Constantine Fomchenkov

Here's my conversation all about @FFmpeg, the legendary open-source software powering most video on the Internet. In the episode, I talk with Jean-Baptiste Kempf and Kieran Kunhya. JB is lead developer of VLC and Kieran is FFmpeg contributor, codec engineer, and the person behind the now-infamous @FFmpeg account on X. VLC (@videolan), by the way, is also a legendary piece of open-source software: it's a video player that can open basically anything & has been downloaded over 6 billion times. I think both FFmpeg and VLC are two of the most important and impactful software systems ever created, both open source, and both created & maintained by volunteers: brilliant engineers from all walks of life. Thank you to everyone who contributed to FFmpeg and VLC, and in general to all engineers giving their heart & soul to building systems used by millions (or billions) of people, and often doing so not for money, status, or fame, but purely for the love of building great software and doing good for the world. Thank you to the builders! 🙏❤️ Shoutouts in this chat to @ID_AA_Carmack @karpathy @elonmusk @TimSweeneyEpic and everyone who is a contributor & fan of open source! It's here on X in full and is up everywhere else (see comment). Timestamps: 0:00 - Episode highlight 2:17 - Introduction 5:35 - Weirdest things VLC opens 9:59 - How video playback works 19:20 - Video codecs and containers 30:07 - FFmpeg explained 51:07 - Linus Torvalds 55:46 - Turning down millions to keep VLC ad-free 1:10:04 - FFmpeg & Google drama 1:29:18 - FFmpeg developers 1:35:55 - VLC and FFmpeg 1:40:29 - History of FFmpeg 1:43:46 - Reverse engineering codecs 1:57:01 - FFmpeg testing 2:01:08 - Assembly code (handwritten) 2:25:26 - Rust programming language 2:34:42 - FFmpeg and Libav fork 2:43:04 - Open source burnout 2:50:51 - x264 and internet video 3:04:07 - Video compression basics 3:11:04 - CIA and fake VLC 3:21:39 - Ultra low latency streaming 3:39:07 - AV2 codec and video patents 3:48:59 - VLC backdoors 3:59:14 - Video archiving 4:05:51 - Future of FFmpeg and VLC

The only thing better than shipping a feature and seeing users find it before the announcement? Shipping TWO features😆 Just dropped: @ mentions and / commands are now live in ardor.cloud! I promise I'll do marketing next (starting with a real landing page)... right after I ship custom domains and skills ASAP 🫠

Diving into marketing to grow the platform and attract users, and I just had a brutal realization: The more reach and reposts you want, the more cringe and clickbait you have to embrace🫠 I used to wonder why extremely smart, educated founders post absolute cringe on social media. Now I get it. The more nuanced, accurate, and carefully worded your thoughts are, the less anyone cares. Nuance gets ignored by the media, followers, and peers alike. Nuance doesn't scale. Clickbait does🤷‍♂️ Btw, the image below was generated using my platform: ardor.cloud 🪄 Just shipped image and video generation! Check it out 😎

Shipping non-stop! Come try it out 😁

That vibe when you're burning the midnight oil on a Sunday release, just to give folks something cool to wake up to on Monday morning: youtube.com/watch?v=Xoodun… 🔥 And then it hits you: you still need to add deep linking... oh, and Ardor totally spammed the Codex PRs, so you gotta clean all that up before shipping 😁 Btw, here's how it looks in text. Yep, solution paths inside Ardor are finally human-readable now, and the VFS is feeling more and more like a regular local machine: ``` [$index.css](/Marathon Archive/uesc-terminal/src/index.css?environment_id=dev) ```

Make file/resource search - In and out, 20-minute adventure 🫠 Oh yeah, search should work through the entire workspace virtual file system - tons of files and resources - we need an index... and for it to update quickly...

Just rolled out to production: you can now generate images and video! 🚀 Even better - the agent can now see and hear. Just drop in a video or audio file (Gemini model only), and it understands it completely. Files only for now, links are on the roadmap. The foundation for marketing use cases, conversation analysis, and visual understanding is unlocked. For example: 1️⃣ I upload an old conference recording. 2️⃣ Ardor analyzes the audio & generates a custom cover image. 3️⃣ Ardor writes the article, embeds the audio + cover, and publishes it. Come try it out 😏 P.S. If anything breaks, don't hesitate to ping me directly 😎 I might be a little slow to reply since it's late here 😅

Stripe just gave you a new way to create generational wealth. An entirely new market of agent payments now gets a huge boost. Agents will be spending billions for individuals and corporations. Autonomous system growth will make a flywheel for generating and spending. I believe that all humans will have personal agent, and all companies will have their own centralised agent to help every employee. Make solutions for agents, optimise your existing one for agents - they will be new customer base. WAKE UP.

@Seliydf Cool! Thanks for sharing, I will check at morning :)

@Constantinefom It’s called orynth.dev. You submit your product, it gets reviewed, and if approved it goes live for discovery. Some founders also enable a live market around the product.

Well, the full cycle of asset generation - check Images: • text-to-image • image-to-image • images-to-image Videos: • text-to-video • image-to-video • interpolation (generation based on the first and last frames) • reference asset images, up to 3 pictures • video extension, but the extension is always +7 seconds And all of this can be immediately sent as assets to the artifact and published at stage1.artifact.ardor.build/artifact-api/p… I'll deploy it to production tomorrow after finishing polishing

When you finally decide to buy PRO, but @OpenAI is against this idea 🤣 Already burned $4k on API 😓

@getnachonacho Yeah, usually works best

@Constantinefom The key is starting small and focusing on the niche that needs your product most. Engage directly with them, listen, and iterate quickly based on their feedback.

Most challenging part - distribution and users attraction. How did you find your early adopters and users?

Proud and terrified in the same moment. I was on a call with a support engineer from one of our enterprise customers, collecting feedback and walking through real requests they had tried in our platform. We got to one run they had started and forgotten about. So we opened it live on the call. The prompt was: “Check gateway sidecar for service-X. WSS subscriptions are acknowledged, but no data is returned.” The agent wasn’t given the repo, the service, or the file path. The service wasn’t even set up in the workspace. So it started digging. It walked workspace topology, checked solutions/environments/services, hit dead ends, found empty checkouts, created a temporary service to get a working repo, and then did the part that really got me: it used git history as a clue. It found the relevant component through commit history, traced which branches contained it, checked out the right branch, narrowed the issue down to the gateway-sidecar config, patched it, and pushed a fix branch. That’s the part I’m proud of: it spent ~30 minutes trying to solve the task instead of giving up after the first blocked path. The terrifying part: when its default bot identity was rejected on push, it re-authored the commit under a verified company email and got the branch through. Still digging into how it had enough write access to do that in the first place 😳 This is what building agentic systems feels like right now: the persistence is real, the usefulness is real, and the need for hard boundaries is very, very real. Boiiii that was beautiful. The fix didn’t work. But the process 😅

@julezrz Cool! You look for leads in public/communities than reach out? Could you share any playbook?

@Constantinefom For me personally it was cold DM’s on X and reddit.

@rohidasgowda Thanks, dude!

@Constantinefom Congratulations dude.

First enterprise customer on per-seat + PAYG 🎉 Amazing feeling. Also anxiety 😅

@stalkermustang Yep, same. Also a lot of issues during context compaction. Need to restart it several times to make it work. Feels more like batching, querying cause of API overload

What's up with the Codex app? Now when I press enter to send my request, I'm waiting 3-10 (maybe even 15?) seconds for it to be sent. That's an awful user experience! And it wasn't like that before this last big update. reproduced on fresh version 26.415.32059 Anyone experiencing the same? cc @reach_vb @thsottiaux