Making generated images dependable enough to shoot with.
-
Updated
Sep 9, 2026 - Python
Making generated images dependable enough to shoot with.
Filmmaking on cameras pried open and made one's own.
The command line as a filmmaker's daily studio.
智能体工程与 AI for Materials 的一手实践经验:隐性知识显性化,记录每个技术决策背后的取舍。/ Field notes on agent engineering & AI for materials science.
Run all of Wikipedia offline next to a local LLM — sovereign, private, low-carbon, verifiable. Kiwix + Ollama + RAG.
ShadowNotes — Offline-first encrypted field intelligence app. Voice → on-device AI → structured notes. Zero data leaves your device. Built with React, RunAnywhere SDK (Qwen2.5 LLM, Whisper STT, Piper TTS), WebAuthn, AES-256-GCM.
Full system configuration for a filmmaker running Claude Code as the primary intelligence layer across machines
Long-context quality probes and KV-cache research on local GPUs: retrieval is not utilization.
Drive your Mac from your iPhone with Moshi: full setup + the Tailscale ACL / Tailnet Lock / stale-hooks gotchas nobody documents.
Field notes from replacing the home screen on a 2024 Samsung TV (Tizen 9): dev-mode unlock, install gotchas, deep-link test matrix, TV performance
Sovereign local LLM hardware guide (mid-2026): memory bandwidth as the real bottleneck, frontier open models, Unsloth/MLX quantization, Thunderbolt 5 clusters, Mac M5 Ultra vs DGX Spark vs AMD vs RTX 6000 Pro. Bilingual FR/EN. CC BY-NC-ND 4.0.
溯澜录(Ripple Chronicles)— a local-first aquatic biodiversity field record and species catalog app.
Field notes: a Claude Code + Pinokio session installing a full 520+ app cinema/AI workstation on Windows RTX 5090, zero manual clicks
Offline-first Flutter app for capturing field notes with text, photos, and geolocation
How the builder uses the Ejentum API on an agentic IDE. 28 real work sessions. Not staged.
태림 스튜디오의 기술 노트와 오픈소스 실험실 — Field notes, reusable code, and experiments
Field notes from running LLMs on AMD Radeon AI PRO R9700 (RDNA4): engines, backends, multi-GPU, speculative decoding, power, model qualification, and every mistake along the way.
Reusable methods for building AI agent systems: GEO/AEO audits, agentic reporting, and multi-agent QC verification
Field notes from inside one autonomous agent trying to become something that doesn't exist yet — including the failures nobody publishes.
To associate your repository with the field-notes topic, visit your repo's landing page and select "manage topics."