The closet is open

The AI is here and it’s here to stay, and being stubborn doesn’t pay off all the time. Not for me. I used to deny it and call it SLOP left and right, and now I run, every single day, an OS that came out of this same AI era, and I’m very satisfied with it.

So I tried to improve my past work with AI, and for a while that was indeed the shit. I did more, easier. Then, after long struggles and failed experiments, it hit me: I was doing more of the same. The same limited output I was producing before, just more of it.

And the biggest limit of all was that I never had time to do something for me. Exactly, literally, for me. Not for some client. Not at the job that pays me so I can then spend that money on something for me. Build something for me. I have a lot of itches and ideas, and 40+ years of parked ones piled up into an unbreakable closet: don’t enter, don’t try!

Now the time is fucking different. I’ve tried YouTube for about 2 years, and I know, because the numbers show it, that I’m bad at it. And it’s ok. I don’t aim to become the next big streamer or YouTube creator.

I just have my idea: do shit, share it with others. That’s it. It’s a thing for me, a thing I wanna do to carve my footprint into the Time endless concrete boulevard.

But every time, I hit the same walls: tools and software too complicated, some quite expensive. Fortunately I had the money, since my job is still well paid. Until I finally saw the wrong fucking thing I was doing: hunting for a tool, or a set of tools, to help me follow my dream.

So I stopped. I stopped searching and I started building. And for a very big chunk of it I did it wrong again. I was still using AI and trying to make it obey my ruling. Which is basically shit: it’s my limitation, but faster. That went on for more than 2 years.

Until I stopped using AI to speed up my imbecility and my limits, and went instead to that forbidden closet. I opened the Pandora box where all my ideas lay rusted and dormant, waiting for the idle era to end.

And that’s now. That’s fucking now. I decided to build my idea: a super fast, easy to use, straightforward, extremely highly opinionated app that does what I dreamed of. Not what I could build, only faster, but what I dreamed of.

I wanted to cut my video recordings fast. I mean freaking fast, no wait time and stuff. Fast, because it’s an easy task. And I wanted visual cues based on transcripts, something DaVinci Resolve and Adobe Premiere Pro do, but they come with shitloads of GB of data and other stuff that I don’t want. And I don’t need.

So here’s the thing: I managed to have this freaking awesome app, called Splicer, which I’ll release for the World to enjoy and take a bite at. But I had a fucking itch about one dependency: ffmpeg. A reputable, very well optimised library, huge history, huge adoption, and it was handling the encoding and decoding of my audio and video streams. But it’s fucking huge. Next to my app’s size, freaking huge.

Then I stumbled across FilmCraft, from the ArtCraft team: an open-source video editor written from scratch in pure Rust, a clean-room take on the Premiere Pro workflow, with its own codecs inside and no ffmpeg anywhere. MIT licensed, on GitHub, for anyone to read and build on.

That’s where I freaking hit the real peak of my Eureka moment. The codecs I need take a tiny fraction of what ffmpeg takes. So I did just that: pito-codecs starts from FilmCraft’s codecs, ported into my own Rust and tuned for what my app needs, and FilmCraft gets its credit in the repo, in the notice and here. Thank you, ArtCraft people. You saved me months and you didn’t even know it.

So freaking World, I give you these, all under one roof I call PITO MD:

1. Splicer: cut fast. Freaking fast. An extremely highly opinionated video editor for your simple needs. It transcribes your recordings in English and Spanish, so you cut by the words. The speech engine and its model ship inside the app, nothing to download after you install it. It runs on Mac for the rich ones (Apple Silicon, macOS 13.3 or later), Windows for the poor ones and Linux for the courageous ones (me included). Free to use: the first 30 days with everything, then exports up to 5 minutes with a small Splicer wordmark. Pro lifts that for $19.99, once, on two devices, so I can buy more coffee. Grab it at pitomd.com/splicer.

2. pito-codecs: super light codecs for everyone who needs lighter and faster codecs than ffmpeg. For what I need, it’s faster and far smaller. Open source, MIT, Linux, Windows and macOS, on GitHub now: github.com/gmrdad82/pito-codecs.

3. And since the benchmark charts came out so damn good looking for the pito-codecs repo, why not pito-compare: pcmp, a small MIT command-line tool that draws comparison charts. Not customizable, but freaking fast. Also on GitHub now: github.com/gmrdad82/pito-compare.

And that’s it for now. I’m gonna prepare a video for those with more time, to check it on youtube.com/@S0FTR3S3T, where I share candies for the eyes.

Gamer Dad out. Have a good one!