Open Source News: Security, AI, and Community Highlights

Open Source News: Security, AI, and Community Highlights

Security & Vulnerabilities A critical Linux kernel vulnerability, GhostLock (CVE-2026-43499), affects kernels 2.6.39 through 7.1, enabling stack use-after-free and local privilege escalation. Patches are being prepared. Expert-tested malware removal tools for 2026 are reviewed, emphasizing open-source options like ClamAV for cross-platform protection. Community & Events Africa Wiki Women Community Roundtable 2026 will discuss responsible AI … Read more

Open Source AI, Kubernetes, and Developer Tools Roundup

Open Source AI, Kubernetes, and Developer Tools Roundup

Insight: The Open Source AI Stack Matures This week’s digest highlights a significant shift: open source AI is moving from experimental to production-ready. From browser-based LLMs (WebLLM) to efficient PyTorch training techniques, the ecosystem is lowering barriers for developers. Meanwhile, platforms like OpenSearch and Kubernetes tools (Capsule) show that open source communities are solving real-world … Read more

PyTorch 2.13 Release Live Q&A

PyTorch 2.13 Release Live Q&A

Video by PyTorch via YouTube
PyTorch 2.13 Release Live Q&A

PyTorch 2.13 introduces updates across attention, compilation, distributed training, memory efficiency, Python support, and accelerator platforms. Highlights include FlexAttention support on Apple Silicon with up to approximately 12x speedup over SDPA on sparse patterns, the CuTeDSL "Native DSL" backend for key GPU operations, and nn.LinearCrossEntropyLoss to reduce peak GPU memory by up to 4x during large-vocabulary language model training.

On Wednesday, July 22, 2026, at 11 a.m. PT, PyTorch maintainers and contributors will provide a brief overview of the PyTorch 2.13 release and answer questions from the community live.

Topics will include:

– FlexAttention support on Apple Silicon and deterministic backward computation on CUDA
– The CuTeDSL "Native DSL" backend for Inductor
– nn.LinearCrossEntropyLoss for reducing peak GPU memory
– torchcomms for large-cluster training
– FSDP2 communication overlap improvements
– Torch wheel support for Python 3.15 on Linux, including free-threaded 3.15t builds
– Expanded ROCm, Arm, and Intel XPU platform support

The live Q&A will feature expert panelists Alban Desmaison, Andrey Talman, and Piotr Bialecki, with Chris Gottbrath moderating.

PyTorch 2.13 includes 3,328 commits from 526 contributors since PyTorch 2.12.

Register for the live Q&A today!

Source

What is knn_vector in OpenSearch? Semantic Search explained (with live demo)

What is knn_vector in OpenSearch? Semantic Search explained (with live demo)

Video by NetApp Instaclustr via YouTube
What is knn_vector in OpenSearch? Semantic Search explained (with live demo)

Learn how OpenSearch’s knn_vector field type powers semantic search — and see it running live on an OpenSearch cluster, all in under 6 minutes.

In this video, Ramya breaks down what knn_vector is, the key parameters you need to know when configuring it, and walks through a complete working example on OpenSearch Dashboards. By the end, you’ll know how to create a vector index, index a document with an embedding, and run your first semantic search query.

▶️ Full blog article (deep dive): https://www.instaclustr.com/blog/understanding-opensearch-vector-field-types-part-1-knn-vector/

▶️ Try NetApp Instaclustr free trial: https://console2.instaclustr.com/signup

▶️ How to create an OpenSearch cluster on Instaclustr: https://www.youtube.com/watch?v=2RU-YqW5zsk

NEW TO OPENSEARCH?

If you’re new to OpenSearch itself, check out the other videos on this channel for the basics — they’ll give you a helpful foundation for what’s covered in this video.

Source

Leveraging WebLLM: Private and Powerful On-Device Chatbots, Warat Wongmaneekit #FOSSASIA Summit 2026

Leveraging WebLLM: Private and Powerful On-Device Chatbots, Warat Wongmaneekit #FOSSASIA Summit 2026

Video by FOSSASIA via YouTube
Leveraging WebLLM: Private and Powerful On-Device Chatbots, Warat Wongmaneekit #FOSSASIA Summit 2026

Run Large Language Models (LLMs) directly in the browser with WebLLM and WebGPU—no cloud APIs required. In this FOSSASIA Summit 2026 session, Warat Wongmaneekit explores how to build fast, private, and cost-effective on-device AI chatbots using open-source models and modern web technologies.

Learn how browser-based AI works, how WebGPU accelerates LLM inference, and the architecture behind deploying high-performance AI entirely on the client side. Perfect for web developers, AI engineers, and anyone interested in privacy-first, edge AI applications.

FOSSASIA Summit 2026 held in Bangkok, is Asia’s leading Open Source tech conference featuring sessions on #AI, #Cloud, #DevOps, #Open Hardware, #Security, #Web #Mobile Technologies, #Web3, and #Databases. Learn more: http://summit.fossasia.org

Session slides: https://eventyay.com/ev/88882f3e/talk/BB8XBKPW3B5W/

#FOSSASIA #FOSSASIASummit #opensource #FOSS

Source

ChatLoopBackOff Episode 79: Capsule with Thomas and Oliver

ChatLoopBackOff Episode 79: Capsule with Thomas and Oliver

Video by CNCF [Cloud Native Computing Foundation] via YouTube
ChatLoopBackOff Episode 79: Capsule with Thomas and Oliver

In this episode, CNCF Ambassador Thomas Schuetz explores Capsule for the first time, with CNCF Maintainer Oliver Baehler live on stream to navigate obstacles and elaborate on its design. Capsule is a sandbox project providing multi-tenancy and policy governance for Kubernetes.

For its core capabilities, Capsule features tenant isolation by grouping namespaces into logical tenant environments within a single shared cluster. Its self-service delegation empowers tenant owners to provision resources safely within boundaries without cluster-admin access. For policy governance, it automatically applies resource quotas, network isolation, and RBAC rules across specific tenants. Finally, it supports a native ecosystem, leveraging native Kubernetes custom resources and admission webhooks without adding complex layers.

Join us for this practical, live dive into Capsule!

Source

How Humans Move and Interact with the World: NymeriaPlus

How Humans Move and Interact with the World: NymeriaPlus

Video by Meta Open Source via YouTube
How Humans Move and Interact with the World: NymeriaPlus

Jakob Engel, Director of Research at Meta Reality Labs Research presents “How Humans Move and Interact with the World at CVPR 2026 in Denver, Colorado. The talk covers the addition to the Nymeria dataset – NymeriaPlus.

🔗 Links & Resources:

Read the full paper: https://arxiv.org/abs/2603.18496
Download the dataset: https://github.com/facebookresearch/nymeria_dataset
Learn more about Nymeria: https://www.projectaria.com/datasets/nymeria/

#CVPR #CVPR2026 #ComputerVision #AI #Surreal #TechTalk #MachineLearning

Source

Why LLMs Still Fail: Reasoning, Robustness, and Safety with Nouha Dziri

Why LLMs Still Fail: Reasoning, Robustness, and Safety with Nouha Dziri

Video by Open Data Science and AI Conference via YouTube
Why LLMs Still Fail: Reasoning, Robustness, and Safety with Nouha Dziri

LLMs can solve Olympiad-level math, complex coding problems, and expert benchmarks—but still fail in surprisingly simple ways.

In this ODSC AI East 2026 keynote, Nouha Dziri, PhD, Senior Research Scientist at Cohere Labs, explores what it takes to build more trustworthy large language models. Drawing on her research in reasoning, post-training, and AI safety, Nouha examines the limits of current LLMs, why “jagged intelligence” creates real-world reliability challenges, and how better training and evaluation can improve model behavior.

The session covers key challenges in LLM deployment, from reasoning failures and out-of-distribution generalization to reinforcement learning approaches that strengthen capabilities and scalable safety systems like automated adversarial testing and moderation.

Watch to learn how researchers are working to advance LLM capabilities while making them safer, more robust, and more reliable in practice.

————————————————————————————————————-

Visit our website and choose the nearest ODSC event to attend and experience all our training and workshops: https://odsc.ai

To watch more videos like this, visit https://aiplus.training

Sign up for the newsletter to stay up to date with the latest trends in data science: https://opendatascience.com/newsletter/

Follow us online!
• Facebook: https://www.facebook.com/OPENDATASCI
• Instagram: https://www.instagram.com/odsc/
• Blog: https://opendatascience.com/
• LinkedIn: https://www.linkedin.com/company/open-data-science/
• X (twitter): https://x.com/_odsc

Source

Ask The Hosts – Episode 38

Ask The Hosts – Episode 38

Video by The Late Night Linux Family via YouTube
Ask The Hosts – Episode 38

Support us on Patreon and get an ad-free RSS feed with some early episodes. https://www.patreon.com/LateNightLinux

The condiment we’d pick if we could only have one. With Allan from 2.5 Admins, May from Linux After Dark, and Shane from Hybrid Cloud Show. This is a short episode because Joe is having a summer break.

Patrons got this this in their feed two weeks ago.

https://askthehosts.com/ath38/

Source