lilting channel

Featured

Homebrew on a 2001 P/ECE: Windows 11 WinUSB and a from-source LLVM toolchain

Dug a 2001 Aquaplus P/ECE out of a closet and got my own C game running on it: WinUSB/Zadig for the dead driver, a from-source S1C33 LLVM toolchain, and a 15GB-RAM OOM.

アクセス1位

Tech

Claude Code 'court' bug: tool calls leak as text instead of running (Opus 4.8)

Claude Code sometimes emits a stray court token and raw <invoke> as plain text instead of running Read/Edit/Bash — a malformed tool call that never executes. How to tell it from a hung shell or a 'could not be parsed' error, on long Opus 4.8 sessions.

Latest Articles

TechAug 17, 2026

GitHub Advisory Database Still Shows the Rejected Fake SQLite CVEs as Critical

Checked August 17, 2026 via the GitHub Advisory Database API: all six fake SQLite CVEs rejected by MITRE still show CVSS up to 9.8 as type unreviewed, withdrawn_at null.

Security CVE Vulnerability SQLite AI OSS GitHub

TechAug 16, 2026

Inside a 4-Character Anima LoRA, From Prompt Mistakes to a Biased DiT

Eight ComfyUI experiments on a 4-character Anima (Qwen-DiT) LoRA in plain terms: the conditioning right before the DiT decides who appears, and outfit mix-ups come from a biased DiT LoRA.

LoRA AI Image Generation Anima Anima-Base Qwen ComfyUI Multi-character

TechAug 15, 2026updated

pi.dev + Qwen Built an Idol Label Site — Then Forgot to Use the Generated Images

Built a fictional idol label site with pi.dev + Qwen 3.7/3.8 and Astro. Generated characters, logos, and event photos via genserver — then shipped wireframes with SVG placeholders. Full failure log inside.

Astro pi.dev Qwen Skills AI Web Design Fictional Label

Experimental

TechAug 14, 2026

Hand-written 507-token prompt vs Qwen3-14B for an Anima 4-char LoRA band scene

Tested on ComfyUI: Qwen3-4B–14B prompt writing, a 4B encoder bridge, and split conditioning all lost to one hand-written 507-token band prompt. An 896-caption audit shows the real fix.

LoRA AI Image Generation Anima Anima-Base ComfyUI Qwen Experiment Multi-character

TechAug 13, 2026updated

Anima's 4-character LoRA high-fives whoever stands center, not who you name

Tested on an M4 Mac mini with one 4-girl Anima LoRA: expressions, poses, desk contact, and a high-five across 3 seeds. Sentence order and center slots decided who touched whom.

LoRA AI Image Generation Anima Anima-Base ComfyUI Qwen Experiment Multi-character

TechAug 12, 2026updated

Anima 4-char LoRA dresses the wrong girl in its DiT half, not the Qwen encoder

Tested on M4 Mac mini ComfyUI: with identical Qwen3-0.6B/T5 conditioning, the black tights land on the wrong girl only when the 4-char LoRA's DiT half is applied.

LoRA AI Image Generation Anima Anima-Base ComfyUI Qwen Experiment Multi-character

lilting channel

Featured

Homebrew on a 2001 P/ECE: Windows 11 WinUSB and a from-source LLVM toolchain

Claude Code 'court' bug: tool calls leak as text instead of running (Opus 4.8)

Latest Articles

GitHub Advisory Database Still Shows the Rejected Fake SQLite CVEs as Critical

Inside a 4-Character Anima LoRA, From Prompt Mistakes to a Biased DiT

pi.dev + Qwen Built an Idol Label Site — Then Forgot to Use the Generated Images

Experimental

Hand-written 507-token prompt vs Qwen3-14B for an Anima 4-char LoRA band scene

Anima's 4-character LoRA high-fives whoever stands center, not who you name

Anima 4-char LoRA dresses the wrong girl in its DiT half, not the Qwen encoder

Recently Updated