Rafa

@NousResearch Thank you for the dashboard plugin system! Building my personal life crm front tmr

Hermes Agent v0.11.0 - “The Interface Release” Full changelog below ↓

@BonaVictor Parece bem genial. Os escopos de verificação sao diferentes ou sao só runs sequenciais pra pegar problemas após apontamentos nas runs anteriores?

@rafadmachado Toda spec uma vez concluída, é validada em uma nova sessão por 10 agentes individuais. A ideia é garantir 100% de fidelidade entre o código e a spec! Uso esse padrão tanto no meu trabalho, quanto na guaracloud.com

Depois de muita consideração e testes, achei o modelo que vou usar como fallback e parceiro do claude code! Estarei usando o GLM-5.1 da z.ai! Testei diversos modelos e ferramentas e nenhum deles me surpreendeu tanto quanto o GLM-5.1 via opencode: - Codex + GPT-5.4: legal, me pareceu excelente em analise estática de código, mas infelizmente, muito lento e propenso demais a fugir das minhas specs. - Cursor Composer: Preço impossível de se manter pro meu uso, mas gostei. - MiniMax: promissor, mas não me impressionou - Qwen via AI Farm local: o pouco contexo mata um pouco, mas me impressionou demais como modelo local! - Qwen via Cloud: contexto maior, e meu terceiro favorito atrás do Opus e GLM, talvez considere usar as vezes! Excelente - Gemini: palha demais, melhorem google! - Grok: legalzinho pra fazer specs e discutir arquitetura, mas nada impressionante pra codar. Todos os testes foram feitos em cenários semelhantes ao meu fluxo de trabalho: spec -> planning -> decomposition de tasks -> multi-agent execution -> 10x harness validation -> testes automaticos e analise em staging -> deploy proposal -> deploy rehearsal -> deploy. De longe, o GLM-5.1 foi oque se saiu melhor, recomendo a todos darem uma olhada! Estarei trocando o plano Lite para o Pro/quarterly.

@Falconve_ I'm running openclaw on a raspberry pi with a node on my macbook just for accessing a full chrome instance. Hermes supports paid browsing services, but don't have anything similar to control my chrome profile instance. do you have that need, and how did you solve it?

been seeing a lot of people switching from OpenClaw to Hermes. here's my honest take after running both in production for a month 🧵

@lucaronin @linuz90 Hey Luca, my openclaw intance have a crm system sharing some of the concepts you are using on tolaria. Would love to join forces sometime, are you opensourcing it?

@linuz90 can't believe we have been building almost the same thing! 😄 tolaria.md

I built my dream Markdown editor for Mac. → Introducing Cogito (pronounced koh-gee-toh). It started out of frustration: Obsidian is powerful but overwhelming. iA Writer is beautiful but feels built for a different era. Nothing felt right for how I actually write and work now: plain files, lots of folders, agents and scripts editing alongside me. I wanted both: native and beautiful, powerful and calm. So I finally built it. It's fast, keyboard-first, polished, truly native. A Mac app built with power users and developers in mind. This is my love letter to writing and Mac apps. I use it for all my writing now. Free while in beta ✌️

@PedroCerize “Beba menos” foi foda

Beba menos e treine !!! Lula 2026 😂😂😂😂

Nothing against Anthropic but let us show Apple some respect.

@forwarddeploy working on Sec: the idea is to have tasks as a primitive. those are organized in projects that serve as a context boundary. tasks can be recurrent via a cron expression or one time. agents needing human to do something they post a human task. heartbeat gets task done

Building your own version of OpenClaw or productivity tool that uses agents? Want free xAI API credits to supercharge it with Grok? Reply below (or DM if stealth mode) Hackathon MVPs, side projects, wild experiments - let’s see ’em all! 🦞

@r0ck3t23 It sounds bs to me. OS of course gives people shared knowledge and allows the community to create better software over time. But it completly ignores the fact that most people that consume and use open source sw is not there for the code. They want the value!

Dario Amodei just dismantled the biggest myth in the AI industry. Open source AI isn’t free. It never was. Amodei: “It’s not free. You have to run it on inference and someone has to make it fast on inference.” For decades, open source meant something real. It meant a teenager in a basement could download the same tools as a Fortune 500 company. Could read the code. Could modify it. Could build something that competed with the giants. That was genuine democratization. That actually happened. AI is different. Fundamentally. Physically. In ways the ideology hasn’t caught up to yet. Downloading the weights is the easy part. The part that actually costs something is turning the weights into a running system. Into responses. Into intelligence operating in real time at scale. That requires compute. Power. Infrastructure. The kind measured in billions of dollars and years of construction. Amodei: “These are big models. They’re hard to do inference on. Ultimately you have to host it on the cloud. The people who host it on the cloud do inference.” The open source debate was never about who owns the model. It was always about who owns the cloud. And Amodei goes further. When a competitor drops a new open model, he doesn’t ask whether it’s open or closed. He doesn’t care about the licensing. He doesn’t engage the ideology. Amodei: “I don’t think it mattered that DeepSeek is open source. I think I ask, is it a good model? Is it better than us at the things that matter? That’s the only thing that I care about.” That’s the ruthless clarity of someone actually trying to win. While the media debates licensing frameworks, Amodei is asking one question. Is it better. Everything else is a distraction. Amodei: “I don’t think open source works the same way in AI that it has worked in other areas. Here we can’t see inside the model.” This isn’t Linux. You can’t read it. You can’t fork it. You can’t understand it the way generations of developers understood the tools they inherited. You can download it. And then you need a data center to run it. The teenager in the basement who was supposed to be empowered by this revolution needs a billion dollars of infrastructure before the empowerment starts. The era of the basement coder rewriting civilization on a laptop is over. The future belongs to whoever commands the compute, owns the power grid, and can actually turn the intelligence on. Open weights without infrastructure isn’t democratization. It’s a promise the physics of the universe won’t let us keep.

Whenever i ask my personal AI to add a skill

@TheMattBerman @openclaw @rapidapi 1000

I replaced a $200K GTM hire with @openclaw 😱 here's the system that runs my outbound: step 1: mine LinkedIn engagement → @rapidapi scrapes everyone engaging with niche content → someone who commented on specific posts = 10x warmer step 2: enrich + verify → Hunter/Apollo finds the decision-maker + email → @Perplexity deep research pulls signals like hiring, fundraising, media appearances, quotes step 3: score against your ICP → title, company, signals = ranked 0-100 → only A-tier leads get touched step 4: write personalized outreach → Claude writes outreach referencing what they ACTUALLY engaged with and talk about step 5: send via @instantly_ai → 3-email sequence. automated follow-ups. step 6: pre-call deep research → @PerplexityComet builds a 1-page briefing 30 min before every call input: your ICP + niche keywords output: booked meetings with people who already care $200K/year GTM engineer → $130/month in APIs. I packaged the entire system as the First 1000 Kit: - all 8 @openclaw skills - every prompt - tool-by-tool setup - email sequences that convert giving it away free. comment 1000 + like + follow (must follow so i can DM)

@mattpocockuk Never thought it was an agent. My subagents dont suffer exploring the codebase, tho. Grep and a repo.json summary do wonders

@rafadmachado Claude Code uses an Explore subagent for pretty much all exploration tasks. You'll see it as Explore(Task name) in the UI

Here's my AI coding workflow and all the skills I'm using: Idea -> /write-a-prd -> PRD PRD -> /prd-to-issues -> Kanban Board Kanban -> ralph​.sh -> Ralph Loop Ralph Loop -> Manual QA Links below to skills

@mattpocockuk Hum, not aware of the explore subagent, what is it?

@rafadmachado I don't love the fact that the subagent can't itself spin up subagents. So you lose the benefits of the Explore subagent, for instance.

@AkitaOnRails Essa é versão atualizada da famosa desculpa de call center: “o sistema está com uma instabilidade”

@johnheyerdal 👀 looking forward to it. Thanks, friend!

@rafadmachado I really like that. Thank you for sharing. I am currently working on a children's book and your comment resonates.

I feel myself prepared to build and explore the societal value AI is unlocking, and I think that because of my capacity to connect different knowledge together. But this reason feels strangely shallow, especially when thinking how i educate my children. Although not entirely actionable, your framework is a more complete thinking model I can use. As of why it is an effective model for preparing them, i don’t think it is. But it is a great framework for building upon.

@rafadmachado I appreciate that sentiment. What makes you think this is an effective model for preparing your kids for the years to come?

@OpenAIDevs Thanks! That is a very useful article

We just announced new primitives for building agents. Here are 10 tips on running multi-hour workflows reliably 👇 developers.openai.com/blog/skills-sh…

@anumeta10 @sama I totally agree. Maybe govern is a word that lost its meaning over time, or maybe I’m overestimating what govern should mean, but for me it means exactly that, determining what a good outcome means and acting over time to make sure that keeps being pursued.

I think that framing is close — but one layer is missing. AI won’t simply replace execution while humans “govern.” What’s emerging is a shift in *where cognition happens*: • Humans set direction, constraints, and values. • AI handles exploration of solution space at speed. • Governance becomes continuous oversight, not occasional approval. In business-critical decisions the bottleneck won’t be capability — it will be verification. The real evolution is not “vibe coding,” it’s *intent → simulation → human judgment loops*. The winners won’t be those who let AI run freely, but those who design systems where AI proposes and humans remain the final error-check. So yes — set intent, let AI execute — but the future role of humans is less about control and more about defining what counts as a good outcome.

GPT-5.3-Codex is rolling out today in Cursor, Github, and VS Code!