Intentando hacer ing. aeroespacial... "Es mi trabajo no estar nunca satisfecho", W. Von Braun
Joined November 2018
- Tweets2.5K
- Following216
- Followers15
- Likes349
AJ retweeted
Ven películas de Guillermo del Toro y luego se sorprenden de que no apoye a Bolsonaro. Escuchan la música de Green Day y se indignan de que sean anti Trump. Leen 1984 y creen que es un libro anti LGBT. Y con esa capacidad de comprensión, nos quieren venir a explicar el mundo.
AJ retweeted
Want to fly multiple drones indoors without expensive motion-capture systems?
Low-Cost-Mocap uses external cameras, computer vision, and ESP32-based hardware to track drones in real time and support autonomous indoor flight.
What would you track with a DIY motion-capture system drones, robots or human movement?
by @IlirAliu_
AJ retweeted
We’ve lost an icon. Margaret Hamilton, a computer science pioneer best known for her work at MIT leading development of the flight software for the Apollo missions to the moon, has died at 90. Read about her legacy via MIT News: news.mit.edu/2026/margaret-h…
ALT Margaret Hamilton is reclined inside a NASA spacecraft module, surrounded by control panels and switches.
AJ retweeted
Así se hizo el stop motion de Lizzy —que aparece en Just Dance Decades of Hits y 2026—, realizado por el equipo de inspira_stopmotion 💃🏻🕺🏻🪩✨ #JustDance
wow, tal vez podría ser el inicio de un tratamiento contra el Alzheimer, el Parkinson, y la catatonia abc.es/ciencia/premio-nobel-…
AJ retweeted
The most important briefings of the day.
Karl Deisseroth, professor of bioengineering and of psychiatry and behavioral sciences, shares the news of his Nobel Prize with his kids.
AJ retweeted
Un japonés contó durante un podcast que en el concurso mundial de robótica siempre hay dos países que suelen estar entre los primeros lugares: Japón y... En eso les pregunta a los demás cuál creían que era el otro país.
Y ellos tratando de adivinar:
🤔 “¿Corea? ¿China?”
Pero el japonés les dijo que no, que era México 🇲🇽
Resulta que él conoce a Jorge Orozco, presidente de la zona de Mecatrónica en México y siempre que le preguntaba cómo les había ido, Jorge le salía con lo mismo: que habían quedado en primero, segundo o tercer lugar.
Hasta que un día el japonés ya no aguantó la curiosidad y le preguntó:
🗣️ “¿Por qué siempre México?”
Y Jorge le respondió:
🗣️ “Ay, no seas malinchista. Los mexicanos somos unos fregones. ¿Sabes qué pasa? Que los europeos y asiáticos planean todo a la perfección, pero cuando pasa algo que no tenían planeado durante el combate, se quedan congelados y no saben qué hacer. Nosotros, los mexicanos, como no planeamos, ya sabemos que algo va a salir mal y sabemos cómo resolverlo.”
Y pues simón, sí somos JAJA.
Según el japonés, esa capacidad de improvisar también se ve en cosas tan simples como cuando le haces una broma a un mesero o a un conductor de Uber y te terminan respondiendo con otra.
Para él, esa creatividad es justamente una de las cosas que nos hace destacar a los mexicanos.
Y cerró diciendo:
🗣️ “Lo único que le hace falta a México es la disciplina. Si juntas eso con la mejora continua, ya la hicieron."
cc @Artemio
AJ retweeted
Vengo a que perdáis un poco vuestro tiempo.
¿Has crecido bajo MS-DOS?
Pues aquí tienes casi 6.000 juegos de los 80/90. Los he traído desde Archive y diseñado como si fueran los 80. Elige tu "monitor", busca el juego e inicia con .EXE o descubre más cosas en "Others".
PLAY-DOS
AJ retweeted
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.
AJ retweeted
🔴🔴Una vergüenza, es una completa vergüenza que Morena haya elegido a la impresentable Clara Luz (perdedora el 2021 contra Samuel García) para que otra vez, sí OTRA VEZ sea su coordinadora para el 2027 en Nuevo León.
No me olvido sus nexos con la secta NXIVM del preso por tráfico sexual Keith Raniere. A quien incluso entrevistaba con alegría.
¿A qué le juega Morena? ¿A regalarle Nuevo León al MC a cambio de su apoyo en el Congreso de la Unión? ¿Y por eso mandaron a una candidata indefendible para que pierda?
Pudiendo competir en Nuevo León y ganar por primera vez, le regalan otra vez el Estado al MC para que Mariana Rodríguez sea gobernadora después de Samuel.
Una vergüenza, Morena. Una completa vergüenza. Que los juzgue la historia y nos los achichincles que no ven más allá de su cheque mensual.
AJ retweeted
#PCエンジン #pcengine #ai #ai_slop #slop_manager
I think the debate about AI in related to programming/development is pretty interesting. But the one that sticks out to me the most.. is coding.
TLDR; "vibe coding" isn't just development without hands-on coding. People are using "vibe coding" to do a lot more than just "coding". Skip to the end for the point of this post (tldr).
As in, defenders of vibe-coding seem to have the perspective that coding is a "means to an end". I think for most modern software.. they're not entirely wrong. Even before AI, there has been this push to do as little coding as possible.. and just "configure" what you need. I mean, Salesforce built an entire billion-dollar platform and motto on it: clicks over code. Tho the irony there, is more complex projects still need coders haha (at the time). But even then, when I saw "frameworks" for front-end stuff.. the pattern I saw was "write as little code as humanly possible". I'm not a full-stack or web-dev "developer". But I did spend more than a year doing full-stack full-time about 7 years ago (Vue was one of the frameworks). Anyway, the trend and pattern was already there. And from a business model it makes sense - why should I hire and pay more for someone with a CS degree.. when I can pay less for a scrapy boot-camp dev - that does "code lite". That is not me saying that.. those were the words of the director on a project I worked on. Her view was CS degree devs always asking for more money, more likely to job hop for better opportunities, etc. Always causing a risk to the projects that ran with a lean crew. And she wasn't wrong. I did exactly that haha - complained and got a big raise.. and then left for better opportunities no long after.
But when it comes to retro systems, what does it mean to take code out of the loop (AI does your coding)? Because outside of counting cycles, there are a lot of "high" level optimizations you still need to apply. There are constraints of the system.. as a system design challenge. Nothing is free. You can have the best/fastest code.. but if you high level design is hot garbage.. code won't fix it (performance, capability, whatever). But a high level optimization can overcome garbage (but functional) code.
So if you take code out of the equation, you're still left with all sorts of things you still need to understand and optimize on a high level - all around the hardware itself. For example, there's only so much bandwidth for updating VRAM on these system (unless you're the NES haha), right? SMS, PCE, MD, SNES - there is only so much you can upload in a frame. So what happens when you have use case scenarios that blow your bandwidth budget? Do you just take the hit and slowdown the game? Well, if your code is already optimal.. then that level of work is on the system design level. Start with converting the dumb schedular for uploads.. to a smart scheduler. Or start imposing new rules for upload slots. Or Partial streaming and lowest priority drops. Etc. You have all these things that need to update.. what can be delayed that won't have a visual or control-feedback impact? If you have a priority system, how do you avoid starvation of lower priority uploads in the queue? Do you implement a gradual priority increment over time for stuff waiting in the queue? This is what high level optimization looks like.
That kind thing, where even if the code was average - your performance optimizations come from a higher level domain - system level design. And I don't mean "system" as physical game system.. I mean system in engineering terms. You could apply this to all sorts of aspects of a game.
My point being, you could be a systems architect - principle or senior principle, and not write any code.. but have the biggest impact on performance and execution. In that respect, code IS just a means to an end. As in, you have that deep understanding, knowledge, and skill.. you're just not writing any code. And.. I have found a few people like this - that have NEVER written any code in their life, but just intuitively understand systems at this level.. and can come up with clever solutions that even experienced codes didn't think of (I'm talking a retro game development here). Not everything is code.. and I say that as someone who has cut their teeth in writing crazy "clever" assembly code that takes into account every cycle.
But vibe-coding without this system level knowledge/experience/understand.. means you're also vibe-"system engineering" and vibe-"system architecting".. not just vibe "coding". I think the general public just sees or thinks only about the coding part of it.. but a lot of it is vibe-"do everything for me".. and not just code. There's a difference between a "tool" and a "genie". Turning on god-mode in a game doesn't make you better at the game.. and skill atrophy sets in, it makes you worse. Or you never had the skill/experience to begin with.
There is a difference between someone who knows what they are doing, has that deep understanding, experience, knowledge and real capability.. and using AI as tool to develop the code. Vs someone that "vibe-coded" a project into existence - with minimal or no experience. Think of it like a veteran Lead that's been given a team of AI workers to perform the tasks, VS the junior engineer that got asked to lead a project with the same AI team. The real Lead is going to be in control of the system design/arch - making corrections, giving thorough directions. The junior SWE is going to be dependent on AI making most (if not all) of those decisions.
(tdlr)
I have been running some experiments on this over the past couple of months. From the systems-engineering and systems architect perspective.. I'll post some interesting findings for what hands-off "vibe-everything" has produced (from the top models on high effort). To be fair, I used the PC-Engine as the platform of choice... since I know it inside and out.. like the back of my hand. I needed to be able to fully 100% properly assess things/results. I think the results are pretty interesting (and not all of it good way, for sure).
I'll post the results in an update. FYI - this isn't "AI sucks", or "AI is great".. just an academic-ish perspective of AI performance with experienced guidance vs hands-off "do it all". If you're a SWE right now in the field, you probably already have some experience with this anyway. But this still might be interesting. At least, I found the results interesting.
Note: Those YT videos where they give it a long prompt to make a game - is a lousy way to evaluate any model hahah. This also doesn't deal with art and music - that's a whole 'nother topic.
