/net/dev

A View from Schuman: Celebrating Google's Signature on the EU AI Act Code of Practice

Sun, 26 Jul 2026 00:00:00 -0400

A 20-year Brussels tech policy veteran's perspective on how Google is empowering users, advancing transparency, and pioneering responsible AI. - If you have spent two decades walking and cycling past the iconic curved facade of the Berlaymont, under the blue European Union flags flying over Rond-Point Schuman, you develop a deep appreciation for moments when technology and public policy move in harmony. This week, Google announced a major milestone in responsible AI governance: signing the official EU AI...

Sub-150ms Teleport & Closed-Loop Reliability: SRE Practices and Performance Engineering Behind Aether

Sat, 25 Jul 2026 00:00:00 -0400

How the Aether team achieved <30s pure TypeScript container builds, <150ms instant session state-teleport recovery, and closed-loop Gemini ADK SRE observability. - When building platforms capable of running autonomous AI engineering agents, classic Site Reliability Engineering (SRE) paradigms face a breaking point. Unlike static web services where request-response lifecycles terminate in tens of milliseconds, autonomous agent trajectories run multi-turn feedback loops, edit codebases, run shell commands, and interact with live cloud infrastructure over prolonged execution sessions. If an underlying worker...

Quantum-Proofing Cloud Identity: Post-Quantum Cryptography (PQC) & Zero-Trust Architecture in Aether

Fri, 17 Jul 2026 00:00:00 -0400

Lattice cryptography, NIST ML-KEM/ML-DSA primitives, gVisor runsc isolation, and Cosign SLSA-3 supply chain attestations. - As enterprise autonomous AI engineering agents and microservices scale to manage critical infrastructure across cloud and edge topologies, the foundational assumptions underlying classical cybersecurity are shifting. Among the most looming challenges facing modern identity platforms is the advent of cryptographically relevant quantum computers (CRQCs). Classical asymmetric encryption algorithms—such as RSA-2048 and Elliptic Curve Cryptography (ECDSA/ECDH)—form the backbone of...

Unshackling Cloud Identity: The Story of Project Aether Genesis

Wed, 08 Jul 2026 00:00:00 -0400

How Google Cloud decoupled Service Accounts and machine identity from legacy global consumer databases into Cloud Gaia. - Behind every enterprise Google Cloud Platform (GCP) workload lies an unseen, hyper-critical foundational layer: Identity. Whether it is a serverless function fetching secrets, an automated machine learning job querying BigQuery, or a Kubernetes pod interacting with Cloud Storage, Service Accounts and machine identities form the nervous system of modern cloud computing. For over a decade, however, these cloud-native...

The Agentic Engineering Paradigm: Software Development at the Speed of Context

Thu, 18 Jun 2026 00:00:00 -0400

From pair programming to multi-agent swarm orchestration in cloud-native enterprise architectures. - Mid-2026 marks a historic milestone in software engineering: the transition from human-centric manual coding to Spec-Driven Agentic Swarm Orchestration. Engineers no longer spend days writing boilerplate CRUD endpoints or manual integration tests; instead, they act as high-level system architects guiding specialized subagent teams. The New Engineering Stack: Intent Specification -> Agent Swarm Plan & Verification -> Multi-Agent Execution...

Watts, FLOPS, and Zero Trust: The Green & Post-Quantum Cloud of 2026

Fri, 10 Apr 2026 00:00:00 -0400

How European cloud centers achieved PUE < 1.1 while deploying NIST post-quantum encryption. - As we cross into 2026, the European cloud landscape stands at the intersection of two massive imperatives: sustainability mandates under the EU Green Deal and Post-Quantum Cryptography (PQC) deployment across critical cloud infrastructure. Dual Imperatives: European datacenters in Benelux must now report real-time Power Usage Effectiveness (PUE) and water usage efficiency while upgrading public key infrastructure (RSA/ECC) to...

From Monitoring to Autonomous Remediation: The SRE Agent Revolution

Tue, 24 Mar 2026 00:00:00 -0400

How multi-agent triage loops reduced incident MTTR from hours to seconds in multi-cloud mesh. - Spring 2026 transformed Site Reliability Engineering (SRE). Instead of waking up human engineers for midnight PagerDuty incidents, cloud platforms now utilize autonomous SRE agent swarms that observe live eBPF kernel metrics, diagnose root causes, and execute self-healing remediation plans in real time. MTTR Breakthrough: Automated anomaly diagnosis and traffic rerouting slashed Mean Time To Recovery (MTTR) across European...

SLSA Level 4 in the Age of Autonomous Coders: Securing the Agentic SDLC

Mon, 19 Jan 2026 00:00:00 -0500

Cryptographic provenance, reproducible builds, and secret-less IAM when agents author code. - As autonomous AI agents began authoring a significant fraction of enterprise pull requests in early 2026, securing the Software Development Life Cycle (SDLC) required updating Supply-chain Levels for Software Artifacts (SLSA) standards for non-human contributors. Core Rule: Code authored or modified by an AI agent must pass through cryptographic provenance attestation (Sigstore / Cosign), ephemeral sandboxed build workers,...

The Phantom in the Pipeline: Defending European Cloud Infrastructure Against Synthetic AI Exploits

Sat, 06 Dec 2025 00:00:00 -0500

Mitigating deepfake social engineering, LLM prompt injection, and automated supply chain poisoning. - By late 2025, ENISA’s annual Threat Landscape report underscored a troubling evolution: cybercriminals and state-sponsored actors began deploying autonomous AI agents to discover zero-day vulnerabilities in cloud software supply chains and launch hyper-personalized deepfake social engineering attacks. New Vector - Indirect Prompt Injection: Attackers hide malicious instructions in public code repositories, web pages, or PDF documents parsed by...

The Rise of Agentic DevSecOps: Reimagining Internal Developer Platforms (IDPs)

Fri, 14 Nov 2025 00:00:00 -0500

How self-healing platform orchestrators and LLM sidecars transformed developer experience. - By late 2025, Internal Developer Platforms (IDPs) evolved beyond static self-service web portals into Agentic Platform Mesh Orchestrators. Rather than forcing developers to manually construct YAML manifests or navigate complex cloud portals, autonomous AI agents now act as interactive copilots for platform engineering teams across Europe. Key Evolution: Developers describe desired service capabilities in high-level domain specs, while...

From Leuven to the Cloud: Microchips, Hardware Roots of Trust, and AI Security

Thu, 21 Aug 2025 00:00:00 -0400

Why cloud security now starts at the sub-2nanometer wafer level in Belgium's silicon hub. - Nestled in Leuven, Belgium, IMEC (Interuniversity Microelectronics Centre) stands as the world’s premier research hub for advanced semiconductor chip design. As AI models scale into trillions of parameters, cloud security has migrated down from software containers to hardware roots of trust and sub-2nm silicon enclaves. Confidential Computing Era: Securing data at rest and in transit is no longer...

Supercomputers on the Canal: EuroHPC and the Rise of European Sovereign AI

Wed, 09 Apr 2025 00:00:00 -0400

Inside the exascale infrastructure powering open European models and green datacenters. - Spring 2025 marked the full operationalization of the EuroHPC Joint Undertaking’s exascale supercomputing network—including JUPITER, Lumi, MeluXina (Luxembourg), and associated AI Factories across Belgium and the Benelux region. Sovereign Compute Power: European researchers and enterprises now have access to GPU clusters operating on 100% renewable energy, training open European multilingual foundation models without depending exclusively on overseas hyperscaler...

The DORA & NIS2 Crucible: Transforming EU Compliance into CI/CD Code

Wed, 04 Dec 2024 00:00:00 -0500

How European fintechs and enterprises turned regulations into continuous automated guardrails. - With DORA (Digital Operational Resilience Act) coming into force in January 2025 across European banking, insurance, and investment firms, alongside national NIS2 deadlines, compliance teams and engineering organizations faced a stark choice: drown in manual spreadsheets or automate compliance as code. DORA Pillars: ICT Risk Management, Incident Reporting, Operational Resilience Testing, Third-Party Risk Management, and Information Sharing. Policy-as-Code...

The Day Cloud Monoculture Broke: Post-Mortem of a European Outage

Sun, 11 Aug 2024 00:00:00 -0400

Why resilience requires cloud diversity, canary deployments, and eBPF kernel isolation. - Mid-2024 witnessed one of the largest IT disruptions in history when a single faulty channel file update pushed to a widely used endpoint security driver triggered bootloops across millions of hosts globally—grounding flights at Brussels Airport, disrupting Belgian healthcare systems, and bringing financial operations to a standstill. The Warning: Single-vendor monoculture and un-staged kernel-level updates represent catastrophic single...

Le Parcours BD de Bruxelles: Une Promenade Artistique au Cœur de la Capitale

Thu, 09 May 2024 00:00:00 -0400

De Tintin à Gaston Lagaffe: exploration du patrimoine mural bruxellois guidée par une carte interactive. - Bruxelles est incontestablement la capitale mondiale de la bande dessinée. Depuis 1991, sous l’impulsion de la Ville de Bruxelles et du Centre Belge de la Bande Dessinée (CBBD), les murs aveugles du Pentagone, des Marolles et de Laeken se sont transformés en une immense galerie d’art à ciel ouvert. Recensées dans l’inventaire officiel du patrimoine mobilier de la...

Shielding the Union: Lessons from Belgium’s EU Presidency Cyber Drills

Mon, 22 Apr 2024 00:00:00 -0400

How cross-border EU threat sharing and cloud SOC automation defended critical infrastructure. - During the first half of 2024, Belgium held the Presidency of the Council of the European Union. Central to the Belgian presidency agenda was strengthening European cyber resilience ahead of national and European Parliament elections. Highlight: Pan-European cyber crisis exercises, coordinated by ENISA and Belgium’s Cyber Command, tested multi-cloud incident orchestration and cross-border threat intelligence sharing under extreme...

Midnight in Brussels: Architecting for the World’s First EU AI Act

Fri, 08 Dec 2023 00:00:00 -0500

Breaking down high-risk AI classification, transparency logs, and cloud ML governance. - On December 8, 2023, after an intense 36-hour negotiation session in Brussels, European Union policymakers reached a landmark agreement on the EU AI Act—the world’s first comprehensive horizontal legal framework for Artificial Intelligence. What It Means for Developers: AI systems are classified into risk tiers (Unacceptable, High, Limited, Minimal). High-risk systems (e.g. in hiring, credit scoring, critical infrastructure,...

GDPR Meets GenAI: Navigating Data Sovereignty in European LLMs

Fri, 18 Aug 2023 00:00:00 -0400

Inside EDPB taskforce guidance and privacy-preserving RAG architectures in the EU. - Summer 2023 marked a watershed moment for Generative AI adoption in Europe. As companies rushed to integrate Large Language Models (LLMs) into customer portals and internal enterprise workflows, European data protection authorities (led by the EDPB ChatGPT Taskforce and national DPAs like Belgium’s APD/GBA) issued stern reminders: GDPR compliance is non-negotiable. The Challenge: Standard public LLM APIs often...

The Atomium in the Cloud: Building Cyber Resilience under Europe's NIS2

Thu, 06 Apr 2023 00:00:00 -0400

How Belgian infrastructure and European cloud providers adapted to tougher EU cyber rules. - As the European Union finalised the NIS2 Directive (Network and Information Security Directive), organisations across Belgium—from Antwerp’s bustling port logistics to energy grids and healthcare networks—faced a pivotal shift in cybersecurity obligations. Key Takeaway: NIS2 drastically expands the scope of covered entities and enforces strict incident notification timelines (24-hour early warning) along with direct executive management liability. The...

mostly harmless

Mon, 27 Feb 2023 00:00:00 -0500

The steps above are the bare minimum configuration - Open carefully. I’m mostly harmless Both menus are completely optional and can be commented out if not required. Use the template provided in the file as a guide. If you now run hugo server you will be presented with a blank Congo. Last advice The pageRef parameter allows you to easily reference Hugo It is the quickest way...