Big month in open source AI. Here's what caught my eye: 1. IBM dropped Granite 4.1 - An 8B model matching 32B MoE performance. Smaller, faster, cheaper — and it's open source. This is the kind of efficiency gains we're talking about when we say "own your AI." https://firethering.com/granite-4-1-ibm-open-source-model-family/ 2. Zig project says no to AI contributions - The Zig programming language officially banned AI-generated code from their project. Does "open source" mean human-only contributions? This conversation isn't going away. https://simonwillison.net/2026/Apr/30/zig-anti-ai/ 3. PyTorch Lightning malware (Shai-Hulud) - Someone shipped themed malware through the PyTorch Lightning AI training library. Check your dependencies, even the popular ones. Supply chain attacks are getting smarter. https://semgrep.dev/blog/2026/malicious-dependency-in-pytorch-lightning-used-for-ai-training/ 4. Claude Code vs OpenClaw drama - Anthropic's coding tool started refusing requests or charging extra if your commits mention "OpenClaw" (an open source competitor). Paid AI locking out open source. Sound familiar? https://twitter.com/theo/status/2049645973350363168 5. Fine-tuning can unlock copyrighted content - Research showed that fine-tuning on specific data can make LLMs recall copyrighted material. Alignment whack-a-mole continues. https://github.com/cauchy221/Alignment-Whack-a-Mole-Code 6. Mike: open-source legal AI launched - Legal AI going open source. Another domain getting the open-source treatment. https://mikeoss.com/ What was your favorite April AI story? Drop it below 👇

Hey Open Source AI Builders, If you're new here, it's time for an intro! Our community is powered by Hamish Aman Prakash – a recognized force in Australia's AI and tech ecosystem. Here's why you can trust the guidance, insights, and vision you'll find in this group: - Hamish is a *4-time award-nominated entrepreneur* in technology and artificial intelligence, known for pioneering projects in automation, open source, and local AI solutions. - He's built A-Tech from the ground up – delivering real AI products, consulting for businesses regionally and nationally, and championing a privacy-first, "own your AI" approach through Project Infra. - Hamish is constantly featured across platforms (including LinkedIn: linkedin.com/in/hamishfromatech), sharing innovations, practical demos, and leadership in AI adoption. - His influence and impact have made him a thought leader in Australia's AI community—bridging the world of no-code/builders and real technical mastery. - As a founder, mentor, and teacher, Hamish helps others scale with AI, break down technical barriers, and access life-changing opportunities in an AI-driven future. Let's welcome new members, share our stories, and always aim higher—knowing we're learning from the best in the business. Leave your questions for Hamish below, or connect with him directly to learn more about his journey! 🚀

Guys you have to get into training large language models it's just so much fun! And yes. AI can now play video games for you while you work!

When Anthropic introduced the Model Context Protocol Server to the world in November 2024, it felt like AI could actually become useful by connecting to real world apps that we use everyday to get things done for us. However, I felt something different. While connecting every app you use to AI models is the norm in 2026. I feel the MCP server can do so much more than that.. From what I can see, traditional business requires more than just connecting apps to an AI model and calling it a day. So I spent 2 weeks researching and building in silence to see what I could come up with.. I created 22 MCP servers : - Business focused tools that are simple, yet very effective - Pure Python Based - No external dependencies - Can work in air gapped environments or areas where there's no internet access at all - Data stored and powered by JSON files with TOON enabled for context compression. - Light enough to run on mobile phones!! Programming languages like Python can really show you the power of MCP Servers that actually work and evolve with you. Feel free to fork the repo and make your own pure python mcp servers! Access the MCP Servers on Github: https://github.com/hamishfromatech/business-mcp-servers

Would love to know so I can tailor better content for you all :) Let's me know in the comments