El comentario que no le gusta a tu ai tech boy fav...
@yummy_vzla de la mano de @metavarce se pusieron la del 10 y ya cubrieron casi todos los ángulos que tu nueva app vibecodeada intenta resolver
Mejor redirijan a la gente hacia la iniciativa de la startup más grande del país
Me quito el sombrero ante @yummy_vzla! su respuesta a la crisis en Venezuela ha sido superlativa
En 48 han habilitado:
•Viajes gratis a hospitales y clínicas en Caracas (con compensación total a los conductores: yummy paga el viaje)
•Héroes / First responders: personal de emergencia viaja gratis registrándose en https://t.co/Igot4XJQVV.
•Donaciones con matching: dona en https://t.co/240AsqLOVO (tarjetas, ACH, pago móvil) y Yummy agrega 25% extra (hasta $100.000)
•Plataforma SOS: reporta daños estructurales con fotos y ubicación (https://t.co/7ZtsgU93OD). Ingenieros voluntarios evalúa las estructuras
•Servicio solo donde sea seguro (evaluando constantemente) para proteger a los yummers
Gracias Vicente y todo el equipo de Yummy, ojalá más empresas hicieran asi sea una fracción de lo que ustedes están haciendo
@Juanegron@cristianmock Creó una app para centralizar la informacion https://t.co/8mdYOCWOBk Necesitamos equipos de rescaste en las zonas criticas. DIFUNDE POR FAVOR 🙏🏻
"Counter"
Porque un ruso se encontro unas ruinas abandonadas y se dio cuenta de que por alguna razón las conocía de memoria: el lugar es igual al mapa De_Dust2 del Counter-Strike
🚨 Claude puede hacer SEO como una agencia de $10.000/mes (gratis).
Y casi nadie lo está usando.
Aquí tienes 7 prompts que todo negocio local debería usar:
🔖Guárdalo. Lo vas a necesitar.
Meet MetaClaw 🦞— Just talk to your agent, it learns and evolves.
💬 Conversations become training trajectories
⚡ Models update live with hot-swapped weights
🧠 Failures generate new reusable skills
💻 No GPU cluster required
Under the hood:
🔄 Online SkillRL training
Training runs asynchronously while the model continues serving.
🐚 Skill evolution
When the agent fails, an LLM analyzes the trajectory and generates new reusable skills.
📌 Skill injection
Relevant skills are retrieved and injected into the system prompt at each step to guide behavior in real time.
Built on Kimi-2.5 via Tinker cloud LoRA
→ fine-tuning costs about $10 and requires no GPU cluster
💻 Fully open-sourced
https://t.co/RR24ZJvZ7T
Built with @openclaw and @thinkymachines
Kudos to the team @richardxp888, Jianwen Chen, @Xinyu2ML, @lillianwei423, @StephenQS0710, Zeyu Zheng, @cihangxie!
OpenClaw meets RL!
OpenClaw Agents adapt through memory files and skills, but the base model weights never actually change.
OpenClaw-RL solves this!
It wraps a self-hosted model as an OpenAI-compatible API, intercepts live conversations from OpenClaw, and trains the policy in the background using RL.
The architecture is fully async. This means serving, reward scoring, and training all run in parallel.
Once done, weights get hot-swapped after every batch while the agent keeps responding.
Currently, it has two training modes:
- Binary RL (GRPO): A process reward model scores each turn as good, bad, or neutral. That scalar reward drives policy updates via a PPO-style clipped objective.
- On-Policy Distillation: When concrete corrections come in like "you should have checked that file first," it uses that feedback as a richer, directional training signal at the token level.
When to use OpenClaw-RL?
To be fair, a lot of agent behavior can already be improved through better memory and skill design.
OpenClaw's existing skill ecosystem and community-built self-improvement skills handle a wide range of use cases without touching model weights at all.
If the agent keeps forgetting preferences, that's a memory problem. And if it doesn't know how to handle a specific workflow, that's a skill problem. Both are solvable at the prompt and context layer.
Where RL becomes interesting is when the failure pattern lives deeper in the model's reasoning itself.
Things like consistently poor tool selection order, weak multi-step planning, or failing to interpret ambiguous instructions the way a specific user intends.
Research on agentic RL (like ARTIST and Agent-R1) has shown that these behavioral patterns hit a ceiling with prompt-based approaches alone, especially in complex multi-turn tasks where the model needs to recover from tool failures or adapt its strategy mid-execution.
That's the layer OpenClaw-RL targets, and it's a meaningful distinction from what OpenClaw offers.
I have shared the repo in the replies!
No sé si esto realmente funcione o no, pero me entristece no haber sabido esto hace años para intentarlo con mi madre antes de que falleciera.
Si ustedes tienen algún familiar enfermo espero que puedan evaluar otras alternativas para ayudarlos a salvarse.