# Luca Berton

> Luca Berton is an AI Platform Engineering educator, Docker Captain, and author of 8+ books on Ansible, Kubernetes, Docker, and cloud-native automation. This file links to LLM-friendly entry points across his 1600 published tutorials, books, and talks.

Curated entry points into the site. Each link points to a clean Markdown version of the page (append `.md`) or a topic hub. Use the Optional section for secondary material when you need a shorter context.

## Services

- [AI Integration & Strategy](https://lucaberton.com/services/ai-integration/): AI readiness assessments, GPU infrastructure design, LLM deployment architecture, model governance
- [Kubernetes Workshops](https://lucaberton.com/services/kubernetes-workshops/): Hands-on K8s training, multi-tenant GPU clusters, operator development, production hardening
- [Automation Strategy](https://lucaberton.com/services/automation-strategy/): Ansible automation at scale, IaC with Terraform, CI/CD, GitOps
- [Cloud Infrastructure](https://lucaberton.com/services/cloud-infrastructure/): Multi-cloud architecture, Kubernetes migration, cost optimization, DR
- [Performance Optimization](https://lucaberton.com/services/performance-optimization/): Infrastructure audits, GPU utilization, observability, SLO-driven operations

## Books

- [Kubernetes Recipes](https://lucaberton.com/kubernetes-recipes/): Practical cookbook for Kubernetes operations and troubleshooting
- [Practical RHEL AI](https://lucaberton.com/practical-rhel-ai/): Hands-on guide to Red Hat Enterprise Linux AI with InstructLab
- [Ansible for Real-Life Automation](https://lucaberton.com/books/): Comprehensive Ansible guide published by Packt
- [Ansible Pilot](https://lucaberton.com/books/): Step-by-step Ansible automation

## Key Pages

- [About](https://lucaberton.com/about/): Background, certifications, and professional experience
- [Courses](https://lucaberton.com/courses/): Online courses on Ansible, Kubernetes, DevOps, and cloud platforms
- [Conference Speaking](https://lucaberton.com/conferences/): Past and upcoming speaking engagements
- [KubeCon Europe 2026](https://lucaberton.com/kubecon/): Talk on multi-tenant GPUs on bare metal, slides, and booking
- [GPU Cost Calculator](https://lucaberton.com/gpu-cost-calculator/): Compare cloud vs on-premises GPU infrastructure costs
- [Contact](https://lucaberton.com/contact/): Get in touch or book a consultation
- [Blog](https://lucaberton.com/blog/): 1600 technical articles on AI, Kubernetes, Ansible, DevOps, and cloud

## Conferences

- [Post KubeCon Japan 2026: Reflections on AI, Cloud Native, and the Developer Journey](https://lucaberton.com/post-kubecon-japan-2026-reflections/): Straight from KubeCon Japan 2026: the shift from CPUs to GPUs, exploding AI models, and the new accelerator landscape — what it means for developers.
- [My August 2026: KubeCon Japan, Red Hat Summit & More](https://lucaberton.com/luca-newsletter-august-2026/): August 2026 roundup: KubeCon Japan, Red Hat Summit, Hiroshima, Amsterdam meetups, and the AI platform trends I'm watching.
- [AI Builders Amsterdam: A Hands-On AI Meetup Recap](https://lucaberton.com/ai-builders-amsterdam-2026/): AI Builders Amsterdam recap: talks on validating side projects, docs for agents, and running local LLM agents at StartDock.
- [Hiroshima Peace Memorial Ceremony 2026: Attending in Person](https://lucaberton.com/hiroshima-peace-memorial-ceremony-2026/): I attended the Hiroshima Peace Memorial Ceremony on 6 August 2026. Here is what the minute of silence, 120 nations, and two hibakusha testimonies taught me.
- [Tech Makers Monthly Tech Drinks + Pitching: Amsterdam Recap](https://lucaberton.com/techmakers-monthly-tech-drinks-pitching-amsterdam-2026/): Recap of Tech Makers' Monthly Tech Drinks + Pitching night in Amsterdam: startup pitches on AI visibility, a pre-seed raise, and founder stories.
- [Let's Touch BASE: Getting Your Taxes Right as a Small Business](https://lucaberton.com/base-lets-touch-base-taxes-small-business-2026/): Recap of BASE Amsterdam's tax workshop for small business owners: VAT as a flow not a cost, common compliance mistakes, and a fireside chat on co-creation.

## AI

- [Fish Audio Raises $52M Seed for AI Voice Models](https://lucaberton.com/fish-audio-52m-seed-ai-voice-models/): Fish Audio's $52M seed funds AI voice models for creators and enterprises — 8M+ users, $21M ARR, and an open-source Fish Speech lineage.
- [BrowserStack Meetup Amsterdam: AI's Impact on QA](https://lucaberton.com/browserstack-meetup-amsterdam-ai-qa-2026/): Luca attended the BrowserStack Meetup Amsterdam 2026, exploring how AI is reshaping QA — from Claude + Playwright to GenAI in quality engineering.
- [Four UI & AI Tools I'm Watching in 2026](https://lucaberton.com/four-ui-ai-tools-2026/): Watermelon UI, Motion Primitives, Manus, and Haikei: four tools I'm watching for building interfaces, motion, autonomy, and design assets.
- [Claude Code Usage Limits: Session vs Weekly](https://lucaberton.com/claude-code-usage-limits-explained/): Claude Code's Usage page shows two limits: a rolling session and a weekly cap across all models, plus temporary boosts that revert to your plan baseline.
- [Soup: Fine-Tune an 8B Model on a 4 GB GPU](https://lucaberton.com/soup-cli-fine-tuning-llm-layer-streaming/): Soup is an open-source CLI that fine-tunes LLMs from one YAML and streams the base model layer-by-layer so an 8B model trains on a 4 GB GPU.
- [Model Soups: Better Fine-Tuning by Averaging Weights](https://lucaberton.com/model-soup-fine-tuning/): Model Soups average N fine-tuned checkpoints into one. No extra inference cost or latency, often beats the single best run. Recipe + PyTorch code inside.

## Platform Engineering

- [EDGECASE 2026: Gotta Catch 'Em All](https://lucaberton.com/edgecase-2026/): EDGECASE 2026 returns September 24 in Hilversum — the Netherlands' biggest Kubernetes event, 400+ cloud native engineers, hosted by Bart Farrell.
- [Astro Responsive Images: 50% Smaller with AVIF, WebP](https://lucaberton.com/astro-responsive-images-optimization/): Use Astro's Image component with formats [avif, webp] and a widths array for an automatic responsive srcset — about half the JPEG size.
- [YAGNI: You Aren't Gonna Need It](https://lucaberton.com/yagni-principle/): YAGNI — You Aren't Gonna Need It — an Extreme Programming rule against building features no current requirement demands. Less code, fewer bugs, faster delivery.
- [Netris Raises $15M to Automate AI Neocloud Networks](https://lucaberton.com/netris-ai-neocloud-network-automation/): Netris raised a $15M a16z Series A to automate AI neocloud networks: hardware-accelerated, deterministic, no AI in the product.
- [Kata Containers 4.0 Reinforces AI Agent Sandboxing with a New Rust Runtime](https://lucaberton.com/kata-containers-4-0-ai-agent-sandboxing-rust-runtime/): Kata Containers 4.0 ships runtime-rs, a Rust runtime replacing Go, hardening AI agent sandboxing with VM-grade isolation and lower startup latency.
- [Infrastructure as Code and GitOps for Networks](https://lucaberton.com/infrastructure-as-code-gitops-network-automation-2026/): Learn how declarative Infrastructure as Code and GitOps deliver safe, auditable, drift-free network changes with Terraform, Argo CD, and Batfish.

## Automation

- [Add a Live QR Code in Word](https://lucaberton.com/word-qr-code-displaybarcode-tutorial/): Use the native DISPLAYBARCODE field in Word to embed a real, updatable QR code that links to lucaberton.com — no add-in required.
- [SaltStack vs Ansible 2026: Speed, Scale, and Simplicity](https://lucaberton.com/saltstack-vs-ansible-configuration-management-2026/): SaltStack vs Ansible for configuration management at scale. Event-driven architecture, agent vs agentless, performance at 10K+ nodes, and migration guide.
- [n8n: Self-Hosted Workflow Automation with 400+ Integrations](https://lucaberton.com/n8n-workflow-automation-self-hosted-kubernetes-2026/): Deploy n8n on Kubernetes — the open-source Zapier alternative with AI agents, code nodes, and enterprise-grade workflow automation.
- [Event-Driven Ansible: 12 New Collections in 2026](https://lucaberton.com/event-driven-ansible-new-collections-platform-enhancements-2026/): Everything new in Event-Driven Ansible and AAP 2.7: 12 content collections, Azure/AWS event bus, gateway-only architecture, PostgreSQL 17, Django 5.2.
- [Ansible Automation Orchestrator (Q3 2026 Preview)](https://lucaberton.com/ansible-automation-orchestrator-ai-driven-it-operations-2026/): Red Hat's Automation Orchestrator (Q3 2026) combines event detection, AI reasoning, and deterministic execution in one governed pipeline.
- [Ansible Intelligent Assistant & MCP Server: BYOK RAG](https://lucaberton.com/ansible-automation-intelligent-assistant-mcp-server-byok-rag-2026/): Red Hat's Automation Intelligent Assistant (AIA) brings BYOK RAG, a hosted MCP server exposing AAP to LLMs, and multi-provider BYOM flexibility.

## Linux

- [Azure Linux 4.0: Microsoft's First-Party Linux in Preview](https://lucaberton.com/azure-linux-4-public-preview/): Azure Linux 4.0 is now in public preview: Microsoft's Fedora-derived, dnf5-powered, secure-by-default Linux built for Azure VMs, AKS, and containers.
- [How to Clean-Install the Latest Fedora on a Lenovo Laptop (No CD Drive Needed)](https://lucaberton.com/reinstall-fedora-lenovo-laptop-usb/): No optical drive, no problem. Wipe a Lenovo laptop and clean-install the latest Fedora Workstation from a USB stick, created on macOS, Linux, or Windows.
- [Scott McBrien on the RHEL Roadmap: 10.2, Zero-CVE Images, and DNF5](https://lucaberton.com/scott-mcbrien-rhel-roadmap-summit-2026/): Scott McBrien, Red Hat, on RHEL 10.2, Fedora Hummingbird zero-CVE images, the Long-Life Add-On's flexible lifecycle, and why DNF5 waits for RHEL 11 in 2028.
- [RHCE Exam 2026: Complete Red Hat Certified](https://lucaberton.com/mastering-the-red-hat-certified-engineer-rhce-exam/): Pass the RHCE exam on your first attempt. Complete study guide covering RHEL 9, Ansible automation, system roles, containers, and real exam objectives for 2026.
- [RHEL 9 Administration Guide: Everything](https://lucaberton.com/red-hat-enterprise-linux-9-for-sysadmins/): Complete RHEL 9 administration guide covering installation, networking, security hardening, systemd, containers, and cloud integration for sysadmins.

## Sports Tech

- [Nike Mind: When Footwear Design Meets Neuroscience](https://lucaberton.com/nike-mind-neuroscience-based-footwear-2026/): Nike Mind 001 and 002 use 22 moving foam nodes and EEG research to change sensory input from the feet, aiming for calm focus rather than speed.

## Open Source

- [Shuchi Sharma on Open Source AI's World of Infinite Possibilities](https://lucaberton.com/shuchi-sharma-open-source-ai-program-office-summit-2026/): Shuchi Sharma on leading Red Hat's Open Source and AI Program Office, operationalizing open source AI at enterprise scale, and a new foundation in the works.
- [KubeCon Japan 2025 Shows Japan's Cloud Native Moment](https://lucaberton.com/kubecon-japan-2025-cloud-native-growth-cncf-press-conference/): KubeCon + CloudNativeCon Japan 2025 sold out in Tokyo. The bigger story is Japan's rising cloud native maturity, end-user adoption, and AI demand.
- [Ghost CMS: Headless Publishing Platform on Kubernetes](https://lucaberton.com/ghost-cms-headless-self-hosted-kubernetes-2026/): Run Ghost as a headless CMS with native memberships, newsletters, and a blazing-fast Node.js backend — self-hosted on Kubernetes.
- [Cal.com: Open-Source Self-Hosted Scheduling Platform](https://lucaberton.com/cal-com-open-source-scheduling-self-hosted-2026/): Complete guide to Cal.com — the open-source Calendly alternative with full API access, custom workflows, and Kubernetes deployment options.
- [Medusa: Open-Source Headless Commerce Engine for Developers](https://lucaberton.com/medusa-open-source-headless-commerce-2026/): Build custom e-commerce with Medusa — modular architecture, multi-region, multi-currency, and extensible with plugins. The Shopify alternative.
- [Gitea: Lightweight Self-Hosted Git Server on Kubernetes](https://lucaberton.com/gitea-self-hosted-git-server-kubernetes-2026/): Deploy Gitea as a GitHub alternative — lightweight Go binary, CI/CD with Gitea Actions, package registry, and Kubernetes Helm deployment.

## database

- [Vik Bhatti (Primer): Move Before You Hit the Write Ceiling](https://lucaberton.com/vik-bhatti-primer-cockroachdb-migration-roachfest-2026/): Vik Bhatti of Primer on zero-downtime CockroachDB migration, rejecting self-managed sharding, and a post-cutover write bug.
- [When Kubernetes Pod Disruption Budgets Aren't Enough for Multi-Cloud CockroachDB](https://lucaberton.com/kubernetes-pod-disruption-budgets-multi-cloud-roachfest-2026/): Why cluster-scoped Kubernetes Pod Disruption Budgets can silently break quorum for multi-cloud CockroachDB, and how a custom cross-cluster PDB fixes it.
- [Surviving a Real GCP Outage: A Multi-Cloud CockroachDB Story from RoachFest London 2026](https://lucaberton.com/gcp-outage-resilience-cockroachdb-roachfest-2026/): A RoachFest London account of a real GCP outage with zero platform impact thanks to a multi-cloud CockroachDB architecture.
- [Spencer Kimball: From 'We're Not Google' to CockroachDB's Agentic Future](https://lucaberton.com/spencer-kimball-cockroachdb-agentic-vision-roachfest-2026/): Spencer Kimball on CockroachDB's agentic future, the old 'we're not Google' objection, and distributed SQL as a mainstream need.
- [Major Tim Peake at RoachFest London: What Space Missions Teach Engineers](https://lucaberton.com/tim-peake-astronaut-lessons-roachfest-london-2026/): Major Tim Peake's RoachFest London keynote on launch day nerves, spacewalk training, psychological safety, and a decision framework for incident response.
- [AI Agents in Healthcare: A Floor Chat at RoachFest London](https://lucaberton.com/ai-agents-healthcare-roachfest-london-2026/): A spontaneous RoachFest London 2026 conversation on where AI agents could take healthcare: pattern recognition, microbots, and cancer detection.

## hardware

- [Jessie Lacome on Lenovo's ThinkSystem SR675 V3 AI Server](https://lucaberton.com/jessie-lacome-lenovo-thinksystem-summit-2026/): Jessie Lacome walks through Lenovo's ThinkSystem SR675 V3, an AMD EPYC server for OpenShift Virtualization and AI headroom.
- [Dell PowerEdge P2PBridge.SL.x: Diagnosing PCIe Enumeration Hangs](https://lucaberton.com/dell-poweredge-p2pbridge-sl-pcie-enumeration-hang/): A host hanging at P2PBridge.SL.x during POST usually means a PCIe enumeration or link-training failure, not a fault in the bridge itself.
- [Chip Design Software: Free Open-Source EDA](https://lucaberton.com/introduction-to-chip-design-with-open-source-eda-tools/): Learn chip design with free open-source EDA tools. Simulate, layout, and validate chips using Magic, OpenROAD, KLayout, and Yosys. Free Coursera course.

## DevOps

- [macOS ENFILE Error: Too Many Open Files — Fix Guide](https://lucaberton.com/macos-enfile-error-too-many-open-files-fix/): Fix the macOS ENFILE file table overflow error when builds fail. Diagnose kern.maxfiles exhaustion with sysctl and lsof, and find the leaking process.
- [Restore a Deleted Google Analytics 4 Property](https://lucaberton.com/restore-deleted-google-analytics-ga4-property/): Deleted a GA4 property by mistake? Restore it from Google Analytics Trash within 35 days and keep the original measurement ID.
- [Fix OpenClaw ERR_STRING_TOO_LONG Session Error](https://lucaberton.com/openclaw-err-string-too-long-bloated-session-fix/): OpenClaw agent fails with 'Cannot create a string longer than 0x1fffffe8 characters'? It's a bloated session JSONL hitting Node's string limit. Here's the fix.
- [Turn Google Search Console Data Into a Growth Plan](https://lucaberton.com/google-search-console-growth-analysis-python-script-2026/): Run one dependency-free Python script on your Search Console export to surface the SEO levers that move traffic: CTR bands and striking-distance pages.
- [etcd Backup & Maintenance for Production Kubernetes](https://lucaberton.com/kubernetes-etcd-backup-maintenance-production-2026/): Protect your cluster state — automated etcd snapshots, defragmentation schedules, performance tuning, and disaster recovery procedures.
- [Argo CD: GitOps Continuous Deployment for Kubernetes](https://lucaberton.com/argo-cd-gitops-kubernetes-deployment-2026/): Production-ready Argo CD setup — ApplicationSets, multi-cluster sync, RBAC, notifications, and progressive delivery patterns.

## RISC-V

- [DeepComputing DC-ROMA: Running DeepSeek on a RISC-V Laptop, No GPU](https://lucaberton.com/deepcomputing-dc-roma-riscv-deepseek-local-ai-2026/): At RISC-V Summit Europe 2026 in Bologna, DeepComputing ran a 7B DeepSeek model on their DC-ROMA Framework 13 mainboard — 100% on CPU, no cloud, no GPU.
- [RISC-V Assembly Tutorial: Your First Program](https://lucaberton.com/risc-v-assembly-language-tutorial/): Learn RISC-V assembly from scratch — registers, instruction formats, and a working hello-world you can run in QEMU. A beginner-friendly intro to the open ISA.
- [RISC-V Cryptography: Scalar and Vector Crypto](https://lucaberton.com/risc-v-cryptography-extensions-scalar-vector/): How RISC-V accelerates cryptography — the scalar crypto (Zk) and vector crypto extensions, AES/SHA instructions, constant-time guarantees, and entropy sources.
- [RISC-V Security: Privilege, PMP & Confidential Computing](https://lucaberton.com/risc-v-security-privilege-pmp-confidential-computing/): How RISC-V handles security — machine/supervisor/user privilege, Physical Memory Protection, the hypervisor extension, and confidential computing.
- [Build a RISC-V Softcore on an FPGA](https://lucaberton.com/risc-v-fpga-softcore-tutorial/): Turn a cheap FPGA into a real RISC-V CPU. A practical intro to softcores like PicoRV32 and VexRiscv — what they are, how the flow works, and how to start.
- [RISC-V Development Boards: 2026 Buyer's Guide](https://lucaberton.com/risc-v-development-boards-2026-guide/): From the VisionFive 2 to the Milk-V Jupiter and Banana Pi BPI-F3 — the best RISC-V development boards to buy in 2026, matched to budget and use case.

## ai

- [Packaging a Hugging Face Model Snapshot](https://lucaberton.com/packaging_huggingface_model_snapshot/): Learn how to download a Hugging Face model as real files and package it into a portable tarball for offline use, archival, or distribution.

## Books and Community

- [FikaWorks Day 2025: Writing Technical Books That Sell...](https://lucaberton.com/fikaworks-day-2025/): Luca Berton shares his journey from engineer to six-time published author at FikaWorks Day 2025—learn how to write, pitch, and scale your impact through.
- [London Book Fair 2024: From DevOps to Digital Storytelling](https://lucaberton.com/london-book-fair-2024/): Join Luca Berton at #LBF24 as he explores tech publishing, AR books, and international innovation at one of the world's leading book fairs.
- [New Delhi Book Fair 2024 Highlights](https://lucaberton.com/new-delhi-international-book-fair-2024/): Explore the buzz from the New Delhi Book Fair 2024 as Luca Berton meets authors, publishers, and tech professionals — with a DevOps twist.

## Optional

Most recent articles (full index at https://lucaberton.com/blog/):

- [Post KubeCon Japan 2026: Reflections on AI, Cloud Native, and the Developer Journey](https://lucaberton.com/post-kubecon-japan-2026-reflections/): Straight from KubeCon Japan 2026: the shift from CPUs to GPUs, exploding AI models, and the new accelerator landscape — what it means for developers.
- [My August 2026: KubeCon Japan, Red Hat Summit & More](https://lucaberton.com/luca-newsletter-august-2026/): August 2026 roundup: KubeCon Japan, Red Hat Summit, Hiroshima, Amsterdam meetups, and the AI platform trends I'm watching.
- [Fish Audio Raises $52M Seed for AI Voice Models](https://lucaberton.com/fish-audio-52m-seed-ai-voice-models/): Fish Audio's $52M seed funds AI voice models for creators and enterprises — 8M+ users, $21M ARR, and an open-source Fish Speech lineage.
- [EDGECASE 2026: Gotta Catch 'Em All](https://lucaberton.com/edgecase-2026/): EDGECASE 2026 returns September 24 in Hilversum — the Netherlands' biggest Kubernetes event, 400+ cloud native engineers, hosted by Bart Farrell.
- [Astro Responsive Images: 50% Smaller with AVIF, WebP](https://lucaberton.com/astro-responsive-images-optimization/): Use Astro's Image component with formats [avif, webp] and a widths array for an automatic responsive srcset — about half the JPEG size.
- [BrowserStack Meetup Amsterdam: AI's Impact on QA](https://lucaberton.com/browserstack-meetup-amsterdam-ai-qa-2026/): Luca attended the BrowserStack Meetup Amsterdam 2026, exploring how AI is reshaping QA — from Claude + Playwright to GenAI in quality engineering.
- [AI Builders Amsterdam: A Hands-On AI Meetup Recap](https://lucaberton.com/ai-builders-amsterdam-2026/): AI Builders Amsterdam recap: talks on validating side projects, docs for agents, and running local LLM agents at StartDock.
- [Four UI & AI Tools I'm Watching in 2026](https://lucaberton.com/four-ui-ai-tools-2026/): Watermelon UI, Motion Primitives, Manus, and Haikei: four tools I'm watching for building interfaces, motion, autonomy, and design assets.
- [Add a Live QR Code in Word](https://lucaberton.com/word-qr-code-displaybarcode-tutorial/): Use the native DISPLAYBARCODE field in Word to embed a real, updatable QR code that links to lucaberton.com — no add-in required.
- [YAGNI: You Aren't Gonna Need It](https://lucaberton.com/yagni-principle/): YAGNI — You Aren't Gonna Need It — an Extreme Programming rule against building features no current requirement demands. Less code, fewer bugs, faster delivery.
- [Azure Linux 4.0: Microsoft's First-Party Linux in Preview](https://lucaberton.com/azure-linux-4-public-preview/): Azure Linux 4.0 is now in public preview: Microsoft's Fedora-derived, dnf5-powered, secure-by-default Linux built for Azure VMs, AKS, and containers.
- [Claude Code Usage Limits: Session vs Weekly](https://lucaberton.com/claude-code-usage-limits-explained/): Claude Code's Usage page shows two limits: a rolling session and a weekly cap across all models, plus temporary boosts that revert to your plan baseline.
- [Soup: Fine-Tune an 8B Model on a 4 GB GPU](https://lucaberton.com/soup-cli-fine-tuning-llm-layer-streaming/): Soup is an open-source CLI that fine-tunes LLMs from one YAML and streams the base model layer-by-layer so an 8B model trains on a 4 GB GPU.
- [Netris Raises $15M to Automate AI Neocloud Networks](https://lucaberton.com/netris-ai-neocloud-network-automation/): Netris raised a $15M a16z Series A to automate AI neocloud networks: hardware-accelerated, deterministic, no AI in the product.
- [Model Soups: Better Fine-Tuning by Averaging Weights](https://lucaberton.com/model-soup-fine-tuning/): Model Soups average N fine-tuned checkpoints into one. No extra inference cost or latency, often beats the single best run. Recipe + PyTorch code inside.
- [Meta Muse Glimmer: A 30B Agentic Model for Your GPU](https://lucaberton.com/meta-muse-glimmer-30b-local-agentic-model/): Meta open-sourced Muse Glimmer, a 30B dense agentic model that runs under 20GB on a consumer GPU via 4-bit quantization and DFlash speculative decoding.
- [Photonic Networks: the Standout Tech at KubeCon Japan 2026](https://lucaberton.com/kubecon-japan-2026-photonic-networks/): From KubeCon Japan 2026: photonic networks could reshape the AI data center by moving data optically — Fujitsu leads this work in Japan.
- [Hiroshima Peace Memorial Ceremony 2026: Attending in Person](https://lucaberton.com/hiroshima-peace-memorial-ceremony-2026/): I attended the Hiroshima Peace Memorial Ceremony on 6 August 2026. Here is what the minute of silence, 120 nations, and two hibakusha testimonies taught me.
- [Kata Containers 4.0 Reinforces AI Agent Sandboxing with a New Rust Runtime](https://lucaberton.com/kata-containers-4-0-ai-agent-sandboxing-rust-runtime/): Kata Containers 4.0 ships runtime-rs, a Rust runtime replacing Go, hardening AI agent sandboxing with VM-grade isolation and lower startup latency.
- [Community & Takeaways from KubeCon + CloudNativeCon Japan 2026](https://lucaberton.com/kubecon-japan-2026-community-takeaways/): From KubeCon Japan 2026: the growing CNCF platform-engineering community in Japan, hallway-track moments, and how AI helped me discover volcanic eggs.
- [Cloud Native in the Real World: Optics, Subaru & Uber (KubeCon Japan 2026)](https://lucaberton.com/kubecon-japan-2026-cloud-native-real-world/): From KubeCon Japan 2026: an optics maker and Subaru running AI on cloud-native infrastructure, plus a walk through the Uber booth on the show floor.
- [Open Source Building Blocks for AI Platforms (KubeCon Japan 2026)](https://lucaberton.com/kubecon-japan-2026-open-source-ai-building-blocks/): From KubeCon Japan 2026: the AI conformance framework for portability, a CNCF+in-house research platform, and dynamic resource composition in CNCF sandbox.
- [Security, Isolation & Sovereign AI on Kubernetes (KubeCon Japan 2026)](https://lucaberton.com/kubecon-japan-2026-security-isolation-sovereign-ai/): From KubeCon Japan 2026: holding AI isolation without killing efficiency, the hybrid namespace/dedicated-tools pattern, and open-source sovereign orchestration.
- [LocalAI LongCat-Video-Avatar 1.5: Local Talking Avatars](https://lucaberton.com/localai-longcat-video-avatar-1-5/): Run an open-source, audio-driven talking avatar on your own GPU with LocalAI 4.7.0 and LongCat-Video-Avatar 1.5 — MIT, Whisper-Large lip sync, no API.
- [Tuning Kubernetes for AI: the Real Trade-offs from KubeCon Japan 2026](https://lucaberton.com/kubecon-japan-2026-tuning-kubernetes-for-ai/): From KubeCon Japan 2026: why AI on Kubernetes needs tuning, not just more GPUs — LLMD caching/routing and the real GPU, memory and electricity cost.