to Linux, so that's definitely still where all the software is actually being run.
Shay Nehmad:Of course, we have to mention the editor wars as well. Jonathan, what code editor do you use?
Jonathan Hall:Begrudgingly, I use Versus Code.
Shay Nehmad:Begrudgingly. I've never been
Jonathan Hall:a big fan, but it gets the job done.
Shay Nehmad:So the question is again phrased not like, what are you using? But rather, what's your favorite code editor for Go? So what's your favorite code editor for Go?
Jonathan Hall:I guess it's Versus Code.
Shay Nehmad:Damn. Yeah. Another one bites the dust. By by the way, this answer, they did something smart, and this answer doesn't count the people who were referred to answer the survey from the Versus Code prompt.
Jonathan Hall:Okay. Yeah. That's clear.
Shay Nehmad:But Versus Code is obviously number one. Goland is number two. Vim and NeoVim number three. So these are, like, the main options right now for Go developers. The the rest are are pretty interesting because you have I'll skip over Zed once again.
Shay Nehmad:I'll get I'll get back to Zed. We have Cursor and Windsurf as options, both of which are Versus forks. And I don't know, I use Cursor, but I think I answered Versus Code just because it's a Versus Code fork. It just has some like, you know, it it goes to a different back end and has some nice features, but I don't know. You know what I mean?
Shay Nehmad:It's it's Versus Code.
Dominic St-Pierre:It's not
Shay Nehmad:something else. Yeah. But, yeah, I use Cursor. Other options here include Zed. Have you tried it out?
Shay Nehmad:I did. I really wanna move to it, like, full time at some point. It's super fast. It's written in Rust. It's really nice and and nice.
Shay Nehmad:I might turn it into my Cloud Code driver in, like, one work tree while I'm working on my main thing in Cursor. Mhmm. 3% answered Emacs. The rest of the Emacs users were still busy configuring their Emacs, which is all they do. And there's still 1% of people who code go in sublime text, which to me is, like, insane.
Shay Nehmad:But sure, you do you. There's a breakdown of cloud environments. It pretty much matches what you see in the marketing reports. AWS number one, GCP number two, Azure number three. Mhmm.
Shay Nehmad:I think, by the way, this is like the other way around in real world. It's just that maybe Go has like, Google Cloud and Azure are are switched in the real world. I think Azure has more market share generally, but maybe not among Go developers. Maybe Go developers like Google more because it's all, like, holistic or whatever.
Jonathan Hall:I don't know. Mhmm.
Shay Nehmad:And, you know, in terms of who gives the best experience for Go developers, again, GCP is pretty high, which is surprising to me because, I don't know, I've I've heard from people working in GCP that they didn't really like it, but, you know, I I guess this is a good answer if you're Google. You don't wanna put Google Cloud as, the last option, and I'll I'll leave it at that. I don't think there's been any in proprietary, but maybe some massaging of the data. And the final section, unsurprisingly, is about AI. Not AI in your software, but AI in your software development process.
Shay Nehmad:Honestly, things I feel it's it's crazy because it's only January 23. I feel things are already different than 2026 AI developing. You're rolling your eyes. You don't seem to agree.
Jonathan Hall:So my biggest thought when I was reading through this stuff about AI is I don't even remember how long ago they asked these questions, but I'm sure my answers have changed significantly since then, And I bet they have for almost everybody who answered. So these answers are probably this is like a snapshot from three months ago that is probably almost irrelevant now. Mhmm. Because my AI habits have changed drastically in the last three months, and they're changing all the time. Like, even just this last week, I learned some new things about how to use Claude, and now I'm making Claude better.
Jonathan Hall:And I've been reading articles and, you know, looking for new ways to improve my workloads. It's kind like we've all been given Google for the first time after never having a search engine before, we're asked to do a survey about it, and then three months later we're asked to look at the results. And like, that's not how I use Google because we're learning all these new things again.
Shay Nehmad:I will say last year they asked about AI assistance kind of differently, so it's very hard to see the trend, and they also point that out. Generally, just to give you, I don't know, if you're a listener and you use AI in a certain way, there's like a spectrum, right? There's a spectrum on one side is, I don't use AI at all. I don't want like Copilot. I don't want anything.
Shay Nehmad:I just use code in the I just code like it's twenty twenty two. And on the other end of the spectrum, have like people like, I have, 10, concurrent instances of a cloud code with WorkTrees and an Orchestrator, and I never look at the code anymore, and I don't even do code review. I just, like, do spec based development. And maybe there's even more extreme than that, like, on the other side of the spectrum that I haven't reached yet. And I think most people do use agents sometimes.
Jonathan Hall:I think that's a difficult question to answer. Like three months ago, would have said, No, I don't use agents, because I just used ChatGPT and I used Copilot. Now I tend to use Claude a lot more, and it does agents whether I ask it to or not. Or maybe you consider Claude itself to be an agent. Know?
Jonathan Hall:I don't think that distinction is clear to most people anymore.
Shay Nehmad:I think agent is an over super as a person who's working and an agentic cybersecurity, like AI cybersecurity startup right now, the word agent is meaningless.
Jonathan Hall:Exactly.
Shay Nehmad:Yeah. But I do think that there is a big difference between I sent a question to ChatGPT via the web interface, and then I copied the answer back into my code, which is like I use LLM as a Google replacement versus I give a task, like a coding task to something on my machine and it runs in a loop and it can run shell commands on its own, or, you know, it might ask me for permission, but it comes up with a shell command to run instead of telling me, please run NPM test or sorry, go test and give me the response. Right? Maybe post the exception here, it's like, no, I'll just run the thing myself and find the exception. I think that's the distinction that most people will agree on.
Shay Nehmad:And overall, are not just happy enough with their AI responses. Mostly people are like not satisfied with the code quality, it never follows practices, it hallucinates, you know, all the various complaints we already know. Again, I think just like a few months after this survey, the landscape has drastically changed because new models have come out and agent skills and MCPs and whatever, but still I don't trust agents as far as I can throw them. I'm still reviewing every single line.
Jonathan Hall:Yeah, for sure. So my thought on that particular question, overall how satisfied or dissatisfied, I'm like, I am both extremely satisfied and extremely dissatisfied. You can't you can't really answer this on a scale. Like, and so so my that would make it you know, you would think that would average out to, like, neutral, but I'm definitely not neutral. I have very strong opinions about how both satisfied and dissatisfied I am with my agents.
Jonathan Hall:So this chart just it's it's it it doesn't convey anything meaningful to me.
Shay Nehmad:One thing that is interesting is how people are using AI tools. Like, we're only assuming it's for writing code, but it's actually searching for answers, writing code, learning new concepts, suggesting fixes or problems, finding problems, reviewing code, writing documentation. This is one example where people say they have AI write the documentation, and I'm like, what the hell? It always creates super lengthy, super wordy, incorrect, duplicated documentation that nobody's gonna read. Like, nobody in their right mind are gonna look at that and be like, this is useful information I read I need right now.
Shay Nehmad:I don't know. I hate it. And I don't understand how, like, 46% of people are like, oh, I use AI for writing documentation today. I guess I I'm not sure it's even better than if they didn't write any documentation at all, but maybe it's slightly better. And I don't know.
Shay Nehmad:For example, generating test data and synthetic data is something I do all the time with AI, and I'm really satisfied with. It seems like only 45% of people use this today. And on the other side of the spectrum, deploying to production and monitoring production environments, basically no one is doing with AI. Everybody's like saying not only they don't hope to use it or they might try AI, they're like, I don't want AI involved in this. Like, let me deploy to prod.
Shay Nehmad:Don't touch it. Which I find I very much agree with. Like, I wouldn't want AI to, you know, monitor my production environment because it's just gonna do whatever it can to make sure that it's okay, like rollbacking and deleting my production database, and now all the pages are faster again. Yeah. So that's the survey from, you know, September 2025.
Shay Nehmad:You know, you can go read it yourself. We pretty much covered, I feel, everything. What is your general sentiment, like looking at this survey and thinking about Go's like future, I guess, I should say.
Jonathan Hall:I'm optimistic that enums and some of these things that I've been complaining about for a while are coming up, and it seems that the Go team is noticing. And we had we had a chat mentioned on the channel that the compiler team meeting has has this on their apparently, unofficial road map for for 2026. So, yeah, I'm optimistic that maybe we'll get enums and some of these other things that it may be union types or something that many of us have been clamoring for for a while. Now that maybe now that we finally put error handling to bed, you know, officially for a while, we can focus on some of these other things. And, yeah, that would make that would that would make me happy.
Jonathan Hall:1.26 has a lot of great new things in it. We talked about it the last a couple weeks ago, so I'm already excited about that. But if we could get enums, neotypes, a few other niceties, it's gonna be a really a really great language to work in, not just the best language that that I complain about. It could be the best language in in many other ways.
Shay Nehmad:Nice. Optimism is good.
Jonathan Hall:Yeah. What do you think?
Shay Nehmad:Same. I'm very optimistic that there will be people who are very passionate about, like, making Go better for agentic development, because that seems to be the thing. And also Google has a horse in that in this race. Right? Like, they want to sell their own models, Gemini their own, like, AI stack.
Shay Nehmad:Right? Gemini and and the Gemini models and whatever. And people have been saying that Gemini three is really, really good. I haven't tried it yet. So I feel like they ought to look at their responses from the survey and be like, okay, I know exactly, like for these 4,000 people that are professional developers, I could like try and get them hooked on AI development with Go, which both strengthens my position as an AI player, which is the only thing which matters for like companies right now, basically, but also generally should make the language better.
Shay Nehmad:Generally, when you try to make things better for agents, you turn out to make them better for everybody because like, oh, tests are faster or documentation is better or best practices are more automated or etcetera, etcetera. It turns out even for just human developers, things turn out better as well. But I honestly think that this survey mostly shows the team keep doing what you've been doing. Like, people are very satisfied with the way things are going. There's no major change needed to keep everybody pretty happy.
Shay Nehmad:Just keep making the language better and faster slowly and surely, know, people will be happy with it. Right. I know I am. Alright. Cool.
Shay Nehmad:So we covered the survey. Have a ton stuff in the backlog, but that's enough for today. Next week's episode is gonna be a live episode at the GO San Francisco meetup. Mhmm. Exciting.
Shay Nehmad:Please send us your voice notes for that.
Jonathan Hall:And if you don't, we're gonna chase you down and we're gonna rip them out of you.
Shay Nehmad:We are gonna DM some people just to get make sure that nobody missed the message. And, yeah, let's do a quick ad break and then jump to our interview with Dominique St Pierre.
Jonathan Hall:All right. For almost three years. We really appreciate the community support we've gotten. Of course, we would be remiss if we didn't remind you or mention for the first time for new listeners how you can best support the show. This is a volunteer effort, but it's expensive.
Jonathan Hall:It's an expensive hobby mainly for editing, but also hosting fees, running our swag store, various things like that. So if you want to support the show, you can do it in a number of ways. The easiest and freest is just to share the show with your friends and colleagues and other fellow gophers. You can also support the show financially, either by buying some swag, buy a Cup A Go mug or a T shirt or stickers, or you could also become a Patreon and support us on a monthly basis. That money just goes to help pay hosting fees.
Jonathan Hall:We are not making money off of this at all. We both pay significant amounts of money every year. To do this, if something makes you feel guilty, just to let you know we're not we're not trying to get rich off of Patreons like some really sophisticated YouTube beasts might be. Not to name any names.
Shay Nehmad:Yes.
Jonathan Hall:This next week, I think we actually mentioned last week that we were gonna do a birthday episode this week. I can't remember. I think this is our three year anniversary episode, but we're postponing the actual celebration until for one more week till we could do it live at the San Francisco meetup that I mentioned. Yeah. So you do have time still to send in a voice note.
Jonathan Hall:Just share your comments and your thoughts about the show. It could be just a quick happy birthday, Brewster. Brewster's our our mascot. It could be sharing your thoughts, something you've learned from the show, something you like about the show, something you don't like about the show, whatever. It doesn't need to be long.
Jonathan Hall:Send a voice note. You can find our email address, or you can do all the other things I mentioned over at cupofgo.dev, or you can send us a voice note on Slack. Send it to me or Shai or or share it in the even the cupago channel if you want to.
Shay Nehmad:Am I missing anything, Shai? Just one small thing is that other than supporting us through Patreon, we don't pay for any advertisements or anything like that. So you can also share the thing about the show. Last week's episode was pretty good, or was it two episodes ago?
Jonathan Hall:Two weeks ago. Yeah.
Shay Nehmad:Yeah. That was a big episode, and that was a lot of fun. So, you know, share the show with co students, co friends, with your agent. Just talk to Claude Coat about it. I don't know if like straight up, if an agent listens to an episode, like downloads the MP3, listens to it, and sends a summary to someone, would you want it to count as a listen and a download or no?
Jonathan Hall:I don't know. I think I think the first problem is why they listen to the MP3 when the transcript's already there. But true.
Shay Nehmad:Like, all the nuances. Maybe it's like someone takes all the episode and then cuts out all the jokes and the ad break.
Jonathan Hall:That's how we can get it done to fifteen minutes.
Shay Nehmad:Yeah. For sure. Okay. So just remember to share the show with people, rate it, whatever. You know?
Shay Nehmad:Just spread spread the word. Yeah. On to our interview. Woo hoo.
Jonathan Hall:And Hugh.
Dominic St-Pierre:Hello there. I'm Demi Sepiak, and you're listening to Go Podcast. So I am a software engineer.
Jonathan Hall:Sorry. Am I am I in the wrong place?
Dominic St-Pierre:Oh, it might be me.
Jonathan Hall:I I thought I I must have gone left when I was supposed to go right. I'm looking for Cupco.
Shay Nehmad:Crossover episode. Woo hoo.
Jonathan Hall:Hey, Dominique.
Shay Nehmad:Hello there.
Jonathan Hall:Thanks for coming. Joining us on the show.
Shay Nehmad:We we need the interviewees to say their name, but you already said you're Dominique Saint Pierre.
Jonathan Hall:Yeah.
Shay Nehmad:So please, you know, introduce yourselves to our listeners.
Dominic St-Pierre:Sure. Sure. So, basically, I'm a, you know, I'm a normal software engineer, I would say, doing that since 2000, and I really appreciate Go for a long time. And I have this this podcast that I started, It's turned four years, by the way, this January. Awesome.
Dominic St-Pierre:Yeah. Thanks. I'm I'm not the the most consistent publisher, though. That's that's the problem.
Jonathan Hall:Is it a problem, though? Do people complain?
Dominic St-Pierre:No. No. It's it's not a problem per se, but, I mean, you know, you know how it is. You always hear things like you should be consistent if you want to grow your audience and whatnot. Kind of have reached that plateau a long time ago myself.
Jonathan Hall:Yeah. For sure. Shy and I are consistent ish, but, yeah, we we take breaks and, you know, it's a volunteer effort. I I don't think you can you can blame a volunteer for for not being consistent.
Dominic St-Pierre:Yeah. Yeah. You you are you are you are too as well. So that, you know, when you are alone, I am alone. I so chasing after interviews is not the easiest thing for me.
Dominic St-Pierre:So that that you know, for for me, that's that's the the hardest part, you know, making sure that either, hey, I have good topics or b, I have a good, you know, person to interview and all all the work that is, you know, that that we need to do to do that. This is this is where it's
Jonathan Hall:It's a fair point. We'll we'll talk more about that being alone here in a little bit. But, Fare, go ahead, Shay.
Shay Nehmad:Yeah. I just wanna say thank you for coming to interview because it's so
Dominic St-Pierre:hard to find interviewer interviewees,
Shay Nehmad:so you made those things easy for us. But I wanna start with what is Go Podcasts? In case people do listen to Cup of Go and they're like, oh, I don't need another Go news program. I wanna start by saying your show is very different. I feel like every episode is more of a deep dive conversation into a specific general topic, not necessarily like this was released last week.
Shay Nehmad:And we mentioned some of your episodes on the show, but I, for one, I, for one, really like it. So I don't know if if Go, if you're looking to expand your Go Podcasting, like, you know, selection, and now that, I don't know, Go Time has been off the air for, a year or something, I highly recommend the Go Podcast. My first question to you is how do you pick the episode topics? Because looking at them, you know, it could be so one time it's like LLM and agents, one time it's CICD, one time it's, alternative to Docker, one time it's testing, one time it's linters, one time is, you know, just interesting conversation with, interesting Go people like John Arundel. You know, it's it's very, there's a very big, you know, like, sort of expanse of topics.
Shay Nehmad:One of them, by the way, being, Datastar, which we mentioned, this, on this week's episode as well.
Dominic St-Pierre:Yeah. Totally. I mean, to to be completely frank, I'm very bad at keeping with what's going on like you like you are doing, basically, if I'm not mistaken, at at the beginning of your episode. So, yeah, that's not for me. So for me, it's what do I, you know, what what do I want to talk this week and what what I did that I think will be worth, you know, the time for for people, and that that mostly it.
Dominic St-Pierre:So that's why that's why sometimes I there's there's just no episode because either, hey, I haven't done anything very extraordinary that that week or b, you know, I could not found any anyone. I I started the I started the podcast. It's it's it's it's interesting, I think, because, frankly, I was not really listening to podcasts myself. Being a blind person, this is kind of crazy because audio was supposed to be my thing, but I I was not. And I I, you know, I was invited twice at Go Time.
Dominic St-Pierre:I I started listening to them a lot, you know, a couple of episodes, but it was it was too long for me. It was way, way, way, way too long. And I knew Justin Jackson, which you might you might know. He is the owner of of Transistor. You you guys are using Transistor as well.
Dominic St-Pierre:So, you know, I've been knowing him for a long time, and he he was always saying, you know, you should you should start a podcast.
Jonathan Hall:He was just
Dominic St-Pierre:and he was he was always doing that to me. So I I said, oh, okay. I I will try that. And at first, it was like fifteen minutes, very, very quick. I I was trying to be you know, take a small let's take error in Go, for example.
Dominic St-Pierre:Or let let's take, I don't know, eight versus memory allocation at first, you know, pointers and whatnot. So I was trying to take some small or at least important topic like that that I was feeling people that are starting in Go might know. And I I was, you know, trying to to do a kind of a blog post, if you will, but an audio blog post, and that that is how I started. But, yeah, these days, I'm I'm just talking about, you know, what's going on in my life at the moment. Basically, that's that's it.
Dominic St-Pierre:If if it if I feel like it's interesting to to, you know, to go first.
Shay Nehmad:And drawing from, like, whatever you were, working on that week or the person you're talking to that week, I guess it does lead to, a very personal perspective on Go development. There's a pair of episodes I wanna point out, and you mentioned, so I wanna talk about where fifty six and fifty seven, which were released in April year last year, sorry, where you talk about, like, being blind in tech. So how how is that? How is being blind in tech?
Dominic St-Pierre:Oh, well, how many hours do we have in your show to to talk about that? I think that yeah. That that's very hard. So I'm legally blind since since birth, but I just I just recently in my life, maybe about I don't I don't even know. Maybe three or four years now, maybe three.
Dominic St-Pierre:I don't I don't recall. I needed to start using a screen reader full time. That was very painful for me. So I, you know, I always have that so that's why you're not seeing me at conferences and whatnot. It's really pretty difficult.
Dominic St-Pierre:You know, that is something that I would really like to do myself, but I I just cannot really move easily. I will need to bring my wife and things like that. You know, I cannot really take a plane alone and things like that. So long story short, it's pretty hard. So I met I met another developer and software engineer that that is also blind.
Dominic St-Pierre:So I think we we kind of did two episodes. Maybe maybe that was painful or not. I don't I don't know. We tried you know, when when I talk about that, the the reaction is, oh, people are just curious. Other software engineer, especially, and things like that.
Dominic St-Pierre:But but don't get me wrong. It's it's pretty hard. It's it's just abysmal, to be to be frank. And if I can say that, because now I have a new audience potentially, If you ever have a blind person coming to an interview, please, please, please let them stay in their environment. I did a couple of interviews in the last three years.
Dominic St-Pierre:I won't name any companies, but, you know, I cannot be outside of my environment. Let me stay in my environment. I will I I swear, I I will, you know, I will I don't know. I I will be able to to show what I can do, but, you know, a blind person cannot leave their environment. This is this is a fact.
Shay Nehmad:Yeah. I think this is a there is a few things, you know, that that the listeners, and also myself, can take from this. So the point about interviewing is interesting. If we have any hiring managers listening, which is possibly, like very possible right now, because, you know, yeah, I'm sure your company has a great return to office procedure and they really wanna use their real estate, But remember, it's not for everybody and there's a lot of benefit in hiring people remotely if they perform better remotely, like you're saying, Dominic?
Dominic St-Pierre:Well, I'm mainly talking about, you know, the initial interviews cycle, the tech interviews and things like that, you know, using using tools. Let's say let's say you're trying to stream some some engineers because you, you know, you don't want to to look at all the the resumes that you are receiving, and now you are using a tool, which I won't name again, but those tools are usually not accessible. You know, I cannot myself I cannot write code in a browser. That is that is not possible for me. So I I did something I did something like that.
Dominic St-Pierre:And, you know, to to reach out for the output, because let's say let's say you're you're writing some whatever the code that you're writing, but, you know, you you write your code, you you click a button, you you you see just to see the output. I need to navigate to the output, return back to this this Texturion or whatever. This is this is not this is not possible. So, yeah, that that's what I was talking about.
Shay Nehmad:I guess the other side of that is if we have any front end ish developers, or even if you can impact the front end roadmap in your company, making sure that software is more accessible and having someone test the software with, like a screen reader tool would be very, valuable. And honestly, low stakes. Like, it's not that hard just to get it prioritized at some point. Right?
Dominic St-Pierre:Yeah. The the the amount the the amount of yeah. It's it's pretty it's pretty simple. Just use header everywhere too. If you are if you are repeating some element, just use h four, h five, whatever.
Dominic St-Pierre:If you are tempted to use a strong element in the HTML, just use an h something, and that will immediately just fix, like, 95 of of the accessibility issues. Yeah.
Shay Nehmad:Good to know. I use Vimeo, which is a thing that lets me click on elements in the browser without a mouse, just for like I don't need it or something. It's just for perceived speed and also makes me feel very leet. And the amount of websites where it's like, oh, you have a fancy React button, but it's not actually a clickable element, so my Vimeo can't fight it. Very frustrating.
Shay Nehmad:But for me, it's just like a momentary frustration, and then I just grab the mouse and I click on it. It must be way harder when it's just literally blocking you from doing an interview.
Dominic St-Pierre:Oh, yeah. Absolutely. And and, you know, take take the Stripe dashboard, for example. You know, they they and at some point, I can I can understand because they have so many things out there, but this thing is is very hard on on the accessibility side? So so it's not it's not just interview, but but, again, we are not there to to fix all the the issues with accessibility accessibility.
Dominic St-Pierre:But let's just say that, you know, there's still a lot a lot of work to do. But but, again, you know, if you if you are interviewing a a blind person, just, you know, just ask them, you know, how how would it be easier for you to to conduct this interview? You know, don't don't impose anything. That that would be immediately extremely helpful. I would.
Dominic St-Pierre:So I I, you know, I would I would be fine just to share my screen. For example, let me stay in my editor. Let me stay in my in my console. Let me stay with my when I say my environment, I mean my desktop environment, if that makes sense.
Jonathan Hall:Sure. I think we should expand extend that courtesy to literally everybody. I mean, there's a lot of
Dominic St-Pierre:Oh, yeah. But that's yeah. That is sure. Are a lot of people
Jonathan Hall:with invisible disabilities or, you know, maybe not even disabilities, but just personality traits or neurodiverse people or whatever who aren't comfortable in certain environments. And trying to shoehorn everybody into the same interview process, I think, could be very harmful.
Shay Nehmad:Generally, when you work on accessibility of a thing, it becomes better for everybody. Exactly. I see it now when I compare, like, normal playgrounds from when I was a kid to the playgrounds that my kid plays on now, which are usually wheelchair accessible. It's just wet better because they need more land and the rides are cooler Yeah. And more kids can play there and less stairs.
Shay Nehmad:Like, it's just better for, everybody. Yeah. Like, I'm sure that people will fix the accessibility of their software. It will turn out with the better looking UI anyway.
Dominic St-Pierre:Well, yeah. It's it's possible. But but at least it yeah. Yeah. Of course.
Dominic St-Pierre:You know, you cannot you cannot have have a a click event on the dev, you know, but but we we are seeing that less and less, but it's it still exists. I mean, the amount of things that I I just cannot click is is still is still, you know, substantial.
Jonathan Hall:So you've been doing the podcast for a while, and I I saw recently that you are looking for a cohost. I'm curious to understand why and, you know, what's what's led to that, and what are you looking for in a cohost?
Dominic St-Pierre:I'm I'm trying to bring some kind of diversity to to the topic. Like I was saying, it's it's mainly me at the moment when I have something I I worked on something or something happened or I saw something that so I think I I I was kind of enjoying myself the the GoTime way of having multiple, you know, co hosts that can can, you know, jump in at some at some point and bring bring bring an episode and whatnot. So so I'm I'm trying to refresh the formula because for me, first of all, because I I I still want to to continue publishing Go podcast, but I don't know. I I need some excitement. I need some some new things.
Dominic St-Pierre:So I think I think it will it would kind of bring some motivation. So what I'm looking for I I post that on Reddit, which which strangely enough, I I I started to add some some traction, and for some reason, the post disappeared. So that that oh, yeah. Well Weird. What what can you do?
Dominic St-Pierre:Very weird. So I'm I'm, you know, I'm I'm not super sure about what I'm looking for. I I I thought my, you know, my my thought process was something like I was wondering, you know, first, I would I would start, you know, just having a quick talk with with the with the person and we could do one or maybe two episodes, just, you know, pilot episode maybe. And but, you know, I I was kind of looking at different angle than me. What I what I'm what I'm saying about that, I'm I'm mostly building, you know, SaaS software software as a service with Go, mostly independent and whatnot.
Dominic St-Pierre:So, I mean, wouldn't that wouldn't that hurt, I think, to have someone that is is working the the nine to five in a in a bigger org? That that that could be interesting. But, yeah, I I was mainly looking at multiple profiles, if that if that makes sense to because it's it's very hard to to ask someone, you know what? Let's jump in on that thing each week. It gets it it yeah.
Dominic St-Pierre:It's not everyone that wants to commit to that kind of of work schedule for a hobby thing.
Jonathan Hall:And have you found any likely candidates? You said you were getting some traction on Reddit. Did it get come to the point where you have some candidates now or you're you're still early in that in that process?
Dominic St-Pierre:Yeah. Yeah. There's a there there's someone that that we will record an episode the next week, actually. There's there's discussion with with someone else. I don't want to name names because, like I said, it's just pilot for now.
Dominic St-Pierre:But, yeah, it's it's you know, it's I I think it's nice. I think the audience that I have, there's there's a couple of people in there that that says you know, especially with the new year, people are
Shay Nehmad:saying, yeah.
Dominic St-Pierre:You know what? I, you know, I could see myself trying to do you know, to talk in a podcast. So that that's why I'm asking people to commit themselves to a weekly thing. It could be it could be just an ongoing, you know, if they want to do one episode per month or per, I don't know, six week, whatever. So that that would be fine with me.
Dominic St-Pierre:So so that that is mostly so, basically, it's it's like having multiple interview interviewee that would that would come back, but, you know, they would they would be closed, basically. I I don't know if that makes sense. It it it seems looks like it could be easier, because, you know, for for example, John Arendall, you know, came, three or four times. There's there's a couple of people, that I interviewed so far that, returned, and I find I found that I I find that very, it it's more
Shay Nehmad:easy. Yeah.
Jonathan Hall:Makes sense.
Shay Nehmad:Other than the podcast, recently, you mentioned you revived, sort of a super base competitor thing called Static Back End. How's that going?
Dominic St-Pierre:Yeah. Yeah. It's, it's an old project I started, and strangely enough, I I I talked twice with the CEO of Supabase actually about that. So I when when I launched that, when I open sourced it, it was basically it seems like it was at the same time that they did. I did not knew them.
Dominic St-Pierre:I just knew them when they when they got their first 30 mil in funding, like, one year after. So, yeah, basically, it's a it's like a back end as a service that I built in Go. It's a it's a project that I I really love, kind of got I don't know. Energy wise, it was it was not really there at the end, so I decided to archive it. I was not able to get traction.
Dominic St-Pierre:I tried to get some kind of VCs at some point, but I I I don't know. It's it's just a passion project. So, yes, I decided to revive it. I think that with the current state of things with people more and more people using LLMs and building prototype that are hosted on, you know, on cloud whatever or OpenAI, I don't really know that those things. But they seems to need they seems to need some some kind of, you know, instant database or back end as a service because those things are not really hosted anywhere, but they still want some persistent.
Dominic St-Pierre:So I I said, you know what? They might be there might be a good timing at this point in time. This is something that I will try to do. I will, you know, I will try to test. I don't know.
Dominic St-Pierre:So it's a, you know, it's basically it's a paid it's a paid hosted services with the fully open source MIT project.
Shay Nehmad:And, like, if I wanted to so if I wanna check it out or or find it, it's I assume it's on GitHub. That my use case is I'm looking for a super base or pocket based, like, alternative. I'm looking for users and auth? Like, what does it provide?
Dominic St-Pierre:Yeah. Exactly. So so there's there's, you know, user authentication, of course. So there there's the the database aspects of the real time as well. So, basically, it's it's like an event driven system a little bit if you if you think about that.
Dominic St-Pierre:So you can react to to a database record changing or something like that. And there's a there's something really nice, I think, that I built, which kind of is similar to what function as a service are. But I'm I'm using Goja, basically, and I built a runtime around that. So when you are writing some kind of, let's say, quote, unquote, server side function in that in that thing, you have access to to a lot of API that that are there that, basically, you know, call the back end functions in in in the same process. So let's say you want to send some emails, you will not do that from the client side or you want to, I don't know, to to send some some kind of or or trigger anything.
Dominic St-Pierre:You you can you can just create some function, but they are not real function as a service. Like, they don't need they don't need a programming runtime. So they are all hosted in memory, if you will. And and that that that is nice. So, yes, the website is staticbackend.dev, and this is on this is on GitHub as well,
Shay Nehmad:for sure.
Dominic St-Pierre:So if anyone wants to, to check it out. So, started to get some, some small small traction at the end, when when I archived it. It's in the last two years there. There's some, contributors that join at some point. Yeah.
Dominic St-Pierre:It's pretty friendly over there, and I think the product is is kind of made mature and whatnot. So so, yeah, it's I I thought it was a nice go project at some point, but, yeah, it it was maybe maybe the gophers are not are not the target audience at all. So I will I will probably focus more on the the the front side people, which which is you know, makes sense at at the end of the day.
Shay Nehmad:Mhmm. I I think a lot more people are becoming, quote, unquote, full stack developers, or at least the expectation is you can run across more of the stack. And having something like static back end or, you know, their competitors is nice. I like that it's like a one file single binary. It's very hostable.
Shay Nehmad:I guess it's very Go ish of you, but that's a very big benefit of getting like a fully hosted backend on one end, but not like a huge cloud, you know, subscription that's super vendor lock in on the other.
Dominic St-Pierre:Yeah. And I I think one other thing that I did that you don't see that often in in any kind of project, I would say, you know, cross languages is that you choose the data store at the end of the day that you want to use. You want to use Postgres, you use Postgres. You want to use MongoDB, you use MongoDB. Want to use SQLite, you use SQLite.
Dominic St-Pierre:So you don't see that often, to be frank, in especially in in Go project to have, well, you know, database slash SQL, of course, but the fact that Mongo is also there, this is this is the the game well, not the game changer. The this this is not something that you are seeing often because it's a totally different for a model to to to store data in in Mongo than than that database slash SQL in Go. So yeah. The the interface for the data data store is, like, 35 function long. So it's pretty dirty, but but, yeah, it's nice.
Jonathan Hall:Cool. Well, Dominique, it's been fun learning about getting some insight into your your project and and your podcast. As we mentioned before we started recording though, we'd like to ask a stumper question. And I don't know, we've mentioned it to you and you were like, so I don't know if you had a chance to think of an answer yet. But this year we're asking our guests what is maybe a favorite?
Jonathan Hall:It doesn't have to be the single most favorite if you can't you know, narrow it down. But what is a favorite third party Go library of yours and why?
Dominic St-Pierre:Yeah. I will I will need to go with SQL SQL C for me. Yeah. Yeah. Being a database person myself and I I like to work with data, I I was writing my my database code at first manually.
Dominic St-Pierre:It took a long long time for me to switch to not not maybe, you know, or adapt SQL c. And since that time, which is kind of recent, but but, yeah, in in the last maybe three or four years, this is mainly my my go to my go to way to to build any kind of database aware program these days. So, yeah, I cannot cannot see any kind of library that that that I'm using more than that.
Jonathan Hall:Nice. Good answer.
Shay Nehmad:I love SQL c. I love, love, love SQL c. It's such a good the way I picked what I unfortunately use TypeScript in the back end right now. And the way I picked how to, like, the way I justify not using an ORM was, hey, I use SQL C in Go projects and it worked really, really well. Let's just try to find the closest thing we can in TypeScript instead.
Shay Nehmad:And oh my god, it's such a good approach to just write SQL and and for when you're working with databases and not try to write your whatever programming language for talking to the database.
Jonathan Hall:I tried SQL C on my oh, go ahead. You you go first.
Dominic St-Pierre:Oh, no. No. I was just going to see. I I I've seen that in in the closure world with with UGG and now Gleam as squirrel, which which is, you know, basically the same kind of idea. By the way, I I'm I'm a huge fanboy of Gleam these day.
Dominic St-Pierre:I don't know. But, yeah, the this this idea of Sequel is is kind of, you know, it's it's at some places other than Go, for sure.
Jonathan Hall:Yeah. Well, I tried Sequel C on a on a personal project a couple years ago, I I suppose. And I I I like it better than an ORM, but I kept bumping into it so many times. I ripped it out, I'm so much happier without it. Oh.
Jonathan Hall:Yeah. That's a that's a hot take for another episode, I suppose.
Dominic St-Pierre:I guess.
Shay Nehmad:Well, thanks a lot, Dominique, for coming on the show. If people want to find the podcast, they can find it at gopodcast.dev. That is gopodcast.dev.
Jonathan Hall:And if they wanna join the podcast as a as a cohost, I suppose they should reach out to you probably at the same place.
Dominic St-Pierre:Right? Yeah. Exactly. Yeah. There there's also the the the Slack channel if if people wants to talk to me or send me an email.
Dominic St-Pierre:My email is is very easy to find on Internet, but it's basically demericsaintpiare@gmail..com. Sorry. Send some French word here.
Jonathan Hall:And the the Slack channel is on the Go For Slack, same as the Cup of Go Slack channel, but it's go podcast. No kebab kids. Just all one word, Go Podcast. You can go hang out over there and
Shay Nehmad:Much easier. Why didn't we do that?
Jonathan Hall:We should have done that. Too late now. We're stuck forever.
Shay Nehmad:Too late now. Well, we'll wait for Go To. When Go To comes out,
Dominic St-Pierre:we'll change
Jonathan Hall:it. Alright. Right.
Shay Nehmad:Thanks a lot, Dominique, for coming on the show.
Dominic St-Pierre:Thank you.
Shay Nehmad:Program exit up. Goodbye.