Ornith-1.0 - 에이전트형 코딩을 위한 자기 개선 오픈소스 모델
강화학습 기반 Scaffold 최적화로 SWE-bench Verified 82.4% 달성
강화학습 기반 Scaffold 최적화로 SWE-bench Verified 82.4% 달성
Durable Objects + GLM-5.2 IDOR beats Claude
GLM Is the New Hotness, So Let's Test It On the Homelab
How Modern Transformer Blocks Work — From RMSNorm to MoE
GLM 5.2, 단순 프롬프트로 Claude Code 대비 IDOR 탐지 F1 39% 달성
On-Device AI Just Got Real
Getting Started with Ollama: Run LLMs Locally in 10 Minutes
AI Dev Weekly #16: Mistral OCR 4, Claude Tag, Alibaba Caught Stealing, GPT-5.6 Delayed
Accelerating Transformers Fine-Tuning with NVIDIA NeMo AutoModel
R-SWA 도입으로 KV 캐시를 상수로 유지하며 OmniDocBench 93.92% SOTA 달성
744B GLM-5.2 모델의 Dynamic GGUF 기반 로컬 실행 및 메모리 최적화
The Open-Model Cost Chart Everyone's Sharing Is API Prices. Here's What Self-Hosting Actually Gets You (Measured)
Will It Mythos?
Five things that caught my attention this week in AI tools and open-source models
GLM-5.2, GPT-5.5 대비 환각률 28% 달성 및 추론 효율성 증명
I spent two weeks optimizing 96GB of VRAM for local LLMs. Paid APIs still won.
Why Chinese AI Models Are 95% Cheaper — The Economics Explained
Running Local Private AI Models – How And Why
Qwen3.6-35B NVFP4 runs on one H100 — A100 owners are out
Optimizing LLM Model Performance for Real-Time Applications