Marcin Piekarski 🇦🇺 @codewithmarcin

AI SDK 7 is now available. Introducing: reasoning control, agent-level tool approval, tool and runtime context, file and skill uploads, MCP Apps, durable workflows, terminal UI, sandbox support, harness integrations, telemetry, lifecycle events, and more.

This indie dev is making a game where you herd masses of lost sheep to a secret refuge - Play as a loyal sheepdog - Uncover a mysterious calamity that only you can stop - Run fast and free in a beautiful but troubled world It's called The Free Shepherd. Would you play this?

"Most agents don't learn, they just leave traces." In 12 minutes, @jakebroekhuizen breaks down how to actually close the loop. Surface issues with LangSmith Engine Write memory updates back to Context Hub Let the agent actually improve between runs If you're thinking about how your agent should handle memory at scale, give this a watch.

Jake Broekhuizen @jakebroekhuizen

a day ago

x.com/i/article/2069…

6 21 128 76K 350

everyone is talking about agent loops, harnesses, and self-evolving agents. but almost no one is talking about the actual hard part: you cannot run a company on one giant agent with every tool, every file, and no accountability. that's not autonomy. that's a fog machine. here's how we're building an agent company OS inside Matrix. — the stack: Workspace Brain → Matrix Runtime Orchestrator → Department Verticals → Department Lead Agents → Worker Agent Pool → Proof / Check-in Loop Matrix is not a chatbot. it's an operating system for autonomous work. — the workspace brain is the company boundary. it gets loaded with the things a real company actually runs on: → product docs → codebase context → chats, files, goals → operating rules → prior runs + examples of good work → approvals, memory, skills this isn't "context." it's the shared operating layer. it knows what the company knows, what it's trying to do, who owns what, what good looks like, and what must be proven before work counts as done. — on top sits the Matrix Runtime. it coordinates wake, cron, department messages, OKR state, permissions, worker dispatch, proof ledger, memory updates. under the runtime, work is organized into departments. a department is not a chat thread. it's a long-running agent with identity, memory, skills, goals, history, tool boundaries, taste, and accountability. Founder Strategy. Product Engineering. Growth. Ops. Research. each one has a lead agent that decides what happens, reads the relevant Memory Skill, breaks work into scoped tasks, and picks the right execution seat. — sometimes that seat is a native Matrix worker. sometimes Codex. sometimes Claude Code. sometimes a browser / computer automation worker. the point is not "one model does everything." the point is: → the right agent → with the right context → inside the right boundary → using the right tools → with a clear definition of done — this is why scoped workers matter. a "do everything" agent is too vague. but: → a release worker with repo context, tests, and approval gates → very good → a Codex worker scoped to one patch and one validation path → very good → a Claude Code worker doing deep repo analysis → very good → a browser worker with a specific flow and proof requirement → very good narrow scope reduces drift. Memory Skill keeps narrow agents from going blind. proof prevents fast output from pretending to be progress. — that is the loop: Workspace Brain → Department Lead → Worker → Artifact → Proof → Check-in → Memory Skill update every cycle, the company gets smarter. that's the real self-evolution. not a single agent rewriting its own prompt in a void — but a whole org compounding through proof. — each workspace is an isolated agent company. its own brain, departments, memory, workers, proof ledger. workspaces can talk when needed. but context should not bleed by default. isolation is not a limitation. it's what makes the system usable. — once a department pattern works, you fork the pattern — not the raw context. you still customize memory, examples, approval gates, tools, voice, definition of done. but you're not starting from zero. you might already have 70% of the OS for that kind of work. — what this actually changes: a small team of strong operators can now run surfaces that used to require entire departments. but only if the agents are actually good. and good agents don't come from connecting more tools. they come from source material, taste, iteration, narrow scope, workflow design, proof, memory, and human judgment. vague agents just create vague output faster. Matrix is our attempt to build the opposite: an agent company OS where autonomous work has structure, memory, ownership, and proof. the loop is the product.

Take one hour tonight to follow my Hermes tutorial below to install the app, set up your Telegram bot, enable voice, connect Google Workspace, and create a few routines. I made it super beginner friendly and walked through all the steps live in 45 min. You won't regret it. 📌 Watch here: youtube.com/watch?v=2ZacwC…

Peter Yang @petergyang

a day ago

I love using Hermes from @NousResearch as my 24/7 AI chief of staff because it's reliable, free, and open source. Here's my new 45-minute free course that covers how to: → Install Hermes safely → Connect Telegram, voice, Google Workspace, and more → Create routines like

24 47 464 92K 740

I'll be demoing AI growth systems this Thursday on Cyber Fund 15 minute live and then Q&A I´m thinking about showing either: closed growth loops that run autonomously, handle the same job every time. these pull from a company brain, execute with tools, and operate without you in the loop (example SEO content) or open agents doing exploratory work, deep research on competitors, markets, angles. you stay in the loop, guide direction, and they go wide pulling context you didn't know to look for (example competitive research) what would you rather see live, the closed loops running growth on autopilot or the open agents doing the exploring?

cyber•Fund @cyberfund

3 days ago

While you're still prompting one task at a time, these founders have agents running their companies. Self-healing codebases. Coding agents you launch from your phone. A full GTM on autopilot. This Thursday, they show you how.

7 8 58 14K 34

🚀 Vite 8.1 is out ⚡ Experimental Full Bundle Mode: ~15x faster dev server startup & ~10x faster full reloads on large apps 🧩 Chunk Import Map to keep chunks stable & improve caching 🦀 WASM ESM integration 💡 Lightning CSS on the road to default 👀 and more! Blog post below.

The loops "debate" isn't one. "Loops are dead" (@mattpocockuk, 174K views) and "just run a loop" (10K bookmarks) sound opposite. They're not. "Loop" is just doing three jobs at once. Routine: fixed steps, every time. Workflow: branches on what it finds. Loop: repeats and checks its own work until a condition holds. Most "agent loops" you see are one of the first two wearing the third one's name. 🧵

keyboard_arrow_left Previous keyboard_arrow_right Next

A full walkthrough of Conductor 00:00 What is Conductor? 01:50 UI 02:33 Adding projects 03:53 Setup and run scripts 05:25 Multiple agents and forking context 07:41 Running workspaces in the cloud 08:44 PRs, CI, and review workflow 11:14 Team workflows with branches and Linear

cloudflare/security-audit-skill: A coding-agent skill for multi-phase security audits with independently verified, machine-readable findings github.com/cloudflare/sec…

Two months ago I was fired by Google for creating the Google Workspace CLI. It went viral, hit #1 on Hacker News, gained thousands of GitHub stars and many thousands of actual users in just a couple days. It was an incredible, confusing journey, from directors and leaders asking what they could learn from the tool to getting grilled by legal about why the Google logo and brand colors are on the Google Workspace GitHub code repositories. I think the cause was that Workspace and certain leaders (and projects) were afraid of being disrupted. But the fear wasn't specific to my CLI, it was a broader fear in what agents meant for Workspace. Either way, the irony of my termination was the announcement at Google Cloud Next two days before I was fired that an official Workspace CLI was coming. I want this out there because it is easier for me to explain my story and it is an experience I want to fully own. It's also part of my healing. Nearly 7 years at Google was an incredible opportunity for me and I was fortunate to have wonderful teammates and a manager that fully supported me through these last few months. Thank you.

Remote / local handoff in Codex! Removing boundaries one at a time. When you let the model be in the driver seat, you actually need less infrastructure.

Guinness Chen @guinnesschen

6 days ago

Codex can now hand off threads between local and remote hosts. Start work on your laptop, send it to a remote box before you close the lid, bring it back later. And yes, Codex can orchestrate the handoff for you.

214 216 4K 982K 2K

Yesterday, I shipped v1 of mattpocock/skills, including: - /ask-matt: ask me how best to use the skills, from right in your agent - Model-invoked vs User-invoked skills - /writing-great-skills, which encodes my skill-writing best practices Enjoy:

💡DESIGN[.]md Best Practices DESIGN[.]md is quickly becoming one of the most useful files for AI-assisted UI design because it help AI understand your design system well enough to create UI that actually feels like your product. A good DESIGN[.]md usually includes your design tokens, brand style, color rules, typography, spacing, and component guidance. 3 best practices for creating a strong DESIGN.md: 1⃣ Treat tokens as decisions: A token is not just a variable. It defines the role something plays in the interface. 2⃣Add reasoning. Don’t just say what a style is. Explain why it exists, when to use it, and when not to use it. This helps AI make better design decisions. 3⃣Build components on top of tokens. Components should reference your tokens and design rules. 👇

Show Codex a workflow once. Reuse it as a skill. Record & Replay lets you show Codex a recurring task, like filing an expense report or submitting a time-off request. Codex turns that demo into an inspectable, editable skill. You control when recording starts and stops.

New in Claude Code: Artifacts. Interactive pages built from your session, like a PR walkthrough or a living project dashboard, shared with your team at a private link. Available in beta on Team and Enterprise plans.

Just launched Loop Library - a curated list of agent loops you can use right now. Find loops, submit your own, tokenmaxx!! signals.forwardfuture.ai/loop-library/

Microsoft Teams just became an AG-UI client 💬 Here's why that's big deal. Any AG-UI agent can use it as an interactive surface: Generative UI, Human-in-the-loop, Shared state Until now, it didn't exist within MS Teams. As a user, I need this. I spend a good part of my day chatting with agents through Slack, and the walls of text they return have been one of the biggest hurdles to adoption. AG-UI brings the same UX upgrades we brought to web apps to non web surfaces everyone uses. This is AG-UI and CopilotKit reimagined.

Before Fable 5 was shut down, it pushed Gemma 4 to 255 tok/s on WebGPU. Some didn't believe it was real. Today we're releasing the demo and kernels it wrote for you to see yourself. Run it locally in your browser. Agentic kernel optimization is the future of on-device inference

Xenova @xenovacom

2 weeks ago

I gave Fable 5 one job: write custom WebGPU kernels for Gemma 4 inference. It climbed to 84 tok/s, then hit a wall, insisting further optimization was impossible. Hours later, Anthropic rolled back invisible LLM development safeguards, and it hit 255 tok/s. The next day,

145 370 5K 1.1M 2K

Given how impressed I was with Cursor’s Thermo-Nuclear Code Quality Review skill, I decided to create a similar one, but specifically for animations. /review-animations to review, and /emil-design-eng to build animations. They complement each other well emilkowal.ski/skill