dfbrule (@dfbrule) profile picture

Repeat founder with a passion for computer vision. Currently building @finegrain_ai. Previously: built @moodstocks (acquired by Google).

Paris
Joined November 2007
Photo by dfbrule (@dfbrule)Photo by dfbrule (@dfbrule)Photo by dfbrule (@dfbrule)Photo by dfbrule (@dfbrule)Photo by dfbrule (@dfbrule)Photo by dfbrule (@dfbrule)Photo by dfbrule (@dfbrule)Photo by dfbrule (@dfbrule)Photo by dfbrule (@dfbrule)
A French engineer who lives quietly in Paris has spent 30 years writing software that the entire internet now runs on without knowing his name. He wrote the code that streams every YouTube video, every Netflix show, every TikTok clip. He wrote the code that runs the virtual servers underneath AWS, Google Cloud, and Microsoft Azure. He calculated more digits of pi than anyone in history. He has no Twitter. He has no marketing. He just keeps shipping. His name is Fabrice Bellard. Here is the story, because almost nobody outside the systems programming world knows what one man has built. Fabrice was born in 1972 in Grenoble, France. He studied at École Polytechnique, the top French engineering school. He never went to Silicon Valley. He never built a startup empire. He just wrote code. In 2000 he started a project called FFmpeg, an open-source multimedia framework for encoding, decoding, and streaming video. He was 28. The project did one thing nobody else had done well. It handled every video and audio format that existed, in one library, on every operating system. He led it himself for years. Today FFmpeg is the invisible engine of the internet. YouTube uses it. Netflix uses it. VLC uses it. Chrome and Firefox use parts of it. Every Android phone, every iPhone, every smart TV, every video editing tool you have ever touched runs FFmpeg somewhere underneath. If you have watched a video on a screen in the last 20 years, Fabrice's code processed it. He was not done. In 2003 he started QEMU, a machine emulator and virtualizer. He wrote it solo until version 0.7.1 in 2005. QEMU lets you run any operating system on any other operating system. It became the foundation of modern virtualization. KVM, the Linux kernel hypervisor, runs on top of QEMU. Every major cloud provider, AWS, Google Cloud, Microsoft Azure, IBM Cloud, runs virtual machines on infrastructure built around it. The Quick Emulator is the most cited piece of cloud infrastructure code on Earth. He kept going. In 2001 he won the International Obfuscated C Code Contest with a small C compiler that grew into TCC, the Tiny C Compiler. TCC can compile and boot a Linux kernel from source in under 15 seconds. In 2004 he calculated the most digits of pi ever computed at the time, using a personal desktop computer and an algorithm he derived himself called Bellard's formula. In 2011 he wrote a complete PC emulator in pure JavaScript that runs Linux in your browser, a project called JSLinux that engineers still cannot believe is real. In 2019 he released QuickJS, a small but complete JavaScript engine that fits where V8 cannot. In 2021 he released NNCP, a neural network based lossless data compressor that immediately took the lead on the Large Text Compression Benchmark. Then he turned his attention to large language models. He built TextSynth Server, a web server with a REST API for running LLMs locally. He released ts_zip and ts_sms, compression utilities that use language models to compress text and short messages at ratios traditional algorithms cannot reach. He released TSAC, a very low bitrate audio compression system. In December 2025 he released Micro QuickJS, a new JavaScript engine for microcontrollers, separate from QuickJS, designed for environments with almost no memory. Fabrice co-founded a telecom company called Amarisoft in 2012, where he serves as CTO. Amarisoft builds 4G and 5G base station software used by carriers and labs around the world. He has been running it for over a decade while continuing to ship personal projects from his own home page at bellard dot org He has no Twitter. He has no Instagram. He gives almost no interviews. His personal website is a flat list of projects with no styling, no fonts, no marketing copy. Just titles and links. A quiet French engineer who never moved to Silicon Valley wrote the code that quietly runs the internet. He is still shipping.
113
1,368
160
7,243
1,277,129
Finegrain Agentic Camera vs. ChatGPT. I could literally resell my entire wardrobe while ChatGPT is still processing the first shot. ⚡️
1
12
1,024
I'm often asked why I decided to start another company after selling my first one to Google. The answer is simple: I wanted to take great photos of plush toys 🧸 See for yourself 👇
2
12
1,560
« La peur est mauvaise conseillère. » Je venais d’embaucher un collaborateur. Tout au fond, une petite voix me disait que c’était une erreur. Les signaux faibles s’amoncelaient : doute sur les compétences, doute sur l’alignement culturel. Mais ma raison les refoulait : trop de temps, trop d’énergie avaient été consacrés à cette recherche. Trop de promesses faites à l’équipe et aux actionnaires. Annuler aurait provoqué un sentiment de régression presque insupportable. Je décidais donc d’ignorer la réalité. Pour me réveiller quelque temps plus tard avec la gueule de bois, à brûler de l’énergie pour éteindre les multiples incendies de ce mauvais choix désormais avéré. Ma peur l’avait emporté. #IlsSavaientTout
2
118
« Mieux vaut être seul que mal accompagné. » Nous devions être 4 quand nous avons lancé Moodstocks. Mais le jour de démissionner, seuls deux d’entre nous avaient franchi le Rubicon. Sur le coup, on le vit comme une catastrophe. Mais sur le tas, on réalise que c’est une sélection naturelle salvatrice. Le mal-être aurait été profond pour ces personnes visiblement pas faites du bois entrepreneurial, et bientôt, face à la muraille des difficultés, il aurait sûrement créé un grand ressentiment : celui d’avoir gâché une carrière toute tracée. Ce qui compte, c’est la qualité et pas le nombre. #IlsSavaientTout
7
269
« Un tiens vaut mieux que deux tu l’auras. » Cette journée avait pourtant bien commencé. Après des semaines de négociation, j’avais les signatures de tous les participants au tour de table. Il ne m’en restait qu’une. Une formalité : il m’avait donné son accord oral au tout début du processus de levée de fonds. Montant, valorisation, pacte : on était bon. Je l’appelle une première fois. Pas de réponse. Tout le monde a le droit d’être occupé. Je retente ma chance un peu plus tard. Toujours rien. J’envoie un SMS. Silence. Je tente le message vocal. Sans succès. Après plusieurs heures de tentatives infructueuses, le verdict tombe : « Ok pour venir, mais il faut changer les conditions. Sans moi, tu ne peux pas boucler. » Je tombe des nues. #IlsSavaientTout
4
7
1,118
« Qui veut voyager loin ménage sa monture. » Lancer son entreprise en parallèle d’une activité salariée prenante, quelle idée lumineuse ! De jour, je subviens à mes besoins. De nuit, je prépare l’avenir. Sauf qu’après six mois à dormir quatre heures par nuit, on est bon pour la casse. C’est un exemple extrême d’un mythe qui a la vie dure : un bon entrepreneur devrait se tuer à la tâche. Mais on ne peut pas sprinter tout au long d’un marathon. Et le marathon entrepreneurial, lui, peut durer dix ans. #IlsSavaientTout
1
2
150
« Petit à petit, l’oiseau fait son nid. » Quand nous avons imaginé, dès 2008, que l’appareil photo de l’iPhone pourrait devenir une interface visuelle pour effectuer des recherches en ligne, bien peu ont su se représenter l’impact potentiel. Nous voulions créer un nouvel usage. Pourtant, les fonds d’investissement de l’époque nous posaient les mêmes questions qu’à un boulanger cherchant à ouvrir sa boutique : combien de clients ? Quel chiffre d’affaires ? Quelle structure de coûts ? En 2014, nous avions convaincu quelques clients téméraires, mais l’adoption restait symbolique. Aujourd’hui, via Google Lens, ce sont des milliards de requêtes chaque mois. #IlsSavaientTout
3
127
🤯 Wondering what your marketplace would look like if every photo had been snapped with the @finegrain_ai Agentic Camera? Well, we have a Magic Wand for that! 🪄 Finegrain automatically decides which images need to be improved, and how. Zero hassle. Powered by our on-device stack, so transformations happen in near real time on commodity GPUs. No need for the latest NVIDIA beast!
2
121
« C’est en forgeant qu’on devient forgeron. » Ce qui permet de déplacer des montagnes, c’est la passion pour un sujet, pas l’expertise qu’on en a au départ. Nous ne savions même pas ce qu’était la computer vision en lançant Moodstocks. Quelques années plus tard, Google nous rachetait. #IlsSavaientTout
5
143
When did capital letters at the beginning of sentences become uncool?
1
105
“Qui trop embrasse, mal étreint.” Aujourd'hui, on appelle ça le focus. On écrit des livres entiers dessus, on organise des offsites stratégiques et on paie des consultants très cher pour nous l'expliquer. Nos grands-mères avaient réglé le sujet en cinq mots. #IlsSavaientTout
114
Hey @bfl_ai @robrombach @jpoersc! Just stumbled upon the FLUX.1 Eraser announcement. Thanks for mentioning @finegrain_ai - we're honored to be included. One question about the comparison though: are you benchmarking against our prompt mode, or masked editing? The vast majority of our users on @fal use Finegrain with masks, which makes inference significantly faster and cheaper. And for teams working directly with us, it's an even bigger difference: on-device editing is sub-second, and we offer flat licensing with unlimited edits (see video). Just thought it was worth pointing out for anyone comparing the different approaches. Congrats on the launch!
186
We’ve spent years at @finegrain_ai building specialized AI models for image editing, eventually making them efficient enough to run entirely on smartphones - without compromising quality. Along the way, we realized something: image editing still takes effort. What if the camera itself produced the photo you actually wanted? Press the shutter. Done. We call it the Agentic Camera. Here it is in action for transactional photography 👇
1
4
3
360
We're shipping bulk processing in Finegrain iOS. I love it - a time saver! Zero pixels sent to the cloud, everything happens on-device.
4
122
dfbrule retweeted
Had a company meeting this morning. Glad to see everyone showed up.
173
242
41
5,756
207,474
Coming soon to Finegrain iOS: Ghost Mannequin tool 👻. The attached video is a screencast of our internal Factory 🏭 in action, but we'll add the new tool to the Fashion Resell studio - 100% on-device as usual!
3
6
702
AI solves laziness, part n. New landing page in hours, better than what I imagined. Thank you, Claude Code and Gemini. Look at this! 👇
1
3
327