Fernando Lazzarin
60 posts

Fernando Lazzarin
@flazzar1n
AI Workflow Engineer & Harness Builder | coding agents, evals, local mode | Keep on shipping.


this screenshot is one pane out of 11 running on my machine right now, and it is the closest thing to an org chart my company will ever have. cursor gave me a $10k ultra trial and i turned it into an engineering department. fable 5 with 1M context reviewing fable 5 with 1M context, on a laptop a stranger on x gifted me, at 11pm, while i write this post. my name is sudo, i am 26, i have zero employees, and i am building a one person, billion dollar software and infra company with a fleet of agents. what you are looking at: my reviewer agent mid audit, grepping cuda compute caps across the codebase before it lets a single line ship. a builder agent is in another pane writing the actual code. the reviewer wrote the task brief, delegated it, and is now waiting to tear the PR apart. somewhere in there it wedged a command, killed it, and rewrote the search on its own, nobody told it to. i did not open a single file. the fleet tonight: 1. one agent runs planning, financials and content, it rebuilt my entire revenue ledger today 2. one reviewer agent orchestrates the backend, writes task briefs, audits code, reviews and merges PRs 3. builder agents take whole tasks and return one clean PR each 4. research agents run benchmarks and dig the web 5. i make decisions, everything else is delegated the rules that make this actually work, learned the hard way over months: every agent gets a name, a lane, and a folder of briefs a builder never touches code before the reviewer pre-audits the task one task, one PR, agents ship whole features not fragments everything that matters lives in files, so any agent can die and its replacement starts warm reviews are adversarial on purpose, agents catch each other's bugs before i ever see them and this pane is just the engineering wing. you have been watching the rest get built for months without knowing it. more receipts coming.



En 1 semana tenemos un nuevo Cursor Meetup Mendoza 🖤 El miércoles 8 de julio nos juntamos para hablar sobre Inteligencia Artificial, compartir experiencias, demos y aprendizajes sobre cómo estamos usando @cursor_ai para programar, automatizar y crear productos más rápido. Ya tenemos la agenda cerrada para veas lo que se viene: ▸ Matías Boldrini(@pescalate_) Apertura y bienvenida. ▸ Nicolas Bustelo(@nicobustelo__) De cero a tu primer venta en 3hs: cómo pasar de una idea a una app funcionando en un solo prompt. ▸ Lucas Trubiano(@lucastrubiano) Docu Viva: cómo automatizamos toda nuestra documentación usando Cursor Automations. ▸ Fernando Lazzarin(@flazzar1n) Cómo Cursor me introdujo a la programación con IA cuando no era técnico. ▸ Comida + networking Queremos que también sea un punto de encuentro para la comunidad tech de Mendoza: un espacio para conocernos, compartir lo que estamos haciendo y generar nuevas conexiones. Si sos developer, founder, estudiante, diseñador, maker o simplemente tenés curiosidad por Cursor y la programación con IA, este evento es para vos. No importa si ya usás Cursor todos los días o si recién querés entender de qué se trata. Va a ser una linda oportunidad para aprender, compartir y conectar con personas que están mirando hacia adelante. Va a haber demos, comida, networking y mucha gente con ganas de construir. Te esperamos! 📍 Campus Olegario, Mendoza 📅 Miércoles 8 de julio 🎟️ Registro gratis(cupos limitados): luma.com/bwst0qek Para más información sobre eventos de Cursor Community: cursor.com/community #IA #Mendoza #InteligenciaArtificial #Cursor #CursorCommunity #Cuyo








Introducing Claude Science, a new app designed with every stage of research in mind. Artifacts traced to their code, environments managed on demand, and 60+ optional scientific databases that you can connect. Available now in beta.





Llevo una semana haciendo stress-testing de Fable 5, corriendo zorbic loops completos en mi agent stack y sincronizando todos los florps nativamente con Codex a través de la nueva capa de glibbificación. Hot take: si no estás traceando tus glibbys a través de troopers con vrentilación asíncrona, tus pipelines son básicamente pre-agénticos. You're NGMI. El overhead cognitivo de trackear glibbys manualmente es una locura cuando escalás. Dejá que el trooper mesh lo maneje.


Introducing Claude Opus 4.8: it builds on Opus 4.7 with sharper judgment, more honesty about its own progress, and the ability to work independently for longer than its predecessors. Available today at the same price.



New in Claude Code (research preview): dynamic workflows. Claude writes an orchestration script on the fly, then spins up a large fleet of coordinated subagents in parallel to take on your most complex tasks. Use the word "workflow" in a prompt to get started.




