gm rise and shine

Cabeza de ratón o cola de león?

College group projects are the easiest proof that communism won't work. You are in more or less the same condition, equal class, equal power, same goal, same constraints. Still you can divide the group in 3: Parasites, mediocre peasants, and leaders.

que droga dura la raspberry pi

most important lesson from years of distributed systems: keep everything on a single machine for as long as humanly possible

if steam launchs and runs mewgenics on my linux machine, bye bye microslop

Telling people vibe coding VMs is a skill issue when you haven't shipped code to production with real money and security in years is peak Dunning Kruger.

Mr President if you can hear me, AFIP is days away from getting a nuclear weapon

running doom on a brain > AGI

Gamedev is such a beautiful and difficult craft. It is such a shame that it is based on crunch and low salaries.

Turns out you can run a LLM with great performance on a box of matches.

At the end of the day "vibe code" is just a fancier, more expensive and convoluted fork.

incredible that we built all this RAG and vector database stuff and it turns out that grep from 1973 works better than all that

que chota le hicieron a claude? Parece un empleado publico: dice cualquier pelotudez con tal de no laburar

All this fuss about "agents", "sub agents", "skills", wathever. And at the end it is just an initial prompt with extra steps.

“GPT2 es demasiado peligroso para dárselo a la gente” - sam altman, 2019 “Mythos es demasiado peligroso para dárselo a la gente” - dario amodei, 2026 os pido que no caigáis en el hype, solo eso

Ahora que Claudio me hace garpar aparte el plan PRO para usar la CLI (unica razon por la que pague el pro originalmente) y ya no me subsidia los tokens. Estoy pensando en gatillar una Mac para correr los modelos en casa. Tendria que hacer costos

Algo que aprendi en 2025 es q hay que automatizar los deploys lo mas posible cuando uno empieza a laburar en serio. Es abismal el tiempo que te ahorras.