Serious Data From Testing LLMs
A data-driven analysis of LLM performance on a simple retrieval task, highlighting the need for evidence-based AI testing.
A data-driven analysis of LLM performance on a simple retrieval task, highlighting the need for evidence-based AI testing.
A curated summary of recent technical blog posts from Microsoft, Azure, .NET, GitHub, and Visual Studio, covering AI, DevOps, and development tools.
Learn to build AI agents in C# using Microsoft Agent Framework and Hugging Face MCP tools for image generation and analysis.
PostgreSQL 19 adds IGNORE NULLS/RESPECT NULLS options to window functions like LAG and FIRST_VALUE, improving data handling.
A guide to building multi-model AI orchestration workflows in .NET 9 using Microsoft Agent Framework, Azure AI, GitHub Models, and Ollama.
Developer details adding experimental JMAP support to pimsync, focusing on challenges with Etag/State handling and serialization libraries.
A guide to building a custom HTTP extension for Bicep, enabling API calls directly from Azure infrastructure templates.
Explains why and how to use CancellationTokens in .NET APIs to stop long-running operations and free server resources.
Explores the unique challenges of testing Generative AI and Large Language Models, contrasting it with traditional software testing approaches.
Introducing Claudiomiro, a Node.js CLI tool that enables Claude AI to autonomously complete complex coding tasks through a structured, looping workflow.
Azure Local Arc Gateway is now GA, providing a centralized HTTPS egress point for Azure Local workloads, simplifying security and reducing endpoints.
A guide on using GitHub Actions to securely checkout a specific branch from a private repository, enabling reusable components in CI/CD workflows.
A curated collection of tech articles covering design systems, data platforms, real-time alerts, microservices, and IoT security.
A look at gradient.style, a CSS tool for creating and sharing HDR gradients with sharable URLs and multi-layered imports.
PostgreSQL 19 introduces GROUP BY ALL, a new SQL feature that automatically groups by all non-aggregate columns, simplifying queries.
Analysis of changes in Claude Code 2.0's system prompt, showing less prescriptive guidance and more trust in the Sonnet 4.5 model.
Explores C++ arena allocation intricacies, lifetime semantics, and corrections to common memory management patterns, referencing C++20 changes.
The exec-maven-plugin now supports Java 25's JEP 512 launchable main methods, allowing execution of simpler void main() programs.
A monthly roundup of 78 curated links on data engineering, architecture, AI, and tech trends, with top picks highlighted.
EF Core 10.0 adds support for the new JSON data type in Microsoft SQL Server 2025, enabling querying and storage of JSON documents.