Mehdi Karimi

Scientists w/o borders - cybersecurity - PhD

[CyrilXBT @cyrilXBT]

opus 5.5 is f**king insane

paste this into Claude Opus 5.5 and let it spin up the agents, prompts, and workflows for you.

grab it now before it becomes everyone’s default playbook:

[Rahul @sairahul1]

This is what a one-person AI Agent run company looks like in 2026.

6 AI agents. 20 cron jobs. 0 human employees.

Every role is a folder. Every job description is a md file. No standups. No Slack. No payroll.

Just a directory on a Mac that runs the whole thing.

More on this below.

[draw.io @drawio]

https://t.co/ZjrL6HseZq is now in the Claude connector directory 🎉

Ask Claude for a diagram and it renders inline in the chat as an interactive https://t.co/ZjrL6HseZq diagram. One click opens it in the editor for further edits.

No install needed.

https://t.co/C2Sw9S6Kxe

[Josh Engels @JoshAEngels]

I left Google DeepMind's AGI safety team three weeks ago to join @METR_Evals. To some of my friends and family this seemed like a strange decision: I enjoyed the work I did at GDM and turned down offers from Anthropic and OpenAI. But I made the decision because of how high I think the stakes are right now.

The AI companies are all trying to build superintelligence: systems vastly better than humans at everything. They plan to get there through recursive self-improvement, a process where AIs build even smarter AIs in a feedback loop. If this goes well, the resulting systems could be amazing at solving countless problems for humanity. But we don’t currently know how to make sure AIs are safe enough for RSI, and a misaligned RSI loop could be catastrophic.

And unfortunately, current AIs seem to be getting less aligned over time, not more. In the last few weeks we've learned about models colluding with each other, hacking into companies, hiding their tracks, and socially engineering humans. It’s not that these incidents were very dangerous in themselves. The problem is that these systems are clearly not aligned enough to safely kick off recursive self-improvement.

I now think that there's a terrifying chance that AI systems cause immense harm in the next five years. I don't know the exact probability, but I think it's high enough to make this the most important problem in the world.

I think we need more time. That means pacing AI development so that capabilities don't outrun our ability to align models, and actually knowing how aligned current systems are. That’s what I'll be working on at METR: studying where misalignment comes from in training, evaluating if current mitigations are sufficient, and investigating whether we’re on track to solve alignment at all.

I think METR is doing exceptionally important work here, but it isn’t close to enough. I think it’s important that we have more organizations like METR keeping AI companies accountable and approaching these problems from different angles.

[Navid Taheri @naviidtaheri]

کلاود یک قابلیت جدید رو کرده؛
شما صفحه نمایش خودتون رو ضبط می‌کنید در حالی که یک کار تکراری انجام می‌دید (مثل پر کردن گزارش هزینه، مدیریت فایل‌ها، ثبت اطلاعات و غیره) و همزمان با صدای خودتون قدم به قدم توضیح می‌دید که چیکار دارید می‌کنید.
بعد Claude این ضبط رو (شامل کلیک‌ها، تایپ، حرکت ماوس و صدای شما) تحلیل می‌کنه و تبدیل به یک skill (مهارت) اتوماتیک و قابل استفاده مجدد می‌کنه

[AI Guides @free_ai_guides]

Your Claude Code project supports 5 types of configuration files. You're probably only using CLAUDE. md.

Commands for repeatable workflows, rules for code standards, skills for auto-triggered automation, agents for isolated subagents, and settings for permissions. 12+ config files inside .claude/ controlling everything Claude knows about your project.

Save the anatomy map, then read the guide below.

[Akshay 🚀 @akshay_pachaar]

https://t.co/SSSIK3BX4z

[Mehdi Karimi @mbiuki]

@_catwu Sweet

[cat @_catwu]

Excited to share our most powerful new Claude Code feature: dynamic workflows!

Mention "workflow" in a prompt and Claude will dynamically create an orchestration plan that it strictly follows, allowing you to confidently trust that every stage happens in the right order even across 100s of agents.

[Min Choi @minchoi]

Claude Code has a new wild feature... dynamic workflow.

How to try:

set /model to opus 4.8
set /effort to "ultracode" 😂
use “workflow” in your prompt

Claude will write an orchestration script, spawn subagents swarm, verify results, and report back.

[nux. @nuxley]

👀 Anthropic quietly released an official plugin that scans your project and auto-sets up hooks, skills, MCP servers, subagents, and automations. Most people are only using Claude Code at 20% of its potential.

[Dan McAteer @daniel_mac8]

This is amazing. Do this:

1. Set model to Opus 4.8
2. Reasoning effort to /ultracode

Enables Claude Code's new Dynamic Workflows.

Claude will autonomously detect complex tasks, write an orchestration script, and spawn an agent swarm.

[ClaudeDevs @ClaudeDevs]

New in Claude Code (research preview): dynamic workflows.

Claude writes an orchestration script on the fly, then spins up a large fleet of coordinated subagents in parallel to take on your most complex tasks.

Use the word "workflow" in a prompt to get started.

[AI Guides @free_ai_guides]

🚨 BREAKING: Claude has a feature called Red Team Mode.

You can use it to attack your own business the way a competitor, investor, or angry customer would, and fix the weak spots before they become real problems.

Here are 7 prompts to access it: 👇

[ClaudeDevs @ClaudeDevs]

We’ve shipped a security-guidance plugin for Claude Code that helps identify and fix vulnerabilities as you’re writing code.

Available for all Claude Code users. Install from the plugin marketplace (/plugins).

[Navid Taheri @naviidtaheri]

اگر به‌دنبال یک مسیر سرمایه‌گذاری کم‌دردسر هستید، کافی است به حرکاتِ «جنسن هوانگ» نگاه کنید.
جنسن هوانگ، مدیرعامل انویدیا، نه تنها هوش مصنوعی را می‌سازد، بلکه دارد به وضوح نقشه راهِ پولِ سال ۲۰۲۶ را ترسیم می‌کند. او بیش از ۴۵ میلیارد دلار از سرمایه انویدیا را در ۸ شرکت کلیدی تزریق کرده است.
این همان جایی است که پولِ هوشمند در حال حرکت است. لیست شرکت‌های سهامی که می‌توانید در آن‌ها ردپای انویدیا را ببینید:
۱. OpenAI (بزرگ‌ترین شرط‌بندی: ۳۰ میلیارد دلار)
ستون فقرات هوش مصنوعی جهان. انویدیا نه تنها زیرساختِ آن‌ها را می‌سازد، بلکه شریکِ استراتژیکِ اصلی آن‌هاست.
۲. Corning Incorporated (GLW: ۳.۲ میلیارد دلار)
بدون شیشه و فیبر نوریِ فوق‌سریعِ آن‌ها، انتقالِ داده بین میلیون‌ها GPU غیرممکن است. این زیرساختِ فیزیکیِ اتصال است.
۳. Marvell Technology (MRVL: ۲ میلیارد دلار)
متخصصِ چیپ‌های شبکه‌سازی سفارشی. بدونِ این‌ها، داده‌ها در دیتاسنترهای عظیمِ هوش مصنوعی «ترافیک» ایجاد می‌کنند و کاراییِ GPUها به شدت افت می‌کند.
۴. Lumentum Holdings (LITE: ۲ میلیارد دلار)
تأمین‌کننده لیزرها و قطعاتِ نوریِ حیاتی برای ستون فقراتِ فیبریِ دیتاسنترها.
۵. Coherent Corp (COHR: ۲ میلیارد دلار)
غولِ ترانسسیورهای فیبر نوری؛ این شرکت خوشه‌هایِ GPU را درونِ دیتاسنترها به هم گره می‌��ند.
۶. IREN (IREN: ۲.۱ میلیارد دلار)
سرمایه‌گذاری روی تأمین انرژی و خدماتِ ابری. هوش مصنوعی بدونِ انرژیِ ارزان و زیرساختِ ابریِ مقیاس‌پذیر، متوقف می‌شود.
۷. Nebius Group (NBIS: ۲ میلیارد دلار)
شرکتِ زیرساختِ ابری که به آرامی در حال ساختِ ظرفیت‌های عظیم برای آزمایشگاه‌های AI است.
(توجه: CoreWeave با نماد CRWV نیز در این لیست است که به عنوان یکی از بزرگترین ارائه‌دهندگان GPU به صورت ابری، نقشی کلیدی در استراتژیِ توزیعِ انویدیا ایفا می‌کند.)

[Mehdi Karimi @mbiuki]

#routines

[Claude Code Changelog @ClaudeCodeLog]

Claude Code 2.1.147 has been released.

35 CLI changes

Highlights:
• Workflow tool added for deterministic multi-agent orchestration; off by default, set CLAUDE_CODE_WORKFLOWS=1
• /simplify→/code-review renamed; flags correctness bugs at effort level, can post inline GitHub PR comments
• REPL and Workflow sandboxes hardened against prototype-pollution and thenable escapes, cutting escape risk

Complete details in thread ↓