Acerca de
Lo que hago
• PHP Developer, experto en UI y arquitectura de la información DB…
• PHP Developer, experto en UI y arquitectura de la información DB…
Servicios
Actividad
480 seguidores
-
Martin Fasani ha compartido estoThis video is for those who "feel to old" to make a complete change of career. This brave woman TuneGirl creates fully analog live performances through Eurorack modular systems. At 63 years old she still has the energy for doing this techno-live performances and is a rising star in the scene, being requested in many venues
-
Martin Fasani ha compartido estoSign-up for the product launch https://lnkd.in/eQ_4q5MY
-
Martin Fasani ha compartido estoMartin Fasani ha compartido estoI vibe-coded a SaaS app in 2 hours, and it hit $10K MRR in a week. There's just one problem... I made this whole thing up. If you’ve been on LinkedIn lately, every day there’s a new post: • "I built my MVP over a weekend, $5K MRR in 3 days." • "I run a $10M ARR SaaS with just a few AI agents." • "I replaced my $500k/yr marketing team with AI" • "I built a $50m ARR SaaS while napping" These posts get 1000s of likes because they sell one very seductive idea: Building a real business is now easy. You can skip the hard parts with AI. That's a lie. Realistically, AI gets you through maybe 20% of your product. The other 80% is still tedious, hands-on, unglamorous work. That’s where the “vibe coding” fantasy breaks. To be fair, AI is genuinely useful for exploration, prototypes, and small utilities. But for complex, long-term systems? Someone still has to guide the architecture, test the output, and fix what breaks. There is no shortcut for that. The reason you never see that version of the story is simple: it doesn't go viral. So here's the honest post nobody writes: "Spent 8 months building my SaaS. Got 3 signups. Pivoted 3 times. Building again. Still learning hard lessons." Not as exciting. But that's kind of the point. What's the unglamorous part of your build that never makes it into a LinkedIn post? ...And if you've actually shipped something nd acquired customers fast with AI, I want to hear it. Convince me I'm wrong.
-
Martin Fasani ha compartido estoMartin Fasani ha compartido estoEl 7 de julio de 2017, Diario EL PAÍS publicaba que Filmin, al cumplir 10 años, salía de los números rojos por primera vez en su historia. Hemos rescatado el recorte de la hemeroteca para recordar que, desde entonces, no hemos dejado de ser rentables La revista Economía 3 (una publicación especializada en análisis financiero y empresarial, con más de 30 años de trayectoria) ha analizado las cuentas en España de cuatro de las principales plataformas de streaming a partir de sus últimas cuentas anuales registradas. La conclusión es que Filmin, la única española del grupo y la única que no pertenece a una gran multinacional, presenta el mayor margen neto. Es decir, es la que saca más rentabilidad a cada euro que factura. Lo importante es lo que hay detrás del dato. Porque la rentabilidad es lo que nos permite seguir siendo quienes somos. La sostenibilidad económica es la condición que hace posible mantener intacto nuestro modelo, basado en un catálogo construido con criterio editorial, una estructura pensada para durar y un compromiso con la comunidad cinéfila que no depende de ciclos de inversión ni de cambios de estrategia. Ser rentables es nuestra manera de garantizar que Filmin no necesite convertirse en otra cosa para seguir adelante. Si creemos que un modelo cultural independiente merece existir, la mejor forma de protegerlo es hacerlo viable. Enlace al artículo en el primer comentario. #streaming #cineindependiente #industriaaudiovisual #sostenibilidad
-
Martin Fasani ha compartido estoNew library comes from a dream idea in order to draw shapes, font and icons using a light JSON domain language #C++ https://lnkd.in/ef-zKQDiGitHub - martinberlin/FastJsonDL: a JSON Domain Language to draw in your Eink receiving the payload from any endpointGitHub - martinberlin/FastJsonDL: a JSON Domain Language to draw in your Eink receiving the payload from any endpoint
-
Martin Fasani ha compartido estoOn the last days I spent 3 hours with Claude and Copilot testing if I can install a symfony based shop as a demo to sell my PCBs. And it worked out: https://shop.fasani.de only a prototype for now, but is fully working, with PayPal and Stripe payments tested. If someone is interested to see my implementation is fully open in Github. Both the server settings and customisations were 95% done with AI. Of course is my take to select images and other stuff that's why still looks ugly. But is a proof that this days you can install anything super fast with the help of artificial intelligence.
-
Martin Fasani ha compartido estoThis actually looks very interesting Espressif Systems https://lnkd.in/eXJzp6sGMartin Fasani ha compartido esto🦎 Inside the New ESP-Brookesia v0.7 Architecture We have just published a new article on the Espressif Systems Developer Portal introducing the new ESP-Brookesia v0.7 architecture. The article walks through the framework’s key modules, including Utils, HAL, General Service, AI Agent, and AI Expression, and explains how they fit together to support the development of next-generation AIoT interaction products. 🔧 It also includes two practical examples to help developers quickly understand how to build on top of ESP-Brookesia and accelerate product development. Read the full article on the Espressif Developer Portal. 👀 https://lnkd.in/ePRcWfV2 #Espressif #ESPBrookesia #AIoT #IoT #AI #ESP32 #ESPIDF
-
Martin Fasani ha compartido estoMartin Fasani ha compartido esto2 years ago, I posted "12 Must-Know GenAI Terms." It went viral. Millions of impressions. Thousands of saves. But here's the truth — that vocabulary is already outdated. The industry didn't just evolve. It completely rewired itself. We moved from AI that answers questions → to AI that takes actions. From models you prompt → to agents that plan, decide, and execute. That shift created an entirely new vocabulary. One that most people are still piecing together from scattered threads and docs. So I built the 2026 edition: 𝟏𝟐 𝐌𝐮𝐬𝐭-𝐊𝐧𝐨𝐰 𝐀𝐠𝐞𝐧𝐭𝐢𝐜 𝐀𝐈 𝐓𝐞𝐫𝐦𝐬 Here's what's inside and why each one matters: 𝟎𝟏. MCP (Model Context Protocol) The USB-C of AI. One standard to connect agents to any tool, API, or data source. 𝟎𝟐. Agent Loop Perceive → Plan → Act → Observe. The heartbeat of every autonomous agent. 𝟎𝟑. Tool Use Agents with hands. The ability to call real-world functions — not just generate text. 𝟎𝟒. Orchestrator The manager. It breaks down complex goals and delegates to specialized subagents. 𝟎𝟓. Subagent The specialist. One job, done well, inside a larger pipeline. 𝟎𝟔. Memory Short-term lives in the context window. Long-term lives in a vector store. Both matter. 𝟎𝟕. Grounding Tethering agent outputs to verified, real-world data — the antidote to hallucination. 𝟎𝟖. Guardrails The safety net. Rules that stop agents from doing things you never authorized. 𝟎𝟗. Sandboxing Isolated execution. Agents run code here before touching anything real. 𝟏𝟎. Human-in-the-Loop The pause button. High-stakes actions wait for your approval. 𝟏𝟏. Context Window Your agent's attention span. The bigger it is, the more it can reason over at once. 𝟏𝟐. Multi-Agent When agents work as a team — each specialized, all coordinated. Save this. Share it with your team. The engineers, architects, and PMs building with AI need this vocabulary now. ────────────────────── Which of these 12 terms do you think is the most misunderstood in real enterprise AI projects right now?
-
Martin Fasani ha compartido estoTesting Doom in Raspberry pi monitor. Obviously not playable, but it was a fun demo. This is actually only to use VIM or a text editor. Project to be available soon here https://lnkd.in/edmbd9Vf
Martin Fasani ha respondido a un comentario
3 semanas
Because v7 is custom and has another connectors, I listed them one by one in Tindie but seems no one reads the specs before buying something
Martin Fasani ha respondido a un comentario
1 mes
Jose Ignacio Marti Gil para el consumo hay actualmente chips muy baratos, que dan el resultado tanto de voltage como % de carga con bastante presición. La idea de hacer procesos AI dentro del dispositivo me parece muy interesante el tema es que habría que usar un ordenador bastante mas potente que el Microcontrolador Espressif que usamos. Adjunto un análisis de lo que consume el producto actualmente, echo con un analizador de Nordic que se llama Power Profiler Kit.
Experiencia y educación
-
Sensoria Lab
Mira la experiencia completa de Martin
Al hacer clic en «Continuar» para unirte o iniciar sesión, aceptas las Condiciones de uso, la Política de privacidad y la Política de cookies de LinkedIn.
Licencias y certificaciones
Proyectos
-
Sensoria firmware
Powered by Espressif-IDF S3 and built in around FastEPD Eink controller, sensoria has all it needs to measure and show an Eink information dashboard
https://github.com/sensoria-iot/sensor-eink -
Epaper weather station
-
Ver proyectoEpaper weather station aims to be an easy starting point for someone wanting to build a panel for a shop, office or home showing real-time-clock (aware of time, day, month, etc) and also added sensors.
This has support with my library CalEPD for about 30 models of epaper (GoodDisplay and Waveshare mostly)
For other more industrial, faster and bigger epapers like the parallel ones using EPDiy board or IT8951 controllers.
Idiomas
-
German
Competencia básica profesional
-
Spanish
Competencia bilingüe o nativa
-
English
Competencia bilingüe o nativa
-
French
Competencia básica
-
German
Competencia básica profesional
Recomendaciones recibidas
3 personas han recomendado a Martin
Unirse para verlo