DevOps'ish 306: Apple Bets on Ternus, Russia Bets on Chaos, and More

Clone Any Voice, in 9 Languages, at 3x Industry Speed
Voxtral TTS is Mistral’s text-to-speech model, 3x faster than industry standard, with voice cloning that goes beyond reading text aloud. It models personality: natural pauses, rhythm, intonation, and emotional range. Try Voxtral TTS now via API, Mistral Studio, or on Hugging Face under Apache 2.0. SPONSORED

How is infrastructure keeping pace with AI in 2026?
Managing IaC or leading platform engineering? IaCConf is the “can’t miss” event featuring 20 top IaC leaders across 13 sessions. Join 5,000+ practitioners to share what’s actually working and swap hard-won lessons. Register now! SPONSORED

DevOpsDays Raleigh
DevOpsDays Raleigh is one of the region’s premier community-driven tech conferences, bringing together engineers, DevOps leaders, and IT professionals for two days of real-world insights, practitioner-led talks, and interactive open spaces. And this year is special as it’s the 10th anniversary! This conference is focused on DevOps, cloud, automation, and modern software delivery. It offers a unique mix of learning, networking, and collaboration with peers tackling similar challenges across industries. Join me in Raleigh April 30th & May 1st!

Cloud Native Days Romania
Two days of cloud native talks, hands-on workshops, and strong community momentum - 18–19 May at the Radisson Blu, Bucharest. Join developers, platform engineers, DevOps practitioners, engineering leaders, and cloud enthusiasts for the 3rd edition of Romania’s community-driven Cloud Native Days, bringing practical Kubernetes use cases and modern cloud native systems to the stage.

The great migration: Why every AI platform is converging on Kubernetes - 66% of organizations running generative AI inference workloads are doing it on Kubernetes, and this post breaks down why the platform became the convergence point for data, training, serving, and agent infrastructure.

Gateway API v1.5 - Six Experimental features graduate to the Standard channel — ListenerSet, TLSRoute, HTTPRoute CORS Filter, Client Certificate Validation, Certificate Selection, and ReferenceGrant — plus a new release train model for more predictable cadence.

Kubernetes v1.36: User Namespaces are finally GA - Rootless container security reaches GA: workloads can now run with non-root identities and namespaced capabilities without requiring recursive file ownership changes, thanks to Linux ID-mapped mounts.

SELinux Volume Label Changes goes GA (and likely implications in v1.37) - SELinuxMount graduates to Beta in v1.36 and will be on by default in v1.37, speeding up volume setup — but shared volumes between Pods with different SELinux contexts may break, so read this before you upgrade.

Tim Cook Steps Down as Apple CEO - Cook exits September 1st after turning Apple from a $300B company into a $4T one; John Ternus, SVP of Hardware Engineering, takes over as CEO.

Meta targets May 20 for first wave of layoffs, additional cuts later in 2026 - Meta will cut about 8,000 employees — roughly 10% of its workforce — starting May 20 while leaving 6,000 open roles unfilled, all so it can fund a $135B AI infrastructure push without visibly hurting the balance sheet.

AI’s New Training Data: Your Old Work Slacks And Emails - When companies shut down, services like SimpleClosure and Sunset now help them sell their internal operational data to AI labs.

Russia hits European thermal power plant in attempted ‘destructive’ cyberattack - Pro-Kremlin hackers targeted a Swedish thermal power plant in what’s becoming a pattern of increasingly reckless attacks on Western critical infrastructure — this one was stopped by built-in defenses, but the escalation isn’t subtle.

The Mother of All AI Supply Chains: Critical, Systemic Vulnerability at the Core of Anthropic’s MCP - Researchers found a fundamental architectural flaw in the Model Context Protocol enabling arbitrary command execution, affecting millions of downloads and potentially thousands of deployed MCP servers.

NIST to limit work on CVE entries as submissions surge - NIST is narrowing its enrichment work to only the most critical vulnerabilities as the flood of bug submissions has become genuinely unmanageable — quiet institutional decay at exactly the wrong time for the security ecosystem.

Amazon ‘strong-armed’ brands into price fixing, says California AG - Newly unredacted court filings show Amazon pressuring Levi’s, Hanes, and others to raise prices on competing platforms to protect Amazon’s margins; trial is set for January 2027.

Grafana offers AI assistant for free, warns users not to go mad - Grafana is making its AI observability assistant free for open-source and on-prem users while quietly expanding into business analytics beyond the monitoring space it built its name on.

The OpenAI Founders On Their Plan To Battle Elon, Compute And Everything Else - OpenAI’s founders lay out their competitive strategy against Elon Musk’s ventures and talk through the compute constraints and organizational challenges that define the next phase of the company.

How Much Do GPU Clusters Really Cost? - A thorough TCO breakdown going beyond sticker-price-per-GPU to examine storage, networking, reliability, support, and engineering overhead across cloud tiers — the real numbers are not comforting.

Duolingo’s Kubernetes Leap - How Duolingo migrated 500+ backend services from ECS to EKS using GitOps with Argo CD, and the hard lessons learned around IPv6 networking and observability integration at scale.

Components of A Coding Agent - Sebastian Raschka breaks down the six key components — repo context, prompt caching, structured tools, context management, session memory, and subagent delegation — that make coding agents far more capable than a plain chat window.

My AI Workflow (Without Losing My Skills) - A practical approach to staying sharp as a developer while using AI tools: deliberate manual coding, careful code review, and keeping planning in your own head rather than outsourcing it entirely to the model.

Load Balancing and Scaling LLM Serving - Standard load balancing erodes prompt cache benefits when you’re running multiple LLM replicas; this post explains cache-aware routing strategies as the practical fix.

Introducing Framework Laptop 13 Pro - The redesigned Framework 13 Pro gets Intel Core Ultra Series 3, a 74Wh battery, LPCAMM2 memory, touch display, and haptic touchpad — still the rare hardware company that ships what it promises.

Ok, what ACTUALLY uses Rust? - An extensive rundown of major adopters from Linux and Windows to Discord and NASA, with an honest note that Rust still trails C++ and Java in overall deployment despite its growing reputation.

PoweredByWho — Who’s Building Data Centers in Your Backyard? - Tracks U.S. data center construction projects, corporate spending, and political influence to help communities understand where AI infrastructure is landing and who’s bankrolling it.

How the Kubernetes control plane works - A solid walkthrough of the API server, etcd, scheduler, and controller manager, and how they connect to manage cluster state — the mental model, not just a doc dump.

AWS Services Infographics - Visual reference guides covering compute, storage, networking, and messaging services for anyone who processes information better with a diagram than a wall of documentation.

sivchari/kumo - MIT - A lightweight AWS service emulator written in Go.

zilliztech/claude-context - MIT - Code search MCP for Claude Code; makes the entire codebase available as context for any coding agent.

drasimwagan/mdv - MIT - A Markdown superset for documents, dashboards, and slides with embedded data and visualizations, shipping with HTML and PDF export, live preview, and a VS Code extension.