Acabo de lanzar esta cuenta para enfocarme 100% en lo que realmente me apasiona:
Automatizar negocios con Inteligencia Artificial.
Me llamo Edwin y ayudo a empresas hispanas a implementar sistemas de IA y automatizaciones que realmente funcionan.
@wilkinmeta Yo no entiendo la gente, yo todo el tiempo he jugado con personajes femeninos en todo los juegos, prefiero ver una mujer mientra juego que un hombre 😂
Claude Security scans now run on Claude Mythos 5, available today in public beta for all Claude Enterprise customers.
Put our most capable security model to work on your codebase, no separate model access needed.
Guarden este tweet.
En 6 meses la infraestructura de SpaceX va a ser el estándar de la IA.
- Grok
- Grok Bot
- Grok Build
- Grok Imagine
- Build Mode…
El que no lo vea ahora, lo va a entender después.
Tengo como 30 PDFs técnicos guardados que nunca abrí y ahora voy a sumar toda esta colección a la pila. Alguien debería escribir un libro sobre por qué acumulamos libros que no leemos. ¿Cuántos tienes tú sin abrir ahora mismo?
Muse Spark 1.2 Contributor is now available on OpenCode Go
this model will train on your data so it requires explicit opt-in and is restricted in some regions
in exchange you get incredible usage limits
This is the AI progress metric that actually matters: not benchmarks, real wet-lab success rates. 2-3x the field baseline on a task with zero room to bluff. The binder either binds or it doesn't.
Designing a binder is an easier process than designing a drug, but it’s a useful proxy. The typical success rate in the field today is between 10% and 15%.
Between 22% and 35% of Claude's designs bound successfully, depending on the setup. Some of its strongest designs bound several times more tightly than the best published de novo binder.
the bar for "ai product news" in 2026 is literally "we stopped throttling our own timers" and it's still more exciting than 90% of model launches this year
Claude Desktop now starts ~2x faster than it did a month ago.
When the app started in the background, its timers got throttled and the JS engine dropped into power-saving mode. We now boot at full speed even while the window's still hidden, along with some smaller perf fixes!
Good. I run small agents, not frontier models, and I already had a scare: OAuth dropped, the fallback model answered a lead with raw internal classifier text. Every critical gate now has a pinned model and monitoring. That discipline isn't optional.
We have paused some frontier RL training to ensure that we can meet the appropriate alignment, security and monitoring standards for the new level of capabilities in front of us. Model progress is now extremely rapid, and we always said we would take action if we felt that model capabilities were outstripping the pace of safety and alignment.
We care very deeply about AI safety. We believe the entire field will have to coordinate on shared safety standards, but will act unilaterally in the meantime.
We expect confidence in safety to increasingly set the pace of AI progress. We are optimistic about the alignment work we are doing, and we remain committed to making frontier capabilities widely available.
https://t.co/51kvKfbfrO
946 personas guardaron este tweet y solo 4 comentaron. Eso significa una cosa: todos van a usar Mono Charts y nadie lo va a admitir en voz alta.
https://t.co/bXfP4EYuWC
¿Tú lo guardaste también o ya lo instalaste?
Charts minimalistas y animados para tu web.
Se llaman Mono Charts:
✓ Los añades con un comando
✓ Hecho con React y Motion
✓ De código abierto
→ https://t.co/HyyyRCVtYS
@frizzante80 Ese es el uso real que no se vende: sacar del cajón manuales que ya nadie lee. ¿Te dio el paso genérico o le diste marca y modelo del coche para el pirindolo exacto?
Le pones nombre a tus subagentes en Claude Code: Architect, Reviewer, Tester y se coordinan solos. Eso hace innecesario montar CrewAI o AutoGen para la mayoría de flujos: ya viene nativo, sin librería extra.
¿Qué tres roles armarías tú primero?
@ClaudeDevs Even Bun's garbage collector learned to wait until you're done talking before it starts cleaning. Next release: it stops interrupting to tell you it disagrees with your prompt.
Six months from now every business profile pic on this app is a Grok render and nobody admits it. The tell won't be the hands anymore, it'll be everyone suddenly having a perfect headshot on the same week.