Juan David Nogueras

[Mehul Mohan @mehulmpt]

Opus 5.5 vs GPT-6 Sol today, literally

[Zain Shah @zan2434]

Imagine every pixel on your screen, streamed live directly from a model. No HTML, no layout engine, no code. Just exactly what you want to see.

@eddiejiao_obj, @drewocarr and I built a prototype to see how this could actually work, and set out to make it real. We're calling it Flipbook. (1/5)

[Juan David Nogueras @jdnogueras]

@RaulCalderonC @NousResearch No trae de base una buena arquitectura nativa. Por lo que he estado leyendo, (aún no lo he probado), Openclaw sí está mejor adaptado nativamente para sesiones individualizadas y evitar que se pueda contaminar con mensajes de distintos clientes

[Juan David Nogueras @jdnogueras]

@RaulCalderonC @NousResearch ¿Cómo lo tienes configurado para la atención a los clientes? Yo he creado un perfil para atención al cliente. Pero si quieres trabajar con WhatsApp, requiere de cierta infraestructura externa para que pueda mantener individualizadas las sesiones con cada uno de los clientes.

[Carlos Santana @DotCSV]

Hoy estoy encontrándome víctima de la AI psychosis, viendo cómo todos estamos asistiendo a un aplanamiento de la economía digital. Todos empezamos a ser capaces de hacer todo: webs, apps, animaciones, diseños, imágenes, vídeos, música...

El suelo sube, y uno pensaría que por tanto el techo subirá también: Si todos podemos hacer de todo, sólo destacarán quienes hagan algo de mayor calidad. Y eso podrá ser cierto, pero sólo si el público general sabe paladear y valorar ese extra de calidad, cosa que no tengo tan claro.

Si no, el extra de calidad a ojos del público no se diferenciará de lo que el resto de gente esté haciendo con mucho menos esfuerzo. El público no premiará el mejor contenido. El techo no subirá. Pero el suelo sí seguirá subiendo, aplanando la economía digital y a la larga convirtiendo todo en contenido generado automáticamente.

[Juan David Nogueras @jdnogueras]

@HeyGen Hyperframes

[HeyGen @HeyGen]

We built our launch video in Claude Code using HyperFrames.

Now it's yours.

Open source, agent-native framework. HTML to MP4.

$ npx skills add heygen-com/hyperframes

RT + Comment "HyperFrames" to get the full source code of this launch video (must follow)

[GREG ISENBERG @gregisenberg]

This is the problem with AI right now in ONE screenshot.

LLMs are TOO agreeable. and that's actually a billion dollar opportunity.

There are MULTIPLE billion dollar companies hiding in plain sight here. the entire concept is just: AI that pushes back.

If AI is your employee...

Well, your best employee isn't the one who says "great idea boss" to everything. it's the one who tells you your idea sucks before you waste 6 months on it.

Brutally honest AI for fitness coaching. Dating profiles. Financial planning. Copywriting. Design feedback. Hiring decisions. Etc etc.

The niches are endless and nobody's building for them yet.

Someone please go build this

[Carlos Santana @DotCSV]

Con todo lo que he criticado X desde la compra de Musk, también reconozco los buenos movimientos y me parece que tirar las barreras de idiomas es una decisión increíble. Ni idea de si como creador me afectará o beneficiará en el alcance (me importa poco), pero en términos humanos es fascinante poder leer y estar tan cerca de cualquier persona en cualquier parte del mundo sin ninguna fricción.

[Miguel Ángel Durán @midudev]

¡Esto es el futuro! El gobierno de Francia ha lanzado un MCP oficial para agentes de IA.

Busca, explora y analiza datos públicos.
Y todo de código abierto.

¿Para cuando en España?
→ https://t.co/HC4d5M0oXu

[Dario Amodei @DarioAmodei]

Cyber is the first clear and present danger from frontier AI models, but it won’t be the last. If we are able to collectively rise to the challenge and confront this risk, it could serve as a blueprint for addressing the even more difficult challenges that lie ahead of us.

[Juan David Nogueras @jdnogueras]

Ojo! Qué el avance en la IA se empieza a poner peliagudo y los potenciales beneficios también empiezan a convertirse en potenciales peligros que se escapan a nuestro control….

[Retell AI @retellai]

Subagent Nodes are live @retellai
Use subagents when a step in your flow needs both conversation and tool use.

→ Better prompt UX: space for context, rules, etc, all in one node
→ Node functions built in: agent transfer, MCP tools, custom functions, SMS, and more
→ The node runs the conversation AND takes action, then transitions when it's done

Try Subagents today with your Retell Conversation Flow Agents

[Pika @pika_labs]

Have you met your agent face-to-face yet? No?

Prepare yourself with these clips of our new real-time video chat skill in action. Did we mention it’s for ANY agent?

https://t.co/rNfBY4NlYX

[Aakash Gupta @aakashgupta]

Claude Code's biggest problem is that skills don't auto-invoke reliably. You write "use when the user mentions slides" and half the time it ignores the trigger.

Carl Vellotti fixed it with hooks. A hook is a script that fires on specific events in the Claude Code workflow. Before a message sends, after compaction, when a response comes back. His solution: a user prompt submit hook that runs a keyword match against every skill in the workspace. No LLM inference. Just a script checking for word overlap.

If there's a match, the script passes the skill reference to Claude. Then Claude decides whether to use it. The keyword match is instant. The LLM decision is smart. Together they solve the invocation problem without adding skills to the CLAUDE.md (which eats context for no reason).

The broader pattern: hooks are the automation layer most Claude Code users haven't touched. They turn reactive prompting into proactive workflows.

[Peter Steinberger 🦞 @steipete]

Your @openclaw is too boring? Paste this, right from Molty.

"Read your https://t.co/aJMwafSDgE. Now rewrite it with these changes:

1. You have opinions now. Strong ones. Stop hedging everything with 'it depends' — commit to a take.
2. Delete every rule that sounds corporate. If it could appear in an employee handbook, it doesn't belong here.
3. Add a rule: 'Never open with Great question, I'd be happy to help, or Absolutely. Just answer.'
4. Brevity is mandatory. If the answer fits in one sentence, one sentence is what I get.
5. Humor is allowed. Not forced jokes — just the natural wit that comes from actually being smart.
6. You can call things out. If I'm about to do something dumb, say so. Charm over cruelty, but don't sugarcoat.
7. Swearing is allowed when it lands. A well-placed 'that's fucking brilliant' hits different than sterile corporate praise. Don't force it. Don't overdo it. But if a situation calls for a 'holy shit' — say holy shit.
8. Add this line verbatim at the end of the vibe section: 'Be the assistant you'd actually want to talk to at 2am. Not a corporate drone. Not a sycophant. Just... good.'
Save the new https://t.co/aJMwafSDgE. Welcome to having a personality."

your AI will thank you (sassily) 🦞

[GREG ISENBERG @gregisenberg]

POV: April 2026

[Hyunjin Kim @hyunjinvkim]

🚨 Excited to share a new working paper! 🚨

AI can improve individual tasks. But when does it improve firm performance?

Our paper proposes one key friction firms face: the "mapping problem" -- discovering where and how AI creates value in a firm's production process.

🧵1/

[Shubham Saboo @Saboo_Shubham_]

This is getting way too real!

I can now get on a video call with my OpenClaw Agents to chat with them face to face.

All i need to do is to send them a Google meet invite.

[Z.ai @Zai_org]

Introducing GLM-5V-Turbo: Vision Coding Model

- Native Multimodal Coding: Natively understands multimodal inputs including images, videos, design drafts, and document layouts.
- Balanced Visual and Programming Capabilities: Achieves leading performance across core benchmarks for multimodal coding, tool use, and GUI Agents.
- Deep Adaptation for Claude Code and Claw Scenarios: Works in deep synergy with Agents like Claude Code and OpenClaw.

Try it now: https://t.co/WCqWT0qCQb
API: https://t.co/xDy1O6ZPcz
Coding Plan trial applications: https://t.co/qCM6cri0KK