Browse every SikiT article in reverse chronological order, or use search and the topic shortcuts below.
Runtime & failure · Delegation & permissions · Models & tooling · Open questions
-
Qwen 3.8 27B thinks by default, and the default effort is xhigh
See the official Qwen 3.8 27B defaults, then change effort or turn analysis off before judging speed or quality.
-
Private AI can protect your input without making every outcome private
See what encrypted inference protects, which privacy layers remain separate, and how to assess a private-AI claim without overstating it.
-
Disconnect a Named Pipe Without a Flush Hang
A Win32 shutdown guide that separates the blocking buffer drain from DisconnectNamedPipe and makes graceful delivery versus bounded discard explicit.
-
Who Spends the Time Your AI Saves?
A four-condition test for when an AI agent saves its user’s time by creating review and response work for other people.
-
How Much Work Can You Delegate to AI? Five Levels from Answer to Finished Work
A five-level decision framework for choosing how much of a real task AI should answer, transform, produce, act on, or repeat.
-
Who Is Free-Riding on AGI?
A concise argument that the free-rider label should be tested against three ledgers: development cost, wider contribution, and control of the result.
-
Did I Mean It Before the Machine Said It?
A philosophical argument about what happens when AI-generated language arrives before settled intention, is attributed backward as an original thought, and is enacted forward as a commitment.
-
How to Stop a Windows Named-Pipe Server from Waiting Forever
A finite Win32 wait-and-cancel path that preserves pending I/O state until Windows reports normal completion, cancellation, or failure.
-
GitHub Copilot Retired Gemini 2.5 Pro and Gemini 3 Flash: What to Switch To
GitHub Copilot retired Gemini 2.5 Pro and Gemini 3 Flash on July 31, 2026. GitHub currently points deeper-reasoning work to Gemini 3.1 Pro (Preview) and faster work to…
-
GPT-5.6 Model Selection Guide: Sol, Terra, Luna, and Reasoning Effort
GPT-5.6 Luna and Terra became cheaper on July 30, 2026. That makes them more attractive for repeatable work, but it does not make model choice a simple “light…
-
GitHub Models Is Retired: What Broke and Where to Move Next
GitHub Models is no longer available. If its playground, model catalog, inference API, or bring-your-own-key setup stopped working, GitHub says the July 30, 2026 shutdown applies to every…
-
ChatGPT’s Unified Desktop App: Chat, Work, and Codex Explained
If a Codex update changed the app’s name and icon to ChatGPT, or left ChatGPT Classic beside it, the installation may be working as designed. OpenAI’s unified desktop…