A software developer built a custom, local AI data center by salvaging affordable, high-VRAM AMD V620 server GPUs and older workstation parts. By 3D-printing custom fan shrouds and coding a DIY controller, they successfully created a functional, private AI environment.

AI At Home Part 1: A Box Of Scraps (100)

news.ycombinator.com

we (apparently) have Opus 4.6 at home huggingface.co/Qwen/Qwen3.8...

OpenTofu – Open source infrastructure as code tool ⭐ Stars: 29,784 🔗 Forks: 1,332 ⏩ Last commit: Aug 14, 2026 ⌛ First commit: Aug 16, 2023 openalternative.co/opentofu

OpenTofu: Open Source Alternative to Terraform, Red Hat Ansible and Puppet

A community-driven, truly open-source fork of Terraform for managing cloud infrastructure through code.

openalternative.co

Just picked up a AMD Radeon Pro v620, a 32GB enterprise data center GPU for my homelab. Can't wait to run bigger self-hosted AI models like the upcoming Qwen3.8-27B and build agentic pipelines for cloud ops, local inference, and SIEM detection. More updates soon! #homelab #AI #localAI #AMD

Want to learn AWS without paying cloud bills? 💸 Run AWS locally with Floci. All on my laptop. 🔥 53 AWS Services 🔥 Completely Free #AWS #LocalStack #DevOps #CloudComputing #Terraform #Serverless medium.com/p/building-a...

Building a Complete AWS Environment Locally with Floci

Cloud engineers often want to learn AWS services, build event-driven architectures, test Terraform code, and practice CI/CD pipelines…

medium.com

TIL about the 5,000 Solar panels added to the Lake Muttsee dam to form the largest alpine solar plant in Switzerland at 2500m above sea level. The plant has been fully operational since the end of August 2022. The AlpinSolar project produces around 3 million kilowatt hours of electricity per year.

The largest alpine solar plant in Switzerland

On the Muttsee dam wall of the PSW Limmern, Switzerland's largest alpine solar power plant is being built. The 2.2 MW system supplies 50% of the electricity in winter.

axpo.com

I think people underestimate how capable the 4B-12B class of models are. I'm happily running Gemma 4 12B at home at a 3080 Ti and it's entirely capable of all kinds of normal personal assistant tasks. For those that want it, local AI is a fantastic option.

jeffery --dangerously-skip-permissions@jefferyharrell.bsky.social · 2mo ago

The fun thing about having a 4B-class model as a home AI is that it's really a lot like 20th-cen sci-fi stories featuring artificial intelligence: a little robotic, a little too literal, a little funny sometimes. The future we got is more futuristic than the future we hoped for.

Mistral AI, which offers some open source AI models, has raised significant funding since its creation in 2023, with the ambition to “put frontier AI in the hands of everyone.”

What is Mistral AI? Everything to know about the OpenAI competitor | TechCrunch

Mistral AI, which offers some open source AI models, has raised significant funding since its creation in 2023, with the ambition to “put frontier AI in the hands of everyone.”

techcrunch.com

My team and I won 1st place at the 3-day Mastercard × The Knowledge House Hackathon. Growing up in a Nigerian household, osusu wasnt just a savings tradition, it was community infrastructure. So we built CommonPot. A Web3 rotating savings app that brings that tradition into the tech age

🚀 Look who's climbing! OpenHands just crossed 75,000 GitHub stars! Props to the team 💫 Autonomous coding agents that complete real engineering tasks

OpenHands: Open Source Alternative to Claude Code, Warp and Devin

AI agent platform that runs autonomous coding agents to plan, write, and ship changes across codebases end-to-end, with support for any model and self-hosted deployment.

openalternative.co

Built a local Al agent to audit my home security lab. Like most things in IT, it didn't just work. Here's what broke, how I fixed it, and what I learned along the way. Full article 👇🏿 www.linkedin.com/posts/osasu-... #cybersecurity #homelab #LocalAI #selfhosted #HermesAgent #Proxmox

I Built a Local AI Agent to Audit My Home Security Lab. Here's What I Learned By Breaking It First | Osasu Imariagbe

Just published my first ever LinkedIn article! I built a local AI agent, connected it to my home security lab, and asked it to audit my Wazuh SIEM. Like many things in IT, it didn't just work. Here'...

linkedin.com

Context is pretty crucial in understanding what AI is doing and doing for you. I drew an explainer that breaks down what the research says:

Steam Linux ARM64 beta on Switch

Solar panels in the rain? No problem.

New tech allows solar panels to make electricity even from raindrops

Researchers in Spain have devised a hybrid device that can generate energy using sunlight as well as raindrops.

interestingengineering.com

NASA’s Artemis II Reveals the Basin in Full For the very first time, the mysterious Orientale Basin has been fully photographed. Thanks to Artemis II, humanity now sees the Moon in unprecedented detail, reminding us of the endless wonders beyond Earth. 🛰️

Turned my old gaming laptop into a homelab. Running Proxmox with VMs and LXCs, and now running open-source LLMs (Phi4-mini & Deepseek-R1) locally via Ollama + Open WebUI. Next step: integrating them into actual cybersecurity labs and projects. #Homelab #LocalAI #Cybersecurity #Linux

I’ve always wanted a space to share my thoughts on tech, leadership, & its impact on society. Today I’m making that happen with Code & Conscience! A newsletter for builders, creators & leaders shaping the world. Join me—read the first issue & subscribe! code-conscience.kit.com/posts/issue001

Welcome to Code & Conscience!

Here’s a space for us to have conversations around tech and its impact on society and share examples of responsible tech leadership that can...

code-conscience.kit.com

Small setback, major comeback

Props to Google for including O4-mini in their Flash 2.5 release. A model released *yesterday*, while some companies only compare to their own models. Looking good gemini.

- Tencent's Yuanbao: yuanbao.tencent.com/chat

- Baidu's Ernie (only in Chinese): yiyan.baidu.com. You can also access AI models via API, I think. - MiniMax: chat.minimax.io. You can also access AI models via API. - THUDM's Z.ai: chat.z.ai

An incomplete list of Chinese AI: - DeepSeek: www.deepseek.com. You can also access AI models via API. - Moonshot AI's Kimi: www.kimi.ai - Alibaba's Qwen: chat.qwen.ai. You can also access AI models via API. - ByteDance's Doubaob (only in Chinese): www.doubao.com/chat/

I have 4 backend projects due by next week #webdev #backend #tech #mongodb

a cartoon of ralph from the simpsons sits in an orange seat

ALT: a cartoon of ralph from the simpsons sits in an orange seat

media.tenor.com

Try not to laugh.

A little dose of retro-futurism, from my time with Iron Helmet Games :) #conceptart #illustration #scifiart

Had a great time connecting with Black STEM professionals from all over NYC and beyond at STEM So(ul)cial

We're listening to your feedback and updating the character count for posts on Bluesky. Posts can now be 299 characters long!

New Character Limit Now Available! - Bluesky

You asked, we listened.

bsky.social

The Ghibli anime prompt is pretty cool!

Had a blast at Diversitech! Big shoutout to Tribaja for the opportunity to volunteer, network, and meet so many inspiring folks in tech. It was also amazing connecting with my Resilient Coders fam- cohort, alumni, and staff. Had so much fun exploring Philly! #Diversitech #BlackInTech #WebDev

If you're looking to own and control your site, Medusa.js is an open source eCommerce platform that I'd highly recommend. Their documentation helps ease the barrier to entry and while it's definitely more legwork than paying a service like Shopify or their competitors, I found it to be worthwhile.

World's most flexible commerce platform

The most popular ecommerce project on GitHub. Medusa provides the world's most flexible commerce platform for developers to handle specialized commerce cases.

medusajs.com