$ cat /var/log/thoughts/tag/open-source

Open Source

Posts tagged with Open Source.

Filters1

16 posts

Qwen 3.8 overthinks on the default reasoning setting

Qwen 3.8 overthinks on the default xhigh setting. A pelican SVG ate 22,276 thinking tokens and 21 minutes. Start on low or off, then raise it for hard jobs.

Strong OpinionsAI CodingDeveloper ToolsSelf-HostingOpen SourceTrend Commentary1 min read · Aug 21, 2026

OpenCode vs Claude Code on the Same Tasks: Same Passes, Different Paths

OpenCode vs Claude Code on four graded tasks. Both hit 4/4. Paths diverged on tools, wall time, and reported tokens. Product lane, not a same-model lab.

Meta AnalysisAI CodingClaude CodeOpen SourceDeveloper ToolsPerformance1 min read · Aug 11, 2026

Eleven vendors back ai-catalog.json manifests under their own domains

Agentic resource discovery launched in June with eleven vendor backers. A scan of 44 domains six weeks later found a working catalog on exactly one of them.

Meta AnalysisMCPAPI DesignOpen SourceDeveloper ToolsTrend Commentary1 min read · Jul 30, 2026

Independent verification of Bun memory usage claims

Bun's Rust rewrite claims lower memory usage. A firsthand build-loop test on the exact version Bun benchmarked checks whether those numbers actually hold up.

Meta AnalysisBunRustPerformanceAI CodingOpen Source1 min read · Jul 29, 2026

Open Source Has a Distribution Problem, Not a Building Problem

Open source distribution decides whether your project gets found. A one-week census of 106 launches, and the title that scored 154 in one sub and 1 in another.

Meta AnalysisOpen SourceDeveloper ToolsTrend Commentary1 min read · Jul 28, 2026

The Open Weights Rules Nobody Can Comply With

Every open weights rule on the table stops working the moment a download finishes. Sort the three asks by when they act and see which ones still hold.

Strong OpinionsOpen SourceMLSecuritySelf-HostingTrend Commentary1 min read · Jul 28, 2026

Gemma 4 Is Google's Best Open Model Yet (Here's How to Use It)

Gemma 4 ships Apache 2.0, 256K context, and four model sizes. Benchmarks beat Llama 4 across the board. Quick use cases with Ollama and Transformers code.

Strong OpinionsAI CodingMLOpen SourceSelf-HostingDeveloper Tools1 min read · Apr 10, 2026