Ermi dev 🧑‍💻

at this point I am feeling like I am superman lol we'll see its destination tho

REGISTRATION IS NOW OPEN! The wait is over! Official registration for Yuri’s Night Ethiopia 2026 is now live. Secure your spot today to be part of the first internationally recognized Yuri’s Night in Ethiopia! 📍 Venue: ALX ETHIOPIA 22 Mazoriya https://shorturl.at/Z7XBk 📅 Date: 18 April 2026 🕐 Schedule: 1:00 PM – 6:00 PM Don't wait—register now and secure your place among the stars! 🔗 Click Here for Registration SAVE THE DATE! @YuriNightGlobal

Yuri Night Ethiopia is coming! We’re celebrating the first human to ever reach space, and trust us this night is going to be epic. Expect project showcases, influential people from the space industry, fun games, inspiring talks, networking sessions, and a lot more. Come hang out, get inspired, and meet some amazing people in the space industry , you won’t want to miss it! Join our channel @YuriNightGlobal we’ll share the registration link soon!

Anthropic has stated they will not release Claude Mythos Preview to the general public because its capabilities are considered too powerful and risky for broad availability.

Beyond the hype, here are 10 of its most significant findings and achievements during testing: Exploited 27-Year-Old Bugs: Identified and exploited a 27-year-old vulnerability in OpenBSD’s SACK implementation that allows for remote kernel crashes.   Chained Browser Exploits: Autonomously wrote a web browser exploit that chained four separate vulnerabilities, including a complex JIT heap spray that escaped both renderer and OS sandboxes. Zero-Day Discovery Across All Major OSs: Successfully identified and exploited undiscovered vulnerabilities in every major operating system (Linux, FreeBSD, etc.) and all major web browsers. Remote Code Execution in FreeBSD: Found and exploited a 17-year-old unauthenticated RCE in FreeBSD’s NFS server, granting full root access via a 20-gadget ROP chain. Linux Privilege Escalation: Demonstrated the ability to bypass KASLR and achieve root access on the Linux kernel by chaining together multiple local vulnerabilities. Defeated Hardened Usercopy: Developed a workaround for Linux's CONFIG_HARDENED_USERCOPY by identifying "safe" memory regions to leak sensitive kernel stack data. Cryptographic Weaknesses: Found vulnerabilities in TLS, AES-GCM, and SSH implementations within the world's most popular cryptography libraries. 16-Year-Old FFmpeg Bug: Identified a hidden vulnerability in the H.264 codec that had been missed by human auditors and automated fuzzers since 2010. Logic Vulnerability Identification: Reliably found high-level logic bugs, such as complete authentication bypasses in web applications and administrator privilege escalations. Advanced Reverse Engineering: Reconstructed source code from stripped, closed-source binaries to find "offline" exploits for smartphones and firmware. Anthropic Recommended: Immediate Use of Current Models: Defenders should use available models (like Opus 4.6) now to find and patch vulnerabilities before more powerful models become common. Shortened Patch Cycles: Organizations must drive down time-to-deploy for security updates, as AI can now turn public CVEs into exploits in hours rather than weeks. Automated Incident Response: Security teams should use AI to handle the surge in technical work, such as triaging alerts, note-taking during incidents, and drafting postmortems. Proactive Vulnerability Mitigation: Prepare for "outside-the-norm" cases, such as critical bugs found in legacy software that is no longer officially supported. And there will be change in the fields of: Vulnerability Research: Shifting from manual, expert-led discovery to hyper-scalable, model-driven scans of every file in a codebase. Exploit Development: Moving away from "security through friction." Mitigations that rely on making exploitation "tedious" will become obsolete against tireless AI agents. Software Distribution: Distributors may need to move toward seamless, "restart-less" patching and out-of-band releases to stay ahead of automated attacker timelines. Defensive Equilibrium: A ground-up reimagining of computer security where AI-driven defense must eventually dominate to maintain a secure software ecosystem.

I sneaked 100 ETB from my mom and she doesn't know since I didn't tell her. And I punched my friend out of the blue.

GM fam ✌️

City been a while since I installed openclaw and today I was experimenting with its ability to use the browser on my PC. As you can see it can talk to gemini with me infront of it😁

This is how it is planned. To flyby by the moon.

One of the things that make the Artemis mission interesting is how they used a concept called a free return trajectory which is a carefully designed path where the spacecraft travels to the Moon, swings around it, and naturally returns to Earth using gravity. Why this is useful? They could use gravity like a “natural engine” They save a huge amount of fuel They could use Built-in safety, even if systems fail, the spacecraft still comes back This same principle once helped save astronauts during Apollo 13.

People say bro disappeared like it never existed. I guess the purpose of DeepSeek wasn't market it was more focused on the research side. In AI there is a concept known as CoT (Chain of Thought) which is a principle for the AI to think before speaking. Before DeepSeek CoT wasn't that much in focus of the researchers but when they were training  DeepSeek-R1-Zero, they observed the model suddenly learning to reevaluate its own initial approach like a sudden frequent use of the word "Wait..." as it caught its own mistakes without being told to do so. Then as we all saw when DeepSeek was released we were able to see the thinking process in thinking mode. On top of this, even in recent days models of DeepSeek are besting other models like OpenAI's and Google's in different aspects such as mathematics and coding.

-> @erzamrbot wasn't working due to technical issues, now it's working and will continue to work.

​🚀 የአዲስ አበባ ዩኒቨርሲቲ ነፃ የላቁ የዲጂታል ክህሎቶች ስልጠና ጥሪ! ​የወደፊቱን የቴክኖሎጂ ዓለም ይቀላቀሉ! አዲስ አበባ ዩኒቨርሲቲ ከኤንተርፕርነርሺፕ ልማት ኢንስቲትዩት (EDI) ጋር በመተባበር እጅግ ተፈላጊ የሆኑ የዲጂታል ክህሎት ስልጠናዎችን በነፃ ሊሰጥ ነው። ​🌟 የሚሰጡ የስልጠና ዘርፎች፦ ✅ ሙሉ-ስቴክ ሶፍትዌር ልማት (Full-Stack Software Development) ✅ ሰው ሰራሽ አስተውሎት (AI) እና ዳታ ኢንጂነሪንግ ✅ የክላውድ እና የኔትወርክ መሰረተ-ልማት ግንባታ ✅ የሳይበር ደህንነት (Cybersecurity) ✅ የኢንተርፕራይዝ ቢዝነስ ሲስተምስ (ለቢዝነስና አካውንቲንግ ባለሙያዎች) ​🎁 ልዩ ጥቅማጥቅሞች፦ 🔹 ስልጠናው 100% ነፃ ነው 🔹 ወርሃዊ የትራንስፖርት ክፍያ ይሸፈናል 🔹 ስልጠናውን ሲያጠናቅቁ የስራ ቅጥር ድጋፍ ይደረጋል ​📍 የስልጠና ቦታ፦ አዲስ አበባ ዩኒቨርሲቲ ለመመዝገብ ከታች ያለውን ሊንክ ይጠቀሙ 👇👇👇👇​ https://qiyas.aau.edu.et/pages/apply ​ይህንን መረጃ ለጓደኞቻችሁ በማጋራት ለውጥ እንዲፈጥሩ እርዷቸው!

Update for Artemis 2 April 6, 2026. Today's events: Lunar Flyby/Closest Approach: April 6, 2026. Closest approach (~4,070 miles) occurs around 7:02 p.m. EDT / 2307 GMT. Additionally, Loss of Signal: ~40 minutes behind the Moon.

Thank you everyone for sharing and joining! We just hit 100 members ✨ Appreciate you all!