Bookmarks

As an organization that has long held its governance in trust by the maintainers of the project, I propose that we establish a transparent public governance plan. It would be beneficial for everyon...

Ruby Community, At the heart of Ruby Central’s mission is our responsibility to steward the open source tools that power the Ruby ecosystem. That commitment is only as strong as the people and processes behind it. Over the past several months, we have been carefully reviewing how RubyGems.org,

Think of it like having a virtual, specialized software development team right inside your editor, orchestrated by the 👑 Roo Commander, powered by Roo Code on VS Code - jezweb/roo-commander

Project management system for Claude Code using GitHub Issues and Git worktrees for parallel agent execution. - automazeio/ccpm

Agent OS is a system for better planning and executing software development tasks with your AI agents. - buildermethods/agent-os

Language service for Docker Compose documents. Contribute to microsoft/compose-language-service development by creating an account on GitHub.

Reproducibility is a bedrock of scientific progress. However, it’s remarkably difficult to get reproducible results out of large language models. For example, you might observe that asking ChatGPT the same question multiple times provides different results. This by itself is not surprising, since getting a result from a language model involves “sampling”, a process that converts the language model’s output into a probability distribution and probabilistically selects a token. What might be more surprising is that even when we adjust the temperature down to 0This means that the LLM always chooses the highest probability token, which is called greedy sampling. (thus making the sampling theoretically deterministic), LLM APIs are still not deterministic in practice (see past discussions here, here, or here). Even when running inference on your own hardware with an OSS inference library like vLLM or SGLang, sampling still isn’t deterministic (see here or here).

A breakdown of Claude's memory implementation, how it differs from ChatGPT, and what these opposite approaches reveal about each company's philosophy.

こんにちは。 ファインディ株式会社 で Tech Lead をやらせてもらってる戸田です。 現在のソフトウェア開発の世界は、生成AIの登場により大きな転換点を迎えています。 GitHub CopilotやClaude Codeなど生成AIを活用した開発支援ツールが次々と登場し、開発者の日常的なワークフローに組み込まれつ…

This is a technical report on three bugs that intermittently degraded responses from Claude. Below we explain what happened, why it took time to fix, and what we're changing.

Cloudflare’s Dashboard and a set of related APIs were unavailable or partially available for an hour starting on Sep 12, 17:57 UTC. The outage did not affect the serving of cached files via the Cloudflare CDN or other security features at the Cloudflare Edge.

A developer's insights on transitioning from React to Solid.js, exploring key differences in reactivity, JSX, APIs, and web components

We Love You, TypeScript

For enterprise software, the software supply chain presents some of the biggiest risks today to data privacy and security.

Scale your Rails app without rewrites. Learn to keep business logic in Ruby and delegate heavy work to specialized tools, built with Go, C, and Rust, for performance and scalability.

pnpm's new minimumReleaseAge setting delays package updates to prevent supply chain attacks, with other tools like Taze and NCU following suit.

The 1Password browser extension is now available in Comet, the AI-powered browser by Perplexity, giving you secure, end-to-end encrypted access to your credentials and sensitive information as you explore the AI-powered internet.

Post-quantum cryptography protects against quantum threats using today’s hardware. Quantum tech like QKD may sound appealing, but it isn’t necessary or sufficient to secure organizations.

A practical guide to securing MCP model–agent interactions: prevent prompt injection, privilege escalation, replay attacks, and data exfiltration with validation gateways, signing, DLP, and scoped creds.

We replaced slow JSON path lookups with Bloom filters in our global routing service, cutting memory usage by 15% and reducing 99th percentile lookup times from hundreds of milliseconds to under 1 ms. Here’s how we did it.

How and why to use the Outbox pattern to build a robust event-driven system.

A deep dive into horizontal scaling: how we sharded our core db without any downtime or dropped requests.

HTML forms library for ruby. Contribute to jeremyevans/forme development by creating an account on GitHub.

Pidfd implementation for the Ruby language. Contribute to mensfeld/pidfd development by creating an account on GitHub.

Find the mime type of files, examining file, filename and declared type - rails/marcel

Simple, extremely lightweight, extensible, configuration management library for Go. Supports JSON, TOML, YAML, env, command line, file, S3 etc. Alternative to viper. - knadh/koanf