92 articles

📰 Dev.to · soy

Articles from Dev.to · soy · 92 articles · Updated every 3 hours · View all reads

All ⚡ AI Lessons (10463) ArXiv cs.AIDev.to · FORUM WEBDev.to AIForbes InnovationOpenAI NewsHugging Face Blog
Gemini 2.5 Flash x Nemotron 9B — Optimal Division of Roles for Cloud LLM and Local LLM
Dev.to · soy 1mo ago
Gemini 2.5 Flash x Nemotron 9B — Optimal Division of Roles for Cloud LLM and Local LLM
Why Combine Them? When designing AI workloads, it is not easy to simultaneously satisfy...
Reduce API Costs for Large-Scale Document Analysis with Gemini Context Caching
Dev.to · soy 1mo ago
Reduce API Costs for Large-Scale Document Analysis with Gemini Context Caching
What is Context Caching? Google Gemini's Context Caching is a feature that caches context...
Skit: The Man Obsessed with Claude Code
Dev.to · soy 1mo ago
Skit: The Man Obsessed with Claude Code
Comedy Sketch: The Man Possessed by Claude Code Characters: Niiyama: The...
Building a Free Research Agent with DuckDuckGo Search + Local LLM
Dev.to · soy 1mo ago
Building a Free Research Agent with DuckDuckGo Search + Local LLM
Why DuckDuckGo + Local LLM? When conducting research, using paid APIs (such as Brave...
Reducing Token Consumption in Claude Code — FTS5 Knowledge DB + Tiered Index Design
Dev.to · soy 1mo ago
Reducing Token Consumption in Claude Code — FTS5 Knowledge DB + Tiered Index Design
Problem If all coding conventions, test commands, and documentation for the entire project...
Building a 5-in-1 App with Local LLM and Flutter
Dev.to · soy 1mo ago
Building a 5-in-1 App with Local LLM and Flutter
Introduction "I want to leverage AI without sending data to the cloud." The biggest...
Leveraging Claude Code's MCP Server
Dev.to · soy 1mo ago
Leveraging Claude Code's MCP Server
Introduction: The Context Switching Problem in DB Operations SQLite is an excellent...
LoRA and FT Are Unnecessary: How to Approach Distilled Models
Dev.to · soy 1mo ago
LoRA and FT Are Unnecessary: How to Approach Distilled Models
Introduction Fine-tuning (FT) a distilled model is either ineffective or leads to...
Lineage of OSS Supporting the AI Development Stack: Its Origins and Creators
Dev.to · soy 1mo ago
Lineage of OSS Supporting the AI Development Stack: Its Origins and Creators
Local AI development environments are built upon numerous open-source technologies. This article...
Running NVIDIA Nemotron-Nano-9B-v2-Japanese Locally: Mamba SSM + Thinking Mode Support
Dev.to · soy 1mo ago
Running NVIDIA Nemotron-Nano-9B-v2-Japanese Locally: Mamba SSM + Thinking Mode Support
NVIDIA Nemotron-Nano-9B-v2-Japanese This is a 9B parameter LLM specialized for Japanese,...
Strategic Data Organization Techniques Using SQLite, JSONL, XML, and TSV: Lessons
Dev.to · soy 1mo ago
Strategic Data Organization Techniques Using SQLite, JSONL, XML, and TSV: Lessons
Introduction PatentLLM (patent search AI) and HanreiLLM (case law search AI) are both...
Python Environment Management with uv: Introduction and Practical Use of a High-Speed Package Manager Replacing pip/venv
Dev.to · soy 1mo ago
Python Environment Management with uv: Introduction and Practical Use of a High-Speed Package Manager Replacing pip/venv
What is uv? uv is a Rust-based Python package manager developed by Astral (Charlie Marsh)....
Automatically Prevent Port Conflicts and Dangerous Commands Proactively with Claude Code's Hooks Feature
Dev.to · soy 1mo ago
Automatically Prevent Port Conflicts and Dangerous Commands Proactively with Claude Code's Hooks Feature
What are Claude Code hooks? Claude Code's hooks feature enables event-driven automation...
Giving a 'Brain' to Minecraft NPCs with a Local LLM — Nemotron + Mineflayer Implementation Notes
Dev.to · soy 1mo ago
Giving a 'Brain' to Minecraft NPCs with a Local LLM — Nemotron + Mineflayer Implementation Notes
What We Want to Achieve Traditional Minecraft bots primarily relied on command-based...
Using Local LLMs as a "Batch Processing Engine" — A Design for Automatically Generating Artifacts from Your Own Data
Dev.to · soy 1mo ago
Using Local LLMs as a "Batch Processing Engine" — A Design for Automatically Generating Artifacts from Your Own Data
Using Local LLMs as a "Batch Processing Engine" — Designing Automated Artifact Generation...
Fast Searching 4 Million Patent Records with FTS5
Dev.to · soy 1mo ago
Fast Searching 4 Million Patent Records with FTS5
Introduction: The Limitations of LIKE Search When searching for "battery" in PatentLLM's...