your AI agent isn't failing because it can't execute
it's failing because it doesn't know when to stop
one bad decision becomes ten more, each made with absolute confidence
AI agents are getting faster at doing things nobody should've approved
Articul8's Bivia is built to make decisions across workflows and escalate when uncertain
sometimes the smartest decision is knowing when to ask for help
🤝 Paid partnership
Introducing Bivia.
A multimodal AI decision model built to complete entire workflows, not just answer individual questions.
We've spent 3+ years building and deploying AI in production across semiconductors, energy, manufacturing, and other complex industries.
Today, we're bringing that experience to everyone.
berni retweeted
ACABAN DE CREAR UNA HERRAMIENTA QUE CONVIERTE A GROK EN UN HACKER
Suena a clickbait. No lo es.
Shannon es un repo open source que apuntas a tu app y hace lo mismo que haría un atacante: intenta entrar
Y si no consigue un exploit real, no te lo reporta. Nada de listas infinitas de "posibles" fallos.
→ Exploits reales
→ Informe en PDF + SARIF
→ Corre en GitHub Actions
→ También en GitLab CI
→ Funciona con tu Grok
Con Grok 4.6 encontró 10 vulnerabilidades verificadas en Photoview, cero falsos positivos
Encontró el mismo fallo crítico que herramientas de $4.000. Puede revisar tu app cada vez que la actualizas.
Y no estás atado a un modelo: Grok, Claude, GPT, DeepSeek o modelos locales
48K+ estrellas y no pagas nada extra si ya tienes Grok.
Repo abajo :)
berni retweeted
Moving speech generation onto the device changes more than latency, it changes how a product handles connectivity and inference costs.
This breakdown of @hojoHQ’s Hojo-TTS-Light puts those trade-offs into practical terms.
berni retweeted
1 MINUTE AND 25 SECONDS BEHIND THE SCENES OF GRIFFIN.
The @tavus model that listens and talks at the same time, reads your gestures and clones a voice from 10 seconds of audio.
Shot with Griffin running live on set. Shared by @jessetrowe 👇🏼
berni retweeted
SI TIENES UNA AGENCIA, ESTO TE AHORRA MUCHAS HORAS
VEED acaba de lanzar su API de doblaje con lip sync.
El mismo vídeo, en otro idioma, con la misma persona hablando.
🔖 Guárdalo.
THE BTS OF AN AI MODEL IS NOW A CONTENT GENRE ON ITS OWN
and honestly, it works
seeing how something gets filmed with griffin running live on set adds a layer of weirdness no product trailer can capture
this is what interests me most about launches now: not the demo, the making of
berni retweeted
🚨 ACABAN DE LANZAR UN MCP QUE TE CONECTA TODOS LOS MODELOS DE IMAGEN, VÍDEO Y AUDIO DESDE CLAUDE
se llama Sunra API.
la idea: una sola integración y desde Claude (o cualquier cliente MCP) llamas a Seedance, Kling, GPT Image, Nano Banana y los que vayan sumando.
→ cada modelo al precio de API de su propio creador
→ sin markup, sin suscripción
→ pay as you go
lo interesante no es el catálogo. es que no tienes que gestionar una API key por cada lab, ni mantener integraciones separadas, ni comparar precios entre plataformas que le meten recargo.
una key. un MCP. todos los modelos al precio del lab.
enterprise: hasta un 40% de descuento sobre el precio de API de los propios modelos.
ESTO CAMBIA EL DOBLAJE PARA SIEMPRE
VEED lanza una API que dobla el vídeo y sincroniza los labios con la nueva voz.
→ la boca encaja con el audio
→ tono y emoción intactos
→ audio de fondo conservado
yo la usaría para sacar mis vídeos en español en otros idiomas sin regrabar.
A small model is interesting; knowing where it fits in a real product is more useful.
This article explores @hojoHQ’s local TTS through the questions teams actually face: hardware, response time, and workload limits.
We pay to run AI in datacenters while the Mac on our desk can already run many of those models.
Neywa Labs comes out of stealth today to use that power: local, open AI on your own machine.
The datacenter, on your desk.
🤝 Paid partnership
Today we're coming out of stealth, introducing Neywa Labs 🚀
We built the infrastructure to build and deploy AI that runs on the devices people already own. Private, offline, unlimited.
One mission: make AI available to everyone in the world.
neywalabs.ai/
WAIT. we also need to check the thing checking the agent? 😭
🤝 Paid partnership
Today, @JudgmentLabs opens for general access.
We believe long running agent systems will generate more exhaust data than the world can currently process.
Hidden within agent traces are rich records of how to make these systems better: what users ask for, when agents stumble, and how agents interact with the external world.
Judgment is the most performant monitoring and improvement layer for long-horizon agents.
We’re excited to see what you build!
A VERIFIED RESULT CAN STILL BE UNFINISHED
lean accepts a theorem proved under an extra assumption
it looks like progress. it isn't
axiom's swarm hit this early:
→ agents assumed prerequisites
→ tens of thousands of lines deleted
→ milestones restarted
→ reviews compare statement vs target
if you ship ai-written code, "passes the checker" and "does the right thing" are two different questions
🤝 Paid partnership
