https://jacek-mar.github.io/ai-dev-blog/ Curated AI and developer news from leading sources en-us Fri, 06 Mar 2026 11:17:02 +0000 https://jacek-mar.github.io/ai-dev-blog/posts/6-2026policypartnering-with-mozilla-to-improve-firefoxs-security.html https://jacek-mar.github.io/ai-dev-blog/posts/6-2026policypartnering-with-mozilla-to-improve-firefoxs-security.html AI models can now independently identify high-severity vulnerabilities in complex software. As we recently documented, Claude found more than 500 zero-day vulnerabilities (security flaws that are unknown to the software’s maintainers) in well-tested open-source software. Fri, 06 Mar 2026 00:00:00 +0000 Unknown Claude https://jacek-mar.github.io/ai-dev-blog/posts/martians-independent-benchmark-tested-13-code-review-tools.html https://jacek-mar.github.io/ai-dev-blog/posts/martians-independent-benchmark-tested-13-code-review-tools.html Last week, Martian released Code Review Bench — the first independent, open-source benchmark for AI code review tools. It tracks over 200,000 real pull requests across GitHub, measures which review comments developers actually act on, and updates daily. The methodology and code are fully open source Fri, 06 Mar 2026 11:17:02 +0000 Brian Turcotte KiloCode https://jacek-mar.github.io/ai-dev-blog/posts/gpt-54-is-now-available-in-windsurfgpt-54-is-now-available-in-windsurf-with-multiple-reasoning-effor.html https://jacek-mar.github.io/ai-dev-blog/posts/gpt-54-is-now-available-in-windsurfgpt-54-is-now-available-in-windsurf-with-multiple-reasoning-effor.html No content available. Fri, 06 Mar 2026 11:17:02 +0000 Unknown Windsurf https://jacek-mar.github.io/ai-dev-blog/posts/look-what-you-made-us-patch-2025-zero-days-in-review.html https://jacek-mar.github.io/ai-dev-blog/posts/look-what-you-made-us-patch-2025-zero-days-in-review.html Visibility and context on the threats that matter most. Written by: Casey Charrier, James Sadowski, Zander Work, Clement Lecigne, Benoît Sevens, Fred Plan Google Threat Intelligence Group (GTIG) tracked 90 zero-day vulnerabilities exploited in-the-wild in 2025. Fri, 06 Mar 2026 11:17:02 +0000 Unknown Google Cloud https://jacek-mar.github.io/ai-dev-blog/posts/computeh4d-vms-now-ga-deliver-exceptional-performance-and-scaling-for-hpc-workloads.html https://jacek-mar.github.io/ai-dev-blog/posts/computeh4d-vms-now-ga-deliver-exceptional-performance-and-scaling-for-hpc-workloads.html Product Manager Senior HPC Technologist State-of-the-art image generation and editing Today, we’re announcing the general availability of H4D VMs, our latest high performance computing (HPC)-optimized VM, powered by the 5th Generation AMD EPYC™ processors. Fri, 06 Mar 2026 11:17:02 +0000 Unknown Google Cloud https://jacek-mar.github.io/ai-dev-blog/posts/small-models-high-quality-inside-bmw-groups-experiments-evaluating-domain-specific-language-models.html https://jacek-mar.github.io/ai-dev-blog/posts/small-models-high-quality-inside-bmw-groups-experiments-evaluating-domain-specific-language-models.html Google Cloud BMW Group State-of-the-art image generation and editing A car you can talk to has been a longstanding dream, whether as the basis for television shows or more recent smartphone integrations. One way of achieving better, more natural voice commands is by incorporating AI foundation model Fri, 06 Mar 2026 11:17:02 +0000 Unknown Google Cloud https://jacek-mar.github.io/ai-dev-blog/posts/grow-your-own-way-introducing-native-support-for-custom-metrics-in-gke.html https://jacek-mar.github.io/ai-dev-blog/posts/grow-your-own-way-introducing-native-support-for-custom-metrics-in-gke.html Senior Product Manager, GKE Software Engineer State-of-the-art image generation and editing When platform engineers, AI Infrastructure leads and developers think about autoscaling workloads running on Kubernetes, their goal is straightforward: get the capacity they need, when they need it, at the be Fri, 06 Mar 2026 11:17:02 +0000 Unknown Google Cloud https://jacek-mar.github.io/ai-dev-blog/posts/the-ultimate-nano-banana-prompting-guide.html https://jacek-mar.github.io/ai-dev-blog/posts/the-ultimate-nano-banana-prompting-guide.html Product Marketing Manager, Gen Media Technical Solutions Manager, Google Cloud State-of-the-art image generation and editing Creating precise, high-quality images often involves endless trial and error. You need a model that actually understands what you’re asking for. Fri, 06 Mar 2026 11:17:02 +0000 Unknown Google Cloud https://jacek-mar.github.io/ai-dev-blog/posts/bringing-robotics-ai-to-embedded-platforms-dataset-recording-vla-finetuning-and-ondevice-optimizatio.html https://jacek-mar.github.io/ai-dev-blog/posts/bringing-robotics-ai-to-embedded-platforms-dataset-recording-vla-finetuning-and-ondevice-optimizatio.html Recent advances in Large Language Models have enabled the transition from text-only reasoning to multimodal systems. First, with the integration of visual perception in Vision–Language Models (VLMs), and more recently with the generation of robot actions in Vision–Language–Action (VLA) models. Fri, 06 Mar 2026 11:17:02 +0000 Unknown HuggingFace https://jacek-mar.github.io/ai-dev-blog/posts/will-it-roast-we-tested-kilo-code-reviewers-roast-mode-on-5-levels-of-terrible-code.html https://jacek-mar.github.io/ai-dev-blog/posts/will-it-roast-we-tested-kilo-code-reviewers-roast-mode-on-5-levels-of-terrible-code.html Kilo Code’s Code Reviews now has a Roast Mode. Instead of polite suggestions, it reviews your PRs with brutal honesty. We’ve previously tested Code Reviews for accuracy with both free and frontier models. This time, we wanted to see how far the roasting goes. We built a clean bookstore API, created Fri, 06 Mar 2026 11:17:02 +0000 Darko KiloCode https://jacek-mar.github.io/ai-dev-blog/posts/50k-ai-coding-model-benchmark-what-actually-matters.html https://jacek-mar.github.io/ai-dev-blog/posts/50k-ai-coding-model-benchmark-what-actually-matters.html Andrew Filev Published: March 05, 2026 · Last updated: March 05, 2026 The market has been flooded with new models lately. In February alone, the Big Three released updates to their models, along with several major OSS labs. Each of them claims success, and it's quite difficult to tell who's actually Fri, 06 Mar 2026 11:17:02 +0000 Andrew Filev Zencoder https://jacek-mar.github.io/ai-dev-blog/posts/introducing-modular-diffusers---composable-building-blocks-for-diffusion-pipelines.html https://jacek-mar.github.io/ai-dev-blog/posts/introducing-modular-diffusers---composable-building-blocks-for-diffusion-pipelines.html Modular Diffusers introduces a new way to build diffusion pipelines by composing reusable blocks. Instead of writing entire pipelines from scratch, you can mix and match blocks to create workflows tailored to your needs! This complements the existing DiffusionPipeline class with a more flexible, com Fri, 06 Mar 2026 11:17:02 +0000 Unknown HuggingFace https://jacek-mar.github.io/ai-dev-blog/posts/things-stand-with-the-department-of-warannouncementsa-statement-from-dario-amodei.html https://jacek-mar.github.io/ai-dev-blog/posts/things-stand-with-the-department-of-warannouncementsa-statement-from-dario-amodei.html A statement from Dario Amodei Yesterday (March 4) Anthropic received a letter from the Department of War confirming that we have been designated as a supply chain risk to America’s national security. As we wrote on Friday, we do not believe this action is legally sound, and we see no choice but to c Thu, 05 Mar 2026 00:00:00 +0000 Unknown Claude https://jacek-mar.github.io/ai-dev-blog/posts/gas-town-by-kilo.html https://jacek-mar.github.io/ai-dev-blog/posts/gas-town-by-kilo.html Earlier this year, Steve Yegge published a 25-page blog post about something he built called Gas Town. If you read it, you probably had one of two reactions: either “this is insane” or “I need to try this immediately.” Both are correct. Gas Town is an agent orchestrator. Not in the hand-wavy “we orc Fri, 06 Mar 2026 11:17:02 +0000 Brian Turcotte KiloCode https://jacek-mar.github.io/ai-dev-blog/posts/kilo-code-weekly-product-roundup-march-4-2026.html https://jacek-mar.github.io/ai-dev-blog/posts/kilo-code-weekly-product-roundup-march-4-2026.html Welcome back to the weekly product roundup! This week brings one-click review suggestions after task completion, Claude Sonnet 4.6 across three providers, two new API providers, GLM 5 support, and an enormous wave of community contributions. After completing a task in Code or Orchestrator mode, Kilo Fri, 06 Mar 2026 11:17:02 +0000 Brian Turcotte KiloCode https://jacek-mar.github.io/ai-dev-blog/posts/kiloclaw-pricing-no-surprises-no-fine-print.html https://jacek-mar.github.io/ai-dev-blog/posts/kiloclaw-pricing-no-surprises-no-fine-print.html We told you at launch that KiloClaw would come with a 7-day free trial for compute. You’ve been running a fully hosted OpenClaw instance — the same agent with 210k+ stars on GitHub — completely free this entire time. No compute charges or hidden meter running in the background. Just AI inference cos Fri, 06 Mar 2026 11:17:02 +0000 Brendan O'Leary KiloCode https://jacek-mar.github.io/ai-dev-blog/posts/cloud-ciso-perspectives-how-google-approaches-critical-security-topics-from-fundamentals-to-ai.html https://jacek-mar.github.io/ai-dev-blog/posts/cloud-ciso-perspectives-how-google-approaches-critical-security-topics-from-fundamentals-to-ai.html VP, Engineering for Privacy, Safety, and Security The latest on security from Google Cloud's Office of the CISO, twice a month. Welcome to the second Cloud CISO Perspectives for February 2026. Today, Royal Hansen, vice-president, Engineering, explains how we tackle today’s thorniest cybersecurity ch Fri, 06 Mar 2026 11:17:02 +0000 Unknown Google Cloud https://jacek-mar.github.io/ai-dev-blog/posts/coruna-the-mysterious-journey-of-a-powerful-ios-exploit-kit.html https://jacek-mar.github.io/ai-dev-blog/posts/coruna-the-mysterious-journey-of-a-powerful-ios-exploit-kit.html Visibility and context on the threats that matter most. Google Threat Intelligence Group (GTIG) has identified a new and powerful exploit kit targeting Apple iPhone models running iOS version 13.0 (released in September 2019) up to version 17.2.1 (released in December 2023). The exploit kit, named “ Fri, 06 Mar 2026 11:17:02 +0000 Unknown Google Cloud https://jacek-mar.github.io/ai-dev-blog/posts/announcing-the-mcp-toolbox-java-sdk.html https://jacek-mar.github.io/ai-dev-blog/posts/announcing-the-mcp-toolbox-java-sdk.html Staff Developer Advocate, Google Software Engineer, Google Engineering teams are moving beyond simple chatbots to build agentic systems that interact directly with mission critical databases. However, building these enterprise agents often means hitting an integration wall of custom glue code, britt Fri, 06 Mar 2026 11:17:02 +0000 Unknown Google Cloud https://jacek-mar.github.io/ai-dev-blog/posts/how-is-ai-changing-the-future-of-software-engineering.html https://jacek-mar.github.io/ai-dev-blog/posts/how-is-ai-changing-the-future-of-software-engineering.html Sergio Published: March 03, 2026 · Last updated: March 03, 2026 Did you know that 62% of developers rely on at least one AI assistant in their workflow? What began as a simple autocomplete feature has evolved into powerful tools that shape how code is written, reviewed, tested, and optimized. Fri, 06 Mar 2026 11:17:02 +0000 Sergio Zencoder https://jacek-mar.github.io/ai-dev-blog/posts/8-best-conductor-alternatives-to-choose-in-2026-comparison.html https://jacek-mar.github.io/ai-dev-blog/posts/8-best-conductor-alternatives-to-choose-in-2026-comparison.html Sergio Published: March 03, 2026 · Last updated: March 03, 2026 Are you looking for a tool to supercharge your code-generation workflow with parallel AI agents? While Conductor has gained buzz for orchestrating multiple Claude Code agents on your Mac, some users have pointed out a couple of downside Fri, 06 Mar 2026 11:17:02 +0000 Sergio Zencoder https://jacek-mar.github.io/ai-dev-blog/posts/9-best-tools-for-automating-ai-workflows-2026-comparison.html https://jacek-mar.github.io/ai-dev-blog/posts/9-best-tools-for-automating-ai-workflows-2026-comparison.html Sergio Published: March 03, 2026 · Last updated: March 03, 2026 Are you looking for a tool that can automate AI workflows from start to finish? As AI-driven processes grow more complex, the right automation tools are essential for efficiently managing models, data, and integrations. Fri, 06 Mar 2026 11:17:02 +0000 Sergio Zencoder https://jacek-mar.github.io/ai-dev-blog/posts/what-is-multi-agent-orchestration-detailed-overview.html https://jacek-mar.github.io/ai-dev-blog/posts/what-is-multi-agent-orchestration-detailed-overview.html Sergio Published: March 03, 2026 · Last updated: March 03, 2026 Did you know that Gartner predicts that by 2028, at least 15% of day-to-day work decisions will be made autonomously by AI agents, compared to almost none today? This shift reflects a move away from single, isolated AI models toward net Fri, 06 Mar 2026 11:17:02 +0000 Sergio Zencoder https://jacek-mar.github.io/ai-dev-blog/posts/prx-part-3-training-a-text-to-image-model-in-24h.html https://jacek-mar.github.io/ai-dev-blog/posts/prx-part-3-training-a-text-to-image-model-in-24h.html Welcome back 👋 In the last two posts (Part 1 and Part 2), we explored a wide range of architectural and training tricks for diffusion models. We tried to evaluate each idea in isolation, measuring throughput, convergence speed, and final image quality, and tried to understand what actually moves the Fri, 06 Mar 2026 11:17:02 +0000 Unknown HuggingFace https://jacek-mar.github.io/ai-dev-blog/posts/hiring-at-kilo-speed-the-process-reflects-the-job.html https://jacek-mar.github.io/ai-dev-blog/posts/hiring-at-kilo-speed-the-process-reflects-the-job.html Our record is 48 hours from first conversation to offer. Yes, this sounds like a flex, but I share it as signal about what we think hiring should feel like. Most hiring processes are slow by design: multiple rounds, take-home projects, panel interviews, week-long delays between steps. More data poin Fri, 06 Mar 2026 11:17:02 +0000 Emilie Schario KiloCode https://jacek-mar.github.io/ai-dev-blog/posts/give-your-agentic-chatbots-a-fast-and-reliable-long-term-memory.html https://jacek-mar.github.io/ai-dev-blog/posts/give-your-agentic-chatbots-a-fast-and-reliable-long-term-memory.html AI Solutions Acceleration Architect Principal Architect State-of-the-art image generation and editing When scaling conversational agents, the data layer design often determines success or failure. To support millions of users, agents need conversational continuity — the ability to maintain responsiv Fri, 06 Mar 2026 11:17:02 +0000 Unknown Google Cloud https://jacek-mar.github.io/ai-dev-blog/posts/designing-private-network-connectivity-for-rag-capable-gen-ai-apps.html https://jacek-mar.github.io/ai-dev-blog/posts/designing-private-network-connectivity-for-rag-capable-gen-ai-apps.html Developer Relations Engineer State-of-the-art image generation and editing The flexibility of Google Cloud allows enterprises to build secure and reliable architecture for their AI workloads. In this blog we will look at a reference architecture for private connectivity for retrieval-augmented gener Fri, 06 Mar 2026 11:17:02 +0000 Unknown Google Cloud https://jacek-mar.github.io/ai-dev-blog/posts/unified-maintenance-a-new-unified-way-to-manage-maintenance-across-google-cloud.html https://jacek-mar.github.io/ai-dev-blog/posts/unified-maintenance-a-new-unified-way-to-manage-maintenance-across-google-cloud.html Product Manager State-of-the-art image generation and editing Managing planned maintenance is critical for ensuring business continuity and application performance. However, as your usage of cloud services grows, staying on top of maintenance schedules can be complex and time-consuming. Fri, 06 Mar 2026 11:17:02 +0000 Unknown Google Cloud https://jacek-mar.github.io/ai-dev-blog/posts/from-framework-to-scale-accelerating-autonomous-networks-at-mwc-26.html https://jacek-mar.github.io/ai-dev-blog/posts/from-framework-to-scale-accelerating-autonomous-networks-at-mwc-26.html VP, PM and GM, Networking, Google Cloud Principal Engineer, Google Cloud Last year, we unveiled our Autonomous Network Operations framework — a blueprint for Communication Service Providers (CSPs) to move beyond siloed automation toward self-healing, "zero-touch" networks. Fri, 06 Mar 2026 11:17:02 +0000 Unknown Google Cloud https://jacek-mar.github.io/ai-dev-blog/posts/from-vibe-checks-to-continuous-evaluation-engineering-reliable-ai-agents.html https://jacek-mar.github.io/ai-dev-blog/posts/from-vibe-checks-to-continuous-evaluation-engineering-reliable-ai-agents.html Developer Relations Engineer I live through the same story with every single AI agent. After weeks of experiments and tests, it works like a charm. Suddenly, someone comes with a question that the agent fails to answer properly. I rush to make a change by tweaking one of the prompts. After a handful Fri, 06 Mar 2026 11:17:02 +0000 Unknown Google Cloud https://jacek-mar.github.io/ai-dev-blog/posts/turn-your-api-sprawl-into-an-agent-ready-catalog.html https://jacek-mar.github.io/ai-dev-blog/posts/turn-your-api-sprawl-into-an-agent-ready-catalog.html Product Manager Software Engineer State-of-the-art image generation and editing In modern cloud architectures, APIs are the fundamental building blocks of applications. However, as organizations scale, these APIs often end up scattered across multiple gateways, teams, and platforms. Fri, 06 Mar 2026 11:17:02 +0000 Unknown Google Cloud https://jacek-mar.github.io/ai-dev-blog/posts/centralized-policy-meets-distributed-logic-getting-to-know-eventarc-advanced.html https://jacek-mar.github.io/ai-dev-blog/posts/centralized-policy-meets-distributed-logic-getting-to-know-eventarc-advanced.html Staff Software Engineer State-of-the-art image generation and editing Enterprise architects often face a fundamental dilemma: choosing between developer agility and organizational control. Development teams need to move fast and deploy independent microservices without waiting for permission. Fri, 06 Mar 2026 11:17:02 +0000 Unknown Google Cloud https://jacek-mar.github.io/ai-dev-blog/posts/exposing-the-undercurrent-disrupting-the-gridtide-global-cyber-espionage-campaign.html https://jacek-mar.github.io/ai-dev-blog/posts/exposing-the-undercurrent-disrupting-the-gridtide-global-cyber-espionage-campaign.html Visibility and context on the threats that matter most. Last week, Google Threat Intelligence Group (GTIG), Mandiant, and partners took action to disrupt a global espionage campaign targeting telecommunications and government organizations in dozens of nations across four continents. Fri, 06 Mar 2026 11:17:02 +0000 Unknown Google Cloud https://jacek-mar.github.io/ai-dev-blog/posts/paypals-historically-large-data-migration-is-the-foundation-for-its-gen-ai-innovation.html https://jacek-mar.github.io/ai-dev-blog/posts/paypals-historically-large-data-migration-is-the-foundation-for-its-gen-ai-innovation.html SVP & Global Head of Data, AI & ML Technology, PayPal Sr Director Data Analytics, PayPal State-of-the-art image generation and editing With the dawn of the gen AI era, businesses are facing unprecedented opportunities for transformative products, demanding a strategic shift in their technology infra Fri, 06 Mar 2026 11:17:02 +0000 Unknown Google Cloud https://jacek-mar.github.io/ai-dev-blog/posts/519-developers-competed-to-build-the-worst-website-ever.html https://jacek-mar.github.io/ai-dev-blog/posts/519-developers-competed-to-build-the-worst-website-ever.html Last week, we ran a simple contest: use Kilo’s App Builder to create the absolute worst website you can imagine. The results were horrifying - we loved every single one. Here’s something that’s been true since the first person opened a text editor: there are a limited number of ways to make somethin Fri, 06 Mar 2026 11:17:02 +0000 Brian Turcotte KiloCode https://jacek-mar.github.io/ai-dev-blog/posts/surgical-precision-with-ast-based-code-editing-in-kiro.html https://jacek-mar.github.io/ai-dev-blog/posts/surgical-precision-with-ast-based-code-editing-in-kiro.html TL;DR: Over the past few weeks, we've been testing a new AST-based code navigation and editing engine that reduces token usage by 20% on our SWE-PolyBench, a benchmark containing feature request examples, the most frequent type of query in Kiro. It also enables precise, resilient, production-grade c Fri, 27 Feb 2026 00:00:00 +0000 Myeongsoo Kim, Shweta Garg, Varun Kumar and Murali Krishna Ramanathan Kiro https://jacek-mar.github.io/ai-dev-blog/posts/27-2026announcementsstatement-on-the-comments-from-secretary-of-war-pete-hegseth.html https://jacek-mar.github.io/ai-dev-blog/posts/27-2026announcementsstatement-on-the-comments-from-secretary-of-war-pete-hegseth.html No content available. Fri, 27 Feb 2026 00:00:00 +0000 Unknown Claude https://jacek-mar.github.io/ai-dev-blog/posts/serving-data-from-iceberg-lakehouses-fast-and-fresh-with-spanner-columnar-engine.html https://jacek-mar.github.io/ai-dev-blog/posts/serving-data-from-iceberg-lakehouses-fast-and-fresh-with-spanner-columnar-engine.html Group Product Manager Director of Engineering Our most intelligent model available yet for complex tasks on Gemini Enterprise and Vertex AI The divide between data in operational databases and analytical data lakehouses is disappearing fast. Fri, 06 Mar 2026 11:17:02 +0000 Unknown Google Cloud https://jacek-mar.github.io/ai-dev-blog/posts/pro-level-image-generation-gets-faster-and-more-accessible-with-nano-banana-2.html https://jacek-mar.github.io/ai-dev-blog/posts/pro-level-image-generation-gets-faster-and-more-accessible-with-nano-banana-2.html VP of Product Management, Vertex AI Today, we’re entering a new and vibrant era of generative creativity with Nano Banana 2. What’s new: Nano Banana 2 is our state-of-the-art image generation and editing model. It delivers Pro-level image generation and editing at the speed you expect from Flash — m Fri, 06 Mar 2026 11:17:02 +0000 Unknown Google Cloud https://jacek-mar.github.io/ai-dev-blog/posts/cloud-agents-the-missing-layer-in-your-devops-pipeline.html https://jacek-mar.github.io/ai-dev-blog/posts/cloud-agents-the-missing-layer-in-your-devops-pipeline.html Promotion: Claude Sonnet 4.6 is free in Cloud Agents for 48 hours, ending February 28th. Try it now → Platform and DevOps teams have spent years wiring together CI/CD, infrastructure-as-code, observability, and incident management. The detection and notification layers are solid. A drift detection t Fri, 06 Mar 2026 11:17:02 +0000 Brian Turcotte KiloCode https://jacek-mar.github.io/ai-dev-blog/posts/a-developers-guide-to-production-ready-ai-agents.html https://jacek-mar.github.io/ai-dev-blog/posts/a-developers-guide-to-production-ready-ai-agents.html Technical Solutions Manager Sr. Staff ML Engineer & Founder of Gen AI Intensive, Google Our most intelligent model available yet for complex tasks on Gemini Enterprise and Vertex AI Something has shifted in the developer community over the past year. AI agents have moved from "interesting research c Fri, 06 Mar 2026 11:17:02 +0000 Unknown Google Cloud https://jacek-mar.github.io/ai-dev-blog/posts/mixture-of-experts-moes-in-transformers.html https://jacek-mar.github.io/ai-dev-blog/posts/mixture-of-experts-moes-in-transformers.html Over the past few years, scaling dense language models has driven most progress in LLMs. From early models like the original ULMFiT (~30M parameters) or GPT-2 (1.5B parameters, which at the time was considered "too dangerous to release" 🧌), and eventually to today’s hundred-billion–parameter systems Fri, 06 Mar 2026 11:17:02 +0000 Unknown HuggingFace https://jacek-mar.github.io/ai-dev-blog/posts/from-dario-amodei-on-our-discussions-with-the-department-of-war-announcementsa-statement-from-our-ce.html https://jacek-mar.github.io/ai-dev-blog/posts/from-dario-amodei-on-our-discussions-with-the-department-of-war-announcementsa-statement-from-our-ce.html I believe deeply in the existential importance of using AI to defend the United States and other democracies, and to defeat our autocratic adversaries. Anthropic has therefore worked proactively to deploy our models to the Department of War and the intelligence community. Thu, 26 Feb 2026 00:00:00 +0000 Unknown Claude https://jacek-mar.github.io/ai-dev-blog/posts/minimax-25-vs-glm-5-across-3-coding-tasks-benchmark-results.html https://jacek-mar.github.io/ai-dev-blog/posts/minimax-25-vs-glm-5-across-3-coding-tasks-benchmark-results.html GLM-5 and MiniMax M2.5 are two new open-weight models now available in Kilo Code. MiniMax M2.5 scores 80.2% and GLM-5 scores 77.8% on SWE-bench Verified, putting them very close to GPT-5.2 and Claude Opus 4.6 at a fraction of the cost. We ran both through three coding tasks in Kilo CLI, where they w Fri, 06 Mar 2026 11:17:02 +0000 Darko KiloCode https://jacek-mar.github.io/ai-dev-blog/posts/gpt-53-codex-is-live-in-kilo.html https://jacek-mar.github.io/ai-dev-blog/posts/gpt-53-codex-is-live-in-kilo.html Forget everything you knew about “coding models” being specialized, terse, or limited to a terminal. GPT-5.3-Codex was recently made available for developers in OpenAI’s Responses API, and it was live in Kilo within minutes. Fri, 06 Mar 2026 11:17:02 +0000 Ari KiloCode https://jacek-mar.github.io/ai-dev-blog/posts/25-2026announcementsanthropic-acquires-vercept-to-advance-claudes-computer-use-capabilities.html https://jacek-mar.github.io/ai-dev-blog/posts/25-2026announcementsanthropic-acquires-vercept-to-advance-claudes-computer-use-capabilities.html People are using Claude for increasingly complex work—writing and running code across entire repositories, synthesizing research from dozens of sources, and managing workflows that span multiple tools and teams. Computer use enables Claude to do all of that inside live applications, the way a person Wed, 25 Feb 2026 00:00:00 +0000 Unknown Claude https://jacek-mar.github.io/ai-dev-blog/posts/we-wasted-4-weeks-on-a-1000month-ai-agent.html https://jacek-mar.github.io/ai-dev-blog/posts/we-wasted-4-weeks-on-a-1000month-ai-agent.html Every SaaS platform you use is shipping an AI agent add-on right now. Your project management platform probably just announced one last week. And if your experience is anything like ours, they mostly don’t work. We spent four weeks trialing the AI agent add-on from our support ticketing platform. Fri, 06 Mar 2026 11:17:02 +0000 Alex Gold KiloCode https://jacek-mar.github.io/ai-dev-blog/posts/kiloclaw-is-now-generally-available-with-500-models-and-a-new-agent-benchmark.html https://jacek-mar.github.io/ai-dev-blog/posts/kiloclaw-is-now-generally-available-with-500-models-and-a-new-agent-benchmark.html Two weeks ago, we announced KiloClaw, a fully managed way to run OpenClaw without dealing with servers, configuration files, or 3 AM crashes. Today, KiloClaw is generally available. In the first two weeks, more than 3,500 developers joined the waitlist. Early access users have already been spinning Fri, 06 Mar 2026 11:17:02 +0000 Brendan O'Leary KiloCode https://jacek-mar.github.io/ai-dev-blog/posts/deploying-open-source-vision-language-models-vlm-on-jetson.html https://jacek-mar.github.io/ai-dev-blog/posts/deploying-open-source-vision-language-models-vlm-on-jetson.html Vision-Language Models (VLMs) mark a significant leap in AI by blending visual perception with semantic reasoning. Moving beyond traditional models constrained by fixed labels, VLMs utilize a joint embedding space to interpret and discuss complex, open-ended environments using natural language. The Fri, 06 Mar 2026 11:17:02 +0000 Unknown HuggingFace https://jacek-mar.github.io/ai-dev-blog/posts/24-2026policyanthropics-responsible-scaling-policy-version-30.html https://jacek-mar.github.io/ai-dev-blog/posts/24-2026policyanthropics-responsible-scaling-policy-version-30.html We’re releasing the third version of our Responsible Scaling Policy (RSP), the voluntary framework we use to mitigate catastrophic risks from AI systems. Anthropic has now had an RSP for more than two years, and we’ve learned a great deal about its benefits and its shortcomings. Tue, 24 Feb 2026 00:00:00 +0000 Unknown Claude