S02E01: Restart, Refresh, Record

03/06/2025, duration: 14:42

Video

Transcript

Hey, welcome to the Debugging Dan podcast. I've been out for a while. This season, I dubbed season two, episode one. I've been out of it since December. I kind of not really burned out on building in public, building in general. So I didn't do any building for a couple of weeks there. And I'm trying to slowly start again. I've been building a lot of cool stuff and I want to share that. And I'm going to be sharing on Blue Sky and YouTube only for personal reasons. Since I found that sharing on a lot of different social platforms doesn't give me a lot of focus. And also for political reasons related to some other social networks. And this will hopefully lower some barriers for myself. Give me more positivity on sharing, interacting with people on those platforms. Which having only a few platforms should make it easier.

I'm also changing the content a little bit. In the past, I did a topic and then I did a get it done video. And now I'm kind of more freewheeling it. So I'm losing or dropping the weekly schedule. Depending on what I have to share, I'll share. And there won't be a real topic there like there was earlier. I'm just sharing what I'm doing, what I'm working on. And that's what I'm going to try to do for this season. And let's see how it goes.

So looking back at December, it was a pretty busy month. Both personally and at work. Because of the end of year stuff. And the holidays for personal reasons. And I also kind of started looking back on the past year, 2024, where I started the podcast. And I found that I didn't really make a lot of progress as a side project entrepreneur. So it was December and it did have a lot more projects that I built. But for all those projects, I didn't have users, no people paying, no real marketing done. I did learn about SEO and a lot of things about billing in general. But on Reddit, I often see posts and they're probably fake. I don't know how I made 1,200 MRR with my weekend project that I built in two days. And those are fake. But when you're in it for the grind, like most people are. And what I do read often is you just need to keep going, pivoting, try something. If it doesn't work, try something new. That's what I'm still trying to do. And I'm hopefully to get that break somewhere. And just have a project that I can really help people with, engage with people. That's what I'm going for.

In December, I pivoted the latest project that I was working on. So I was working on DanCuts, which was a video editing application. And I did it first in Next.js. So I learned Next.js as a framework. And then I found at some point that video editing takes a lot of storage. It takes a lot of processing. And I was not really up for investing in that before having it being a real product. So I pivoted that. And that was one of the last podcasts that I did, I think. I pivoted and I started building an Electron app. So I started building Electron. And at some point, I think in January, I got kind of out of my burnout. And I figured I want to share more. How can I create content that I can share easily?

And then I figured I might be able to share my screen, what I'm programming on. I don't really have the time and the focus to do live streaming because that should be longer periods consecutively being behind the desk or behind the computer. And I often do the work that I do on the couch or on the table in the living room where my family is around. And I often walk away from my screen or do other stuff. So that doesn't really fit with live streaming, in my opinion. But I can build something that captures the screen. So it makes a screen recording. And I figured and I checked with ChatGPT whether it's possible to determine whether the computer is idle. Then I can pause the recording. If the screen is locked, I can pause the recording. And if I can do that automatically, I can later on just stitch it all together and then create a timelapse, for example, of what I've been building on.

So I moved to doing that and figured when I'm done with that, I'm going back to Dan Cuts. It's in the related field. It's also about video editing. So I learned with Electron how to capture desktop or Windows. Of course, I also want to capture multiple Windows, multiple desktops, multiple screens even. Since on my own laptop, I have two screens on the laptop itself and one screen I use for previewing when I'm building and the other screen for coding. So I want to be able to capture those screens at the same time and render them into one window. So that's what I'm still working on.

But after that, I moved on to a different thing because as a sidestep, I was learning Diffusion Studio Core. That's an NPM package that offers video editing via code in the browser. So in Chrome, you can just write JavaScript, load a video, add some operations to that, trim the video, crop the video, add images, overlays, things like that. And it's a very, very cool package. I encountered that at some point via Reddit. Somebody shared that. And I started building with that. And then at the beginning of January, I believe, they shared, hey, we have a V2 version. We've changed a lot of things under the hood. And I was very enthusiastic.

But at some point I found out, hey, they also changed the licensing model. So it used to be open source, Mozilla public license. And now they changed it to closed source and they have a commercial license. So you're allowed to use the library for free on personal projects. But if you have something commercially that you're making money with, you need to pay them a license, which is a couple of hundred dollars per seat, I believe. And that doesn't really fit me because I want to do things commercially with the video editing library. So I had three options. I could just start using the latest V1 version, which still had the open source license. I could find something else. Or I could also take the open source code from Diffusion Studio and build on that. And I could also choose to build something myself.

And I chose the latter option because working with Diffusion Studio, I learned that they built it on top of Pixie.js, which is a 2D screen rendering library, which is also open source. And I figured I could use that to build my own things because there were some issues with Diffusion Studio, which I didn't really like, how they were doing things. I figured I could build something that fits me more. I already learned a ton of stuff working with Diffusion Studio. So I started on danvidstudio is what I called it. That's what I'm still working on. And I haven't copied any code from Diffusion Studio, so all that I've gleaned from it is by using their API, so I haven't really looked at their source code. I do know where I can find it, but I'm doing things myself together with Cursor.

So I'm building that. I'm adding unit tests. I'm adding integration tests. So before I create a build, I can open up my web browser and watch a couple of pages being loaded where the operations on the video are being executed. I can see if a watermark is being placed, if everything's working correctly, so that saves me a lot of time in bug hunting. And as I mentioned, I use Cursor now, which has been a real productivity booster. I use Cursor to manage the unit tests, to fix the code coverage if it drops below a certain level. And what I have found is that Cursor is really cool, but it also has a tendency to perhaps go off the rails and go into a real wrong direction or that it sends you off into a suboptimal solution, which you're not really aware of.

So in the thing that I did with building DanCapture, it pointed me in a direction for screen capturing with Electron, which is working, but it's not the best solution. So at some point I encountered some things while reading the Electron docs and I figured, wait, this is not what I have, but doing it like this is a lot better. So then I started changing that in the implementation itself that I found that's really something that you really need to take into account when using Cursor. It's not ideal, but it's an awesome tool that I really enjoy.

And I've been almost, I'm almost ready to release a first version of danvidstudio. So it's been working in the tests and I've been integrating it in DanCapture, so my screen capture software, where I've moved out the Diffusion Studio core stuff and replacing that with DanVid Studio, which has a kind of different API. I'm just at the point that I can run the application and then the composition, the video composition that it makes is still incorrect. So there are some changes in my API and some different behavior in my library than theirs that I need to figure out and then fix. So when I have it working in DanVid Studio, no, in DanCapture, I'm going to release it as a beta. I'm going to share it on Reddit, perhaps find some other people that are interesting, that find it interesting. And I'm going to continue working on it because currently I only have video support, which was the easiest with PixieJS.

And I still need to add audio support since, well, that's a bit more difficult to add since it's not something that PixieJS supports natively and renders on the canvas. And I don't need it yet for DanCapture because I don't capture any audio there. But that's some of the things that I'm going to look at once I have it done when DanCapture is running and I can use it locally. Because then I want to use DanVid Studio in DanCuts and then I need more functionality like also audio, for example. So that's what I'm going to be looking into after that. And I'm not sure yet if DanCapture is going to be something that I'm going to release as a product. But at least for DanCuts, that's the goal.

And yeah, that's what I wanted to share. So I'm going to be actively sharing more publicly or I'm going to try to. I've been trying to do that for a couple of weeks, but it's going off and on. I'm going to be recording more podcasts. I have a new setup here. I created a new backdrop. And I have a setup that I'm pretty simple to set up. So that lowers the threshold, which is always important for me, I found.

So season two, episode one. Stay tuned if you want to have some background information on this specific podcast and else like, subscribe, share with your friends and get the word out. So what I wanted to share after that is this is basically the second time that I record this podcast because I'm an idiot. And I'm doing this on a different computer than I did earlier. And I was setting up OBS and I think I thought that I muted desktop audio, but I muted the mic audio. So I was into like minute 13 or something like that. Then I found out that it was recording the screen, but it wasn't recording audio. So I was an idiot there, but I'm now re-recording it. And I think I normally don't practice the podcasts. I do have a list of notes here that I wanted to talk about, which kind of helps. So I'm probably going to be doing that more often.

But probably the wording here is also more flowing because I've already done it before. And I also got some chat GPT title suggestions. So I asked it, Hey, I'm going back into, um, back into season two of my podcast. I've been out for a while. Can you give me some title suggestions? And I really liked those. So I'm not sure which I'm going to pick yet. I think it's going to be called restart, refresh, record, but I also got some other suggestions without it, which I liked, which I didn't want to, uh, to keep you abstained from. So other title suggestions, we're back in the game. Here we go again, picking up the mic challenge accepted. So restart, refresh, record the comeback episode, like the comeback kid, but differently. I guess, uh, if that's a thing, diving back in my con, let's go round two starting again, but smarter and, um, reading this out loud. I really liked them. I need to pick one.

I think I'm also going to do a short one on a thumbnail. Um, so yeah, um, that's what I wanted to do more of a background thing. And, um, uh, I'll see you on the next one. Bye. Bye.

---

Let me know if you need anything else!