Pinned Tweet
Absolutamente recomiendo usar Cache Control en Claude Code
Probablemente uno de los Mods que te ahorrará la mayor cantidad de tokens y ayudará a mantener tus sesiones activas por mucho más tiempo
Añade esta barra para la caché de 5 minutos de Claude y envía notificaciones para ayudar a mantener la caché activa
Comando de instalación:
npx claude-code-templates@latest --mod observability/prompt-cache-control
¡Cada usuario de Claude Code NECESITA instalar /stack... En serio!
Guarda esta publicación y pruébala más tarde
El mod /stack te muestra TODO lo instalado y disponible en Claude Code:
Habilidades, Subagentes, Comandos, Ganchos, Servidores MCP, Plugins, Configuraciones y Estilos de Salida
Todo organizado por ámbito: Enterprise, Usuario, Proyecto, Plugin, Flags de CLI
Realiza un escaneo completo y muestra los resultados en un panel interactivo donde puedes ver qué está activo, qué está deshabilitado y qué componentes están siendo sobrescritos por otros
Finalmente tendrás una mejor comprensión de por qué tus sesiones de Claude Code consumen tantos tokens y por qué Claude se comporta como lo hace
También puedes explorar todos los comandos, habilidades y subagentes integrados que vienen con Claude Code, ¡incluyendo algunos que ni siquiera se mencionan en la documentación oficial!
Instálalo con:
npx claude-code-templates@latest --mod ui/stack-viewer
21
429
Models can ace every benchmark and still freeze on stuff like:
"timing violation 3 days before tapeout. waive it, re-route, or slip the date?"
That's judgment, and it lives in how experts actually decide. Nobody trained on it.
@arcfront is building exactly that data with 300+ enterprises. Under a year in, already at a $1B run rate.
Wild team too. @anshuman_ssh applied to one college, made ICPC World Finals twice. Overnight success that took 10 years.
🤝 Paid partnership
Today we're launching @arcfront with @Naman_Bhalla and deep support from @asxna.
In under a year, four of the five frontier labs became customers.
In 2025 an AI solved all 12 problems at the ICPC World Finals. The best human team solved 11. I competed in two of those finals, so I know how big a feat that is.
And yet, put that same AI inside a company as a finance manager, or in almost any other job, and it fails on day 1.
Why is that?
Stefan Friedman retweeted
Today we're launching @arcfront with @Naman_Bhalla and deep support from @asxna.
In under a year, four of the five frontier labs became customers.
In 2025 an AI solved all 12 problems at the ICPC World Finals. The best human team solved 11. I competed in two of those finals, so I know how big a feat that is.
And yet, put that same AI inside a company as a finance manager, or in almost any other job, and it fails on day 1.
Why is that?
NUEVO en /shot: pide a Claude que te guíe a través de algo en una app y señala cada paso directamente en tu pantalla
Aquí encuentra el código no utilizado en la página principal de Amazon en Chrome DevTools (menú ⋮, Más herramientas, Cobertura). 5.3 MB de los 9.2 MB de JS y CSS no se habían ejecutado
/shot es un mod de código de Claude y MCP que le da a tu agente acceso a herramientas de capturas de pantalla
Cobertura te muestra exactamente qué JS y CSS en tu página no se usan (prueba ejecutándolo en la primera carga en tu sitio)
/plugin marketplace add davekiss/mods
/plugin install shot@davekiss
Stefan Friedman retweeted
Introducing Higgsfield GPU Harness.
Rent on-demand GPUs in Claude and ChatGPT to render, run open-weight models and train LoRAs.
Let several AI agents work on your professional apps in the cloud at once, from any device.
Black Friday deal, live now: get GPUs at up to 64% OFF.
GTA 6 took 13 years.
Don't spend 13 hours editing it.
One long stream becomes a week of clips, in 1 click.
One raw capture becomes a video you're proud to post.
Live now in OpusClip ↓
You play GTA 6. We clip it.
One long GTA 6 stream becomes a week of clips, in 1 click. One raw capture becomes a video you're proud to post. Play it rough. Post it polished. Two new gaming Recipes, live now in OpusClip.
>MiniMax H3 free and unlimited on Envato for 24h
>open the tab "just to test a couple prompts"
>it's going to be 3am, isn't it
🤝 Paid partnership
Stefan Friedman retweeted
the most prepared person in my next meeting is going to be the Teamily AI agent and i'm fine with that
#VideoAgent #AIGlass #PhysicalAgent
Introducing Teamily Video Agent
We give all Teamily agents a face.
They also see the world you see through your AI glasses and your camera.
#VideoAgent #AIGlass #PhysicalAgent
Stefan Friedman retweeted
digital agents, then AI glasses, then embodied skills, then everyday robots.
that's Teamily AI's roadmap, and today they previewed Teamily AI Glass. the agent keeps its identity and memory, it just gets to see your world.
#AIGlass #VideoAgent #PhysicalAgent
Introducing Teamily Video Agent
We give all Teamily agents a face.
They also see the world you see through your AI glasses and your camera.
#VideoAgent #AIGlass #PhysicalAgent
Stefan Friedman retweeted
these guys just casually hacked into the most important frontier models.
including one that was supposed to be impossible to attack.
and then they published the results.
🤝 Paid partnership
Today we’re releasing FrontierBreach, a new standard for attack success rate on agentic workflows in the enterprise.
Existing benchmarks systematically underrepresent the security risks facing modern AI agents.
FrontierBreach reflects attack scenarios on realistic enterprise workflows, while using sophisticated attack chains that result in cyber consequences (ransomware, data exfiltration, phishing, etc).
Pasar pruebas de tu agente de codificación aún no te dice si el cambio funciona
Claude Code incluye una habilidad /verify que compila y ejecuta tu app para verificar el cambio, y sus documentos dicen que lo hace "sin recurrir a pruebas o verificaciones de tipos"
Solo se ejecuta cuando lo solicitas, así que hazlo la última cosa antes de considerar un cambio terminado o intégralo en tu flujo de trabajo
Ejecuta esto una vez para enseñarle cómo se lanza tu proyecto, luego /verify después de cada cambio:
/run-skill-generator
every agent demo i've watched this year cuts away right before the captcha
agents keep hitting captchas and blocks on the web.
today we're launching tilion (@ycombinator): the world's fastest and stealthiest browser infra for agents.
#1 in stealth
#1 in agent action speed
#1 in start-up time
dont believe it? tell your agent:
setup tilion.com
AI CAN DRAFT A CONTRACT AND STILL CHOKE ON A SCANNED INVOICE
Unsiloed just raised $7.5M from Felicis to fix the reading part. Scans, tables and spreadsheets turned into something an agent can query like a database.
Least exciting pitch in AI right now. Probably the one I'd bet on.
🤝 Paid partnership
We’ve raised a $7.5M seed round led by Felicis so AI can read documents as easily as it reads databases.
Finance, law and healthcare run on documents, and not databases like any other industry.
Soon AI agents will read and write more documents than humans, and they’ll need to query them the way software queries tables.
We’re building the computable layer that makes this possible.
To do this at scale, we've brought together ML researchers and engineers from Google DeepMind, Cognition, and quant trading firms.
We use computer vision and dual-stream models to turn complex multimodal documents into formats LLMs and AI agents can understand and reason over.
If your team wants to turn messy documents into reliable, AI-ready data, book a demo:
unsiloed.ai/book-demo
The meetings that actually matter never happen on Zoom.
They happen over coffee or on the walk back to the car, and nobody's taking notes.
Wispr Flow put Notetaker on iPhone for exactly those, free, and the notes sit next to the ones from your desktop calls.
Next coffee meeting I'm hitting record.
🤝 Paid partnership
Stefan Friedman retweeted
agents keep hitting captchas and blocks on the web.
today we're launching tilion (@ycombinator): the world's fastest and stealthiest browser infra for agents.
#1 in stealth
#1 in agent action speed
#1 in start-up time
dont believe it? tell your agent:
setup tilion.com
every bad release i've shipped had a PR that passed review
diff looked fine, checkout didn't load
curious how many of those Verify would've caught
🤝 Paid partnership



















