Cloudflare Analytics Engine, Workers + more with Ben Vinegar

Hey, guys. This is weird. We were we were just in a Slack channel together, and now we're recording this.

Yeah. Sure.

My name is Ben Vinegar. I've been working in software and web development for, painfully, like, 20 years. I I really don't like how that number has gotten pretty high lately. And, quick bio.

Trying to retread a little bit of that last episode. Mhmm. But, you know, I got into software to make video games. That was my original original career ambition.

Didn't really worked out. Got into web development. Was an early employee at the startup called FreshBooks and then discussed doing embedded comments.

And for the last eight and a half Yarn, I've been at Sanity Wes I've done everything from work on our client SDK development, specifically the JavaScript stuff, working on the product UI, building the product. And then I also kinda became our 1st engineering manager and and had a stint as our VP of engineering for 4 years, and now I work with you all. Yeah. Now you get to,

heard the heard the podcasters.

So, yeah, it's a nice fun. Yeah. Yeah. And do you wanna do you wanna give a shot at doing a, a century,

presented by? Oh, man.

Well, this I don't know. This is getting a little a little questionable here, but I'll I'll give it a shot.

Look, this podcast is brought to you by Sanity.

If you're a software developer and you're pushing things into production, your software is gonna break. I've been building some software lately and it it broke a lot faster than I was expecting, frankly. It's a really harsh reminder of how easily that happens. And, Sentry is just something that helps you understand how your software is performing in production, whether it's telling you about errors and exceptions that you have, or telling you that your website is slow, gives you tools to just sort of become informed, really, which is the most important thing. And then gives you tools to remediate those problems and and get your software back on track. I love it. Very well done. If you want to give Sanity a try, head on over to century.i0forward/syntax.

Sign up and get 2 months for free. Man, that was excellent, Ben. Thank you. I mean, it's super reasonable considering, you know, you've worked at Sentry for so long now. So you know the product.

I do know the product. I've also heard you heard you guys say that say a version of that line probably, like, you know, 50 to a 100 times at this point. So You've given us notes. More than that. I Yeah. Like, how many times? Like like, Century sponsored the podcast for,

Oh, yeah. But I said I've heard. Oh, you've heard. Yeah. Yeah. Yeah.

Yeah. I I think even some background, like, I'm always interested when people make new stuff in this industry. And when workers first came out, I thought I thought those were super neat, and I played around with that. I wanna say in the summer of the pandemic, I think I went Deno. This could have been like 2020. Just sort of interested in in Wes new infrastructure comes out, like, what what does that make possible? Mhmm. Right? I kinda relaunched my my my personal website, which is no. I'm not gonna plug that. We'll see. Yeah. Let's what is it? Benz.ca.

Yeah. I mean, I think this I think this even goes back to syntax, which is, you know, working with you all, wanted to do more in HTML and CSS, and and and and, you know, revamping the website really early way to get started with that.

And I I didn't wanna use Google Analytics for a variety of reasons. For example, Century stopped using Google Analytics. We've stopped, using marketing cookies on our website. We've written about that, and I kinda wanna do the same thing.

And I was looking at, you know, okay, if you don't want Google Analytics, how do you actually know how many people have visited your website? Turns out Vercel solutions are commercial paid offerings, which, you know, Scott at like, I don't know, 5, 10, $15 US a month, which seems, I don't know, a little look. My website gets, like, 30 visits a month. Yeah. Right? Okay.

Just to be clear here. And something about just sort of spending, you know, more money than my hosting provider. Right? Like, I'm not paying anything to host the website, but to actually host my analytics, that Scott money. And I don't know. I I mean, those Yarn those commercial products. Fathom, Plausible, they're good products. I wanna be clear, but they do charge money. Mhmm. And I was just sort of curious why there wasn't, you know, just something sort of easy to use. Now I understand there is some baked in analytics with Vercel and other platforms and stuff, but, you know, I was just using some some old cooked website provider and I wanted my own thing that I could drop on my page. Yeah. You don't wanna be,

I don't know, reliant upon your host just for your analytics, at least in my opinion, especially

you wanna have it somewhat portable. Right? Yeah. I I wasn't I, you know, I wasn't using a baked in solution. So even if even if I don't even think my head space was there, Scott, to be like, oh, I'm gonna use this built in thing. I was just I just totally went to what is an external Google Analytics thing that I could drop.

So that's the preamble to I was lamenting this problem with, some other, you know, early century people. And, you know, one thing about centuries is that we ingest a lot of data, and I've learned a little bit about the challenges of that. You Node, what what a high throughput ingestion stat kind of looks like. And I am not an expert to be clear.

You know? I just I just wave my hands a lot. But I do know a couple things. Okay? Which is if you wanna ingest a lot of data really quickly, you wanna write that immediately.

You don't wanna have to, you know, have an endpoint that handles data. Node you Scott query a database. You just wanna write it and then let some other job deal with it later. So usually that's a queue right. So this is often why you'll see Kafka rolled up into sort of or or similar kind of like, systems like that to just sort of accept data into a big queue. And then you got a bunch of workers that work on that data, write it to a database, maybe do some some kind of computation or whatever.

This is a long preamble. I'm sorry. But No. It's interesting. I was interested in sort of, like, what would an open source kinda, like, analytics thing look like, but I knew that those kind of systems need to ingest data quickly and at scale.

And there JS no way in hell that I wanna learn how to run Kafka or how to run all this complicated infrastructure. Yeah. You know? Riffing around, I'd be like, you know, it'd be really neat if somebody if there's just some hosted service that understood this pattern, that would let me just sort of ingest data at scale and write it, and I could read it back in kind of a in kind of a queryable way. So I started exploring tools online.

The story here is it began with a problem. I don't know that I necessarily started with Cloudflare.

However, I do like I do like Cloudflare stuff, so I definitely went there pretty early. And then I found this service called Analytics Engine, which was in beta at the time.

And I didn't quite understand it when I first saw it, but effectively, this is what I just described. It is a system that lets you write data, write data events onto onto a schema that they've defined, and I can come back to that. They queue it up. They ingest it. They write it into their own ClickHouse, managed ClickHouse

database that they're running. So it is ClickHouse?

That's what they put on their docs. As far as I understand, that's the underlying technology where all the sort of, you know, all the all the event data is written. And what is ClickHouse

Yeah. So ClickHouse is a columnar columnar.

Here's another one to you know, how do you pronounce this? You know, column store database.

It's an open source project. It was spun out of Yandex several years ago.

Sanity, and I think Posthog, like, a bunch of a bunch of products are built on this, analytics products.

It is if you think about the name, it's it's ClickHouse. It's designed to store clicks. This is it is like an analytics sort of, database. Mhmm. What's different about it from MySQL or Postgres or those sort of traditional relational databases is that the data is stored in columns instead of rows. Mhmm. If you think about when you're querying a Postgres database, you're you're kinda saying like, Vercel.

Select star from this table, and then okay. This is the the most basic query, ever, but postgres will basically just like or the underlying database engine will sort of, you know, scan every single row, collect all of those. Right? Now let's say that you wanna select 1 attribute. You're like, oh, select browser name from clicks from Postgres. K? Postgres still has to go through every row to accumulate those names. Mhmm. Does that make sense? So, you know, at the end of the day, there is still a file system. There's there's or how this data is stored in memory is that it has to rip through every single row of data. But a column store database is different because it's storing sort of all the unique column events in a row column oriented. So if you wanted to sort of figure out every single unique column name, you know, this would be great with some with some visualizations on the screen. I don't know if this is making sense, but, you know, it doesn't have to go through every row of the database. It just has to go through the unique values of that of that column. This is a real deep dive into this right away, and I hope I'm not butchering this whole ESLint. No. It's super interesting. Yeah. We first heard about ClickHouse when we heard that Century was using ClickHouse. And that was the 1st time they crossed my my radar in general.

But I've never even, like, really thought about a column oriented databases being a thing. Is is it specific you said it is like an analytics, like, widely used for analytics companies. Is it something, like, that's practical to be used outside of analytics?

I think so this is the like, I have not really used this very much. I'm more this is a really tricky thing where it's sort of, like, I I guess I sort of managed a team that's worked with it, but that doesn't mean that I know, you know, it's API or or exactly those details.

But it does what I understand is that it does have a SQL like language.

You know, this is the part where a producer could pull this up and verify this for us.

Bos, basically, how you work with the class is very different from how you would work with MySQL or or Postgres. It's not a situation where you can just, like I got this data and I'll just load it up over here and it will work work the same way with the same queries. You have to work with it differently.

It has different sort of advantages and limitations.

Is it used outside of that setting? It probably shouldn't be. It's really Mhmm. You Node, I I I'm not an expert, but I think you really would wanna derive some real benefit that you're getting out of using it. Like, it's not a real alternative, I think, for 95% of what web development is out there. Right. Does that make sense? Yeah. Cool. So I was checking out this analytics engine docs, and I think you know what? I think how even Cloudflare built it was more, like, hey, if you're building products on Cloudflare, you can track your own analytics from your products into the system and you can bring that back to your users. Like, you could use this for usage tables, you know, if you've got some kind of event based system or or who Node. Right? So they were thinking about it that way. At least that's how the docs kind of read. But I was looking at it and I thought, you know, this could be why couldn't we just sort of use this as a as a way of forwarding web browser traffic analytic events? It just kinda works. That's the cool thing about it. You know? You just kinda blast it with data, and it ingest it all. It stands up. It doesn't seem to really fall down, and then you can read it back. It does expose a Vercel like sort of query language that you can use to read back that data. And you can group it, you can sum on it, you can do a whole bunch of cool math on it. And that's really what is driving counter scale. It is a very sort of like, very simple application.

Couldn't maybe just like 2,000 lines of code that is it is a it is a script that you drop on the page. That script forwards events to a Cloudflare worker.

That worker writes to analytics engine, and then the dashboard reads back from that engine using a SQL dialect. Wow.

So analytics engine has a fixed schema. And what that means is and my understanding is that ClickHouse also needs schemas just like Postgres or MySQL. You gotta define what's in here.

So wait. Quick answer is you gotta write your own stuff. You gotta it it's sort of like a free form object store for you to store data that you want. Mhmm. So counter scale JS, like, the script is basically pulling browser device, stuff like that, and then pushing that into analytics engine.

But analytics ESLint, actually, just like the schema JS sort of it's it's like blob 1, blob 2, blob 3, blob 4, blob 5, double 1, double 2, double 3. It's basically just got free form text and number columns that you can use that they've already defined in their schema that are that are sort of fixed, but you can make them whatever you want. Right? I don't know if that kinda makes sense. Yeah. It's wild.

It's wild to think about. You know?

You basically have your own code do translation for what you want those things to be. Okay. Right? So when you write them, I happen to know that browser name is stored in blob 6. K. And Blob is just a free form string column. K.

And they have basically something like, I don't know the exact number, but you Node, 24 blob columns you can use, and 24 double columns you can use, and that's it. There are a bunch of limitations on this database that are like, you know, even an individual row I think can only be 5,000 characters. So it has a whole bunch of wacky limitations that you have to be aware of. But if you if the software that you're writing works within those constraints, then you get all these incredible properties. Right? You can write sort of near infinite data points. You can aggregate a 1000000 data points and read them back and be like, oh, a a 1000000 people visited my website yesterday, and it will probably return, you know, in a second or 2. That's pretty wild. Okay.

Well, this is trivia time, Wes. I wasn't I wasn't prepared for this.

I'm not pulling up the ClickHouse docs here. Right? But in my mind, these are effectively just sort of, like, they Yarn unders they're under under the hood. There are these primitive types that exist inside of ClickHouse.

Right? And double is some kind of number, integer, whatever. Yeah. And and blob, I think is it's either I mean, it could just be binary data. It could be a string. Right? I think it's equivalent to the Blob. I think there's a Blob column type in Postgres.

Yeah. That is similar. Right? So it's just kind of exposing these really raw primitives you can choose what to do with.

The number columns doubles, like, you can do some and average. You can actually do math. Okay. So it it does it does allow for you to do some SQL like computation inside of the analytics engine. So when you query it, you actually send it a SQL query. If I wanted for example, every unique browser name out of this thing. It's like select, and I happen to know that I've that I I personally as a developer have put browser name on blob 6, which is a string.

Then I could say, you know, select count, comma, blob 6 from this data set, and then group by blob 6. And then I'll get a I'll get in the results effectively, like, a a table output of every unique browser name and then the count. Right? Oh, okay.

Wild that you have to

use the blob blob 6 instead of, like, naming the columns. Right? Yeah. If you look at the code that I've written, it is it is a bunch of string, you know, shoving column names in there to make the code

look somewhat like it makes sense. Yeah. Yeah. I pulled up the schema, and it's, like host blob 1 user agent blob 2. It's like that name.

Gee, I don't know. Wes this an analytics engine or is this no. This is just on ClickHouse's engine or ClickHouse.

Yeah. In this case oh, you're right. And they have an pnpm.

Yeah. If you look up, there's, like, a SQL API for for this Analytics ESLint. Okay. They only support Blob and Double. They do expose this index column, which they describe as the index value that was logged with the event. The value in this column is used as a key for sampling. They also sample data. This is part of how the system works for you at scale. Basically, if they ingest a ton of data points and that and that data looks similar, they will, you know, sort of sample that data and they don't have to store every single data point. I'm getting a little deep here. I'm sorry. Oh, yeah. It's interesting. Yeah. The the sampling, we talked to Armin from Sanity on it, and we're just like, how

All these services use sampling. So, when I when I launched this project, you know, somebody came out of the woodwork and was like, you better explain that this uses sampling. And I'm like, okay. I'll I'll I'll put an issue about that. And the concern there, it just means, like, at high throughput, the system is gonna throw out some numbers or it's gonna make estimates is how I would kind of simplify it. But everything does that. Google Analytics is not storing every data point of every visitor that ever visited your website.

Right? It's not effective for that. It's not effective for them to do that. I also think this is why I'm sure you I'm sure you guys have encountered this. You ever have a situation where you have a bunch of sort of, like, analytics trackers or counters and and or or been in a company where it's like, why don't these add up? Yeah. Why is it that your system is telling me a different number than this system? And everyone's kinda pointing fingers at each other.

I don't know. Maybe I've been in a bunch of sort of meetings and stuff where that that's happened.

Yeah. The answer sampling. Well, it is a It could be among among a great many other things, but that's one of them. Yeah. So are these all decisions that were influenced by working on Century? I mean, I know Century and an analytics tool have a Scott in common. So

I think it's that so when I did look at Open Source Analytics trackers, like, I think Plausible is open source, and they run it as a as a commercial offering as well too. And so I I, you know, I poked at that source code just to sort of understand how they did it. And I think at least an early version of that was Postgres.

Right? So if you think about it, most people, if they sit down, I'm gonna write an analytics tracker, and you think about what you Node. It's like, oh, yeah. Okay. I'm gonna write this to a data write this to a database.

But what happens when that just sort of explodes with data? It's not really the best. You know, I I've seen enough from my work to to say, like, at low volumes that's perfectly fine. But at high volumes, you know, it's not really relational data, if that makes sense. It's not really what that database was designed for.

And while I don't really think this is gonna be a problem for the 30 people who visit my blog every website, I sorry, every month. I guess I wanted to just sort of explore what what a, like, a fire and forget system would look like. Like, could you just press a button and deploy this kind of software stack? And if somehow you showed up on the top of Reddit tomorrow, it would be okay, If that makes sense?

Yeah. Right? The hug of death.

Yeah. We talked about this on our on the last time I was on the show, but I'm a very lazy Vercel, and I or I I work hard to be lazy. And what that means is I don't wanna have to monitor my databases and sort of figure out how to back them up or figure out I just don't wanna do any of this. Right? Yeah. I just wanted a software system that I could put live, forget about it. And I've been running this for about, I think 4 months now, 3 months, and I've literally never had to do anything with it. So it's it's sort of fulfilled that from my side.

So yeah. I think I just kinda looked at this as like, what would be the system that would allow for this type of software to exist in a in just a fire and forget kind of way. And I think that analytics engine is

The way that it works, it is it is a worker and and a analytics engine. Right? So every time that, like, a user visits your website and sort of request that tracking script, the tracking script makes a request to a worker.

That worker does some really light, you know, data munging, and then it writes a data point to Analytics Engine.

So that's 1 worker invocation, and that's 1 Analytics Engine write per Analytics ESLint. Does that make sense? K? As I understand, on Cloudflare's free plan right now, you know, for free. If you just sign up, you get a 100,000 worker in vacations per day.

I think analytics engine also gives you you get a 100,000 rights as well per month on the free plan.

So, you know, in the magical world where you just got the perfect amount of data, or a 100,000 people visited your website every day, is anyone gonna need that? Yeah. That's a lot.

And and to be clear for the audience who who might not, you know, have have used a lot of these things, you don't have to have your website hosted on Cloudfly. This is just for the tool itself. The tool itself, counterscale, you host it on a worker.

So Cloudflare is only there to host the tool, which is free to use outside of your usage.

That is an important clarifying comment that I probably should have opened with earlier. Yeah.

It it's basically, like, 3 commands from the command line. Like, you clone you clone this, you run a bunch of commands, you generate an API key, and then it's just up there and it's just working, you know, and then you drop the script in your website. So, yeah, this is just referring to just this thing. Just counter scale. This is what it Scott to host it. Those numbers are absurd. Right? Like It's absurd. Yeah. Yeah. The

I don't work there. Yeah. However, I'm gonna I'm gonna offer my own random theories.

You know, one is that they actually manage their own infrastructure, by which I mean the hardware. Right? So they're not they're not built on AWS or Azure or Google, like, they are their their own kinda cloud provider.

So that's one. Another reason is that they've developed this kinda unique technology, v eight isolates, which is what this worker stuff is built on that is sort of quicker to run. And it also has, like, different performance characteristics, like, you know, spinning up a a serverless worker is really fast. And I most people who've worked with, like, Lambda's or or GCP's, equivalent thing. GCP functions JS, you you know, they're spinning they're spinning something up, a different kind of environment, and it's slower.

And then I think the third thing is they can have a set they might have a completely different strategy of being, like, super cheap for hobbyist, you know. Yeah. A lot of people say enterprise is paying for it. Yeah. I have no insight into that to be clear. Like, that could be true. That could be false. I'm just I wanna be clear here. We're not in a in a in a complete love fest for this for this infrastructure.

And then I think the 4th thing I would add is this, is that I just I don't think the DX of Cloudflare, you know, is as good as Vercel. Mhmm. Just, you know Not even close. Yeah. Yeah. Yeah. And and I I do think that there have been painful parts of of working on this. Like, the stack traces that I get out of the out of the workers are pretty painful, to be honest. Mhmm. And I've kinda dorked around with trying to get source maps working the way that I want. Counter scales also built in remix. That's another random thing. I'd never done that before. So I didn't have a lot of familiarity with that stack, and it's got its own it's got its own little server as well as I understand, or it had its own kind of build tool, which they they're switching over to Vite now.

Someone please correct me if I'm wrong. No. They aren't. That's true. That's all accurate. Everybody's becoming a Vite app these days.

Yep. You know, so I while building this, like, I was getting some pretty gnarly stack traces, and debugging stuff was hard. Just wanna be clear. Right? Yeah. But I think, you know, for me, I think in Node ways, this was a little bit of a an exercise and a and a and a mission to see if it was possible, and I was willing to sort of put up with those those challenges,

but I've always had just a, a fair enough easy experience with adapter patterns. Right? You just push it live. I I think I I wonder a lot about that Hano JS since Cloudflare.

Maybe. Pretty interesting stuff. Yeah. Go ahead. Let's take a detour here into, Syntax and Sentry 1 year later.

You Node, Syntax, we joined Sentry last March, end of March.

We did our our big announcement on our 600th episode. We had David on to talk about that.

And it has been 1 whirlwind year of a lot of stuff. Half of maybe not half of which, a quarter of which Ben Ben was, on sabbatical for. But when he when Ben showed up, we were kind of I don't wanna say directionless, but we were a team of, you know, folks who have been doing this all very, I don't know. I don't cowboy coding type of deal. You Node, we're we're working on the site kind of we're very siloed. We're working on our projects. We're working on our site. We're used to being stallions, as Wes would say. And we we definitely needed a a tamer in in house, and and Ben kinda showed up and whipped us into shape. And it ever since Ben has showed up, it it feels like the amount of progress we've made as a team has just kind of blown blown up completely. So, Ben, do you wanna give, you know, a quick overview of your role on Syntax and, like, what you do for the team for people who don't know? You know,

my title is general manager. So, technically, you know, I I I like to describe myself as the as the corporate wrangler. The the sort of go between, between century and syntax. And I think it's worth talking about that because we you talk about on the show, but probably never heard it from someone like me that's, like, you know, we we run syntax pretty independently. Wes and Scott and the team run, you know, content and figure out what they wanna talk about, what they wanna write about. And I think that, you know, going back to my Node, my role is to sort of empower you guys, these guys. I'm like, who am I talking to here? You know, to just kind of achieve what they wanna do. So, gee, what has that been? I mean, I think an early moment for us was getting everybody together and into an on-site, and really just talking about, like, what do we want this to be in a year?

You know, I'm trying to set the stage for listeners here, but it was like you know getting into a room talking about what we want to do and then and then working backwards like what would we need to, what we need to do between now and this kind of, like, result that we wanted. We wanted to do some more stuff on YouTube. We wanted to reach more people. We wanted to make better episodes. Right? Mhmm. And that turned ESLint, hey. Well, let's let's hire a producer. You know? What does that look like? Yeah.

Yeah. No. I I think that that initial getting us into a room and talking about, you know, actual tangible SMART goals or whatever Yeah. For us, even though you didn't come to us and say, hey. We need SMART goals.

But it it was more or less like, you know, what what can this be and where can we get? I think that was the biggest thing for us.

But, you know, one thing that I think we've all been very appreciative on your end and on centuries end in general is that there's been a a distinct do not mess up the vibes angle for from everyone. You Node, we all know what could happen if the vibes are off or or or if, you know, corporate overlords do come in and and kind of ruin what's going on. But I think that has been a focus of century from the Scott, and, specifically, Ben has always been, I do not want to mess up the vibes. And, I think you've done a really awesome job with that. So, you know, hopefully, the audience feels the same way that the goal here is just to take what we are doing and do it better and and scale it up and be more efficient.

What did you do to not mess up the vibes? You join a you join an established media product.

Yeah. You don't mess up the vibes. It's it's it's good that we started this episode with by the way, someone fell through, so we threw Ben in here rather than, like, here's the apology tour.

Right? Yeah.

I swear Wes didn't ruin it.

The way I think about it is I feel like this is like the trope movie thing that you've that we've all seen a 100 times. This just feels like you've seen this situation depicted many times where this kinda great thing joins a bigger joins a bigger company.

And and then and then people kinda meddle with it. Right? And, you know, sometimes I maybe could see that perspective because maybe there are a bunch of reasons why they feel that they need to meddle with it. It's like, just let them cook. I don't know. Yes.

I'm trying to it's, I think, you know, there's there's a lot of, like, don't fix what's not broken. Right? And so, I guess, I don't really have a coherent answer for this, Scott. I'm sorry.

No. No. I think that Just Let them Cook describes it all. I mean, if we if we go back to our tried and true Wes of, method of using food metaphors for everything, you are basically the general manager of, of a restaurant here, and a restaurant here, and Wes and I are cooking up. And now we we Scott another chef in. We got a, you know, expediter in. We got all the, we got all the parts ESLint, and you're the Node ordering up the the food and making sure we have what we Node. And yeah. Yeah. Like,

Mhmm. Yeah. I forgot about the store. That's 1. Yeah. Yeah. You know, thank you to all the people who bought from the store, by the way. And if I emailed you and told told you that sometimes the shirt was laid, I'm sorry about that. I'm sorry.

Okay. My bad.

But we tried our best. Yeah. That fulfillment software stuff was a wild task, and and I think Wes and I would have been very overwhelmed very quickly handling trying to handle all that stuff on on top of doing the show. Personally, though, for you, how does it feel like going from a software project to a media project? Like, what what how different is that? That is a very good question and probably one that I get all the time when I explain to people what my job is, which usually takes, like, 10 minutes,

at at parties and so on.

I'm like, oh, what do you do?

I'm on trying to avoid going, like, crazy deep on this. I think that, you know, my experience as a as a software manager and as a senior engineering manager person is finding really great people, and then getting everybody aligned on ideally on, like, what they wanna do. K? It doesn't always work out perfectly that way in engineering, but, you know, as much as possible, try to figure out what what people are interested in and try to find a great fit for them. And then but not really get in the way. You know? I it's kinda strange because I I wrote a book on JavaScript, so I, you know, I I certainly have opinions on, you know, how many spaces there should be in a tab and and codeters and all that.

But those are really minor very minor details that are really not gonna affect anybody's success really. Really just kind of the big picture stuff.

You know, that's what I would be doing with with with with software people. And I think that that idea, I don't find is too different from what you all do. So that's why I have very little to say about sometimes micro criticisms on on the shows. Although, sometimes I do, I share some feedback, but mostly from a, hey, what do you think about this? I guess what I'm trying to say is I don't know if for if for everybody this would make a lot of sense. But I think for the way that I like to work with software development teams, which is really high level, hold people to goals, get people to match them with their interest as much as possible, like, does that sound like software engineering, or does it sound like just sort of general, you know, working with teams? I think it's more of the the latter.

That's a good point. And and even just you expressing all of that made me realize how similar that is because that is all all you're doing with us in a a very positive way. So yeah, I I would never put that together, but yeah.

And not too different. Yeah. I mean, there's a lot of you know, strangely, I have a lot of anxiety about being on this podcast, but first but I have no I have no anxiety about blasting open source code.

You know? Yeah. So there's there's certainly some differences. I wanna be I wanna be clear.

You had to download TikTok.

I had to. Yeah. I do think that Node thing, working with you all that is and and working in, I guess, more media is look. I think the software developers love learning. I Think there's something it it it's kind of the job. Right? The technology is always changing.

That's part of why this exists.

The techniques are always changing, like the tools.

So I think, you know, to get into, podcasts and YouTube and that kind of stuff, hey, that's learning too. At least it is for me, and I I enjoy, you know, I've learned so much working, with you all. Like, I'm I'm really the student here to be frank, and to level up my CSS as I've as I've commented earlier.

You know, it's been a lot of fun. So I think, you know I don't know. I know there's a great there's a personal answer for me is that I've I've personally really enjoyed just kind of learning new skills just as I think many software developers would. You know? So

I don't know. Nice. Oh, one last1.

How do you sell syntax internally to Century? I I know a lot of people might be wondering about, like, is it clicks? You're tracking clicks? Like, how do you how do you sell what we do here to a company that has very real, income from products?

Yeah. That's a super good question. And I think even maybe a chance to, you know, offer some transparency on that. Right? I think that one thing that I learned from being a senior software engineering manager is internal storytelling is really important.

So a lot of things I would do with software teams is really encourage them to, you know, talk about their successes whether we, you know even anything from upgrading, like, a really gnarly dependency to, you know, optimizing some code.

And we still do that a lot internally at Sentry, and I think you can see that even on Twitter from the the software people or the product people who work on those things. They often talk about it still.

And so I think that that's that's it's very similar to what we do. Right? Our goal is really just to to grow the reach of the podcast, and, you know, lightly sprinkle some century on there, which which we do during the intro.

So, really, like, most of how we communicate the success is really syntax's success. Right? If we're doing great on platforms, if we're if our listeners are up, which they have been, we just sort of tell that story internally and try to explain, you know, what is it that we're doing that we think is driving this. We don't have marketing leads or, you know, hard numbers that we're trying to hit.

I think the company is really just sort of pretty down with look. We're pretty transparent and clear that Century produces the show. The show continues to be good. We hope that they the audience recognizes that, and that's that's what we're going for. Mhmm.

Beautiful. Scott and West's wild ride here.

Yeah.

Yeah. We have Caitlin rocking the newsletter and,

making sure everybody knows what we're doing. Yes. I think even as a even to to bring this, like, full Vercel.

We started with this talk about counter Scott, this project, and, Cloudflare stuff.

I think something that I left out of that is, you know, not only did I just kinda wanna make this for my website, but I think in working with you all and wanting to, you know, have a great show and to have great education about what's going on in our industry for developers, like, as a general manager, I feel I feel the need to, you know, invest my time and to work with new tools and just try to reactivate myself into what's going on. And I think that that was a little bit, you know, a small driver and working on that project. So I do wanna kinda add that. Yeah. Absolutely. And I think that's I mean, it's very fitting with what we're kind of about here.

Node of the episodes, I think, that's coming out, it's coming out very soon. It's out by the time you listen to this, I think, is is all about you have a wild idea. How do you quickly explore that idea? And that's, like, very much the feel of, like, what we're trying to do around here. So love that. You Node, beyond just managing the team, you're diving into what we're doing in those regards. But also you're making videos, you learn video editing, you learned the social media platforms we're pushing out to. It's like, I think that all all really, really helps.

Being invited onto a podcast an hour before Oh, yeah. Wes recorded it. Yeah.

We need the syntax you know how there's the epic stack? We should have the syntax stack, and that also includes, you know, all the video and audio, you know, descriptions, tools.

Warp.

Synstack.

I was just thinking of this one because we warp talking we were talking about the syntax stack, and we're talking about Randy and and production. And I think it would be fun to even have Randy talk about microphones and stuff. Right? Like Yeah. Look, here's a little preamble to the to the sick pics, but couple decades ago, you have to buy nice suits to to go into the office or or nice wardrobe, blouses, etcetera. Right? Now mostly a lot of people are working remotely from their homes. And on the topic of this sort of stack and and audio production, like, you know, Randy's talked about, you know, $50 microphones that are really great. You know? Like, there there are tools out there that could kinda like raise your professional acumen, inside of your company. I think that's interesting.

And on that note, I guess, because I was thinking about this. There's a software that's yet to be released called Boom. It's a it's an independent little company, here in Toronto. Getboom.app.

I think it's launching in a week or 2. And this is sort of like it's a virtual camera that lets you just sort of put professional overlays onto your onto your camera output so that you can just kinda like go into Zoom calls and and Google Meet, and you can have your title on there. You can have a logo or you can add some filters. Right? It's sort of Cool. Yeah. It's like a like, having, like, a Snapchat camera, but for the professional setting is is what they're going for. I happen to have a a beta version. I got an early access, and I was playing around with it with CJ actually earlier earlier this week. I think it's pretty neat. So that's that that's my sick pick. Little weird because it's not you know, might be out by the time this this launches, but,

Counter scale? Yeah. That's an easy one. I, and I really wanna preamble this. Like, this is this is really like a personal project that that that is not some super deep thing. I I, you know, I designed it for for my 30th 30 visitors a month blog to be clear. Right? But it looks great. Yeah. It looks in in functions. Great. Yeah. I think this plug is hey. Check it out. It's counter Scott dot dev. Is there's a website. You can actually browse the dashboard and take a look, and then there's link to GitHub in there. But if you think this is interesting, like, I would love for people to open up issues, pull Wes, try it, you Node. I really love for this to be a community project, not a me project.

So that's my sick pick. Awesome.

See you guys. See you see you on