Silvio Balduzzi retweeted
What's new in AOS 7.6? 👀
⚡ 307 TB all-flash node support
🔒 Stronger security
🛡️ Better cluster resiliency
💾 External storage enhancements
🔧 Hundreds of stability improvements
See everything included in the latest release: next.nutanix.com/product-upd…
#EnterpriseIT #Nutanix
HAN REPLICADO EL MÉTODO DE RESOLUCIÓN DE PROBLEMAS DE FABLE 5 EN UN PLUGIN QUE CUALQUIER MODELO PUEDE EJECUTAR
159 ejecuciones de agentes después, Sonnet + el plugin igualó a Fable 10/10 en una tarea de research.
El plugin tiene 3 partes:
→ fable-method: da a cualquier modelo un bucle estructurado de resolución de problemas con umbrales de fallo estrictos
→ fable-loop: ejecuta tareas completas con agentes adversariales verificando el trabajo
→ fable-judge: trata cada “listo, todos los tests pasan” como no verificado y lo vuelve a ejecutar todo de forma independiente
Los resultados son concretos:
→ Haiku pasó de 0/4 a 4/4 detectando un test incorrecto antes de “arreglar” código que estaba bien
→ Sonnet + el método igualó al propio Fable en una tarea de research de 5 partes: 10/10
→ El judge llevó a Haiku de 3/5 a 5/5 detectando fraudes plantados en un informe de finalización falso
Lo que más confianza me da es lo que NO hace.
En tareas normales con modelos capaces, no aporta nada.
Lo midieron, lo documentaron y el README lo dice.
El valor se concentra en las trampas: tests incorrectos, falsas afirmaciones de tarea completada y modelos débiles trabajando sin supervisión.
Aporta disciplina, no conocimiento.
La iteración dice mucho de la calidad:
→ La v1 falló su propio test: 0/4
→ La v2 falló: 1/4
→ La v3 por fin pasó: 4/4
El repo incluye todas las versiones fallidas con las transcripciones del judge en bruto.
Y los flowcharts no salieron de Fable describiendo cómo piensa. Grabaron las tool calls reales de agentes Fable y corrigieron el método cada vez que la autodescripción no coincidía con el comportamiento real.
La observación le ganó al self-reporting 3 veces.
Si usas Claude Code con modelos más baratos o quieres reforzar tus workflows de agentes, vale 10 minutos de tu tiempo.
Para instalarlo en Claude Code:
/plugin marketplace add Sahir619/fable-method
/plugin install fable@fable-method
Enlace al repositorio abajo👇
Silvio Balduzzi retweeted
Chris Rasmussen from the Nutanix Technical Marketing Engineering team presents an introduction for the new Nutanix SDK for Microsoft C#, expanding the Nutanix development and integration offerings. Learn how to get started now! nutanix.dev/2026/06/30/annou… #nutanix #sdk #v4api
Silvio Balduzzi retweeted
Did you know Nutanix now has an official C# SDK for our v4 APIs? To get things started, here's a very quick & introductory code sample that demonstrates using the SDK; more in-depth/tutorial C# SDK content is coming soon! nutanix.dev/code_samples/v4-… #nutanix #csharp #sdk #v4api
Silvio Balduzzi retweeted
Nutanix Cloud Manager (NCM) 2.x offers a unified platform to build, operate, and govern across hybrid environments with sovereign-ready options.
Learn more: bit.ly/3OioGpK
Introducing Claude Opus 4.7, our most capable Opus model yet.
It handles long-running tasks with more rigor, follows instructions more precisely, and verifies its own outputs before reporting back.
You can hand off your hardest work with less supervision.
Silvio Balduzzi retweeted
Announced at #NEXTconf, “Powered by Nutanix: Verified Solutions” provides a framework for Service Providers to deliver services aligned to Nutanix best practices and validated standards — bringing greater consistency and trust to customer-facing offerings: bit.ly/3Qqnd14
Silvio Balduzzi retweeted
What an incredible start to #NEXTconf! Nutanix highlighted strong ecosystem momentum, marking the first year with 100+ partners participating across infrastructure, end‑user computing, AI, and security.
Check out the full roundup of announcements: bit.ly/4siCgaA
Silvio Balduzzi retweeted
We’re thrilled to announce the winners of the 2026 Nutanix Partner Awards! 🎉
We’re proud of our global partner ecosystem and congratulate this year’s winners for setting the standard of excellence.
Join us in celebrating their success: bit.ly/4v9c70j
Silvio Balduzzi retweeted
We’re excited to welcome the Nutanix Partner community to #NEXTconf in Chicago!
This event wouldn’t be possible without the support of our phenomenal group of sponsors. Looking forward to a jam-packed week. 💜
StarWind V2V / P2V Converter Version 9 (build 848) – Multi-VM Conversions, Full CLI Support, Hot Migrations, and Cloud Improvements Make This Free Tool a Datacenter Must-Have vladan.fr/starwind-v2v-p2v-c… via @vladan @starwindsan
Silvio Balduzzi retweeted
🚀 Level up your Nutanix homelab! Just dropped a step-by-step guide on installing Prism Central 2024.3.1.13 on Community Edition 2.1.
From vanilla CE to full Prism power!
👉 is.gd/VEGOui
#Nutanix #PrismCentral #Homelab #vCommunity
Silvio Balduzzi retweeted
Long-context AI just got easier. Claude Opus 4.6 & Sonnet 4.6 now support 1M token context (GA), flat pricing, and 600 images/PDF pages per request. msft.it/6016Qmu6O
Silvio Balduzzi retweeted
MINISFORUM MS-A2 + ENPBA - GiovanniDominoni.it | You can Now Supercharge your Mini PC with 25GbE & Extra NVMe Storage! –#vExpert GiovanniDominoni.it dy.si/s2WFLU
Silvio Balduzzi retweeted
Nutanix Central is now available as an on-premises deployment—giving customers a powerful, unified management plane within their private datacenter. We’ve consolidated the complexity of distributed clusters into a single on-prem command center: bit.ly/4aTdtUV




























