▶ Videos →

📰 Dev.to · shinji shimizu

10 articles · Updated every 3 hours · View all reads

All Articles 121,436 Blog Posts 128,060 Tech Tutorials 31,113 Research Papers 24,281 News 17,495 ⚡ AI Lessons

Fitting LLM Reply Suggestions Into Every Provider's Prompt Cache — Without Structured Output

Dev.to · shinji shimizu 🧠 Large Language Models ⚡ AI Lesson 1mo ago

Fitting LLM Reply Suggestions Into Every Provider's Prompt Cache — Without Structured Output

How inline markers beat structured output for adding reply suggestion chips to a streaming, prompt-cached voice roleplay chat — without breaking any provider's

My high-res image-to-video kept OOMing — turns out I was decoding outside no_grad

Dev.to · shinji shimizu 1mo ago

My high-res image-to-video kept OOMing — turns out I was decoding outside no_grad

A 96GB GPU couldn't run 1024x768 I2V (83.5 GiB peak). The 54 GiB wasn't the model — it was an autograd graph from consuming a lazy VAE decode iterator outside n

Building a Sarcastic AI English Tutor with Persona-as-Code and Gemini Audio Input for Pronunciation Correction

Dev.to · shinji shimizu 🧠 Large Language Models ⚡ AI Lesson 1mo ago

Building a Sarcastic AI English Tutor with Persona-as-Code and Gemini Audio Input for Pronunciation Correction

How I built a high-attitude AI English tutor using persona-as-code design, Qwen3-ASR for multilingual STT, and Gemini audio input for real pronunciation feedbac

Five Years Later, I Finally Have 96GB VRAM — What It Actually Unlocks for Agent Loops

Dev.to · shinji shimizu 🤖 AI Agents & Automation ⚡ AI Lesson 1mo ago

Five Years Later, I Finally Have 96GB VRAM — What It Actually Unlocks for Agent Loops

Not a GPU unboxing. A real look at what 96GB VRAM enables for multi-model agent pipelines — and where it still hits its limits.

Turning a 1-Line Idea Into a 40-Second Short with a 10-Beat Local Video Pipeline

Dev.to · shinji shimizu 🛠️ AI Tools & Apps ⚡ AI Lesson 1mo ago

Turning a 1-Line Idea Into a 40-Second Short with a 10-Beat Local Video Pipeline

Full pipeline: Gemma 4 31B expands a one-liner into a 10-beat script, HiDream generates images, LTX-2 I2V renders clips, and ffmpeg assembles everything — all o

Running LTX-2.3 Alongside TTS on a Single 96GB GPU with a Cold-Start Architecture

Dev.to · shinji shimizu 🧠 Large Language Models ⚡ AI Lesson 1mo ago

Running LTX-2.3 Alongside TTS on a Single 96GB GPU with a Cold-Start Architecture

How to go from 86 GiB idle VRAM (instant OOM) to 0 GiB idle / 40 GiB peak by using a cold-start design for LTX-2.3 on one RTX Pro 6000 Blackwell.

Cutting LTX-2 22B Peak VRAM by 40% with fp8_cast — and Why optimum-quanto Was a Trap

Dev.to · shinji shimizu 🧠 Large Language Models ⚡ AI Lesson 1mo ago

Cutting LTX-2 22B Peak VRAM by 40% with fp8_cast — and Why optimum-quanto Was a Trap

How fp8_cast reduced LTX-2 22B peak VRAM from 40 GiB to 24 GiB in cold-start mode, and why optimum-quanto silently breaks the transformer.

HiDream Skeleton Mode: Prompt Beats OpenPose Ref — 8 Patterns Benchmarked

Dev.to · shinji shimizu 👁️ Computer Vision ⚡ AI Lesson 1mo ago

HiDream Skeleton Mode: Prompt Beats OpenPose Ref — 8 Patterns Benchmarked

Benchmarking HiDream-O1-Image skeleton mode across 8 patterns reveals 3 counterintuitive findings about openpose refs, resolution drops, and shift values.

Replicating a Language-Learning Comedy Short with Claude Code — Gemini as a Multimodal Sub-Agent

Dev.to · shinji shimizu 🧠 Large Language Models ⚡ AI Lesson 1mo ago

Replicating a Language-Learning Comedy Short with Claude Code — Gemini as a Multimodal Sub-Agent

Building a local GPU + Gemini 3.1 Pro hybrid pipeline that generates publishable comedy Shorts from a single line of text in under 60 seconds.

HiDream-O1-Image 3–8x Faster: Benchmarking Steps, CFG, and Resolution

Dev.to · shinji shimizu 🧠 Large Language Models ⚡ AI Lesson 1mo ago

HiDream-O1-Image 3–8x Faster: Benchmarking Steps, CFG, and Resolution

Real-world timing benchmarks for HiDream-O1-Image Full — tuning steps, guidance scale, and resolution to speed up iteration without killing quality.