@alanwest

Why Your LLM Classification Pipeline Fails on Edge Cases (and How to Fix It)

14h ago · 6 min read · A Harvard study recently made waves: OpenAI's o1 model reportedly diagnosed 67% of emergency room patients correctly, compared to 50-55% accuracy from triage doctors. Whether or not that number holds up under scrutiny, it highlights something develop...

Why Every Website Wants to Access Your Local Network (And What to Do About It)

18h ago · 5 min read · If you've been browsing the web recently, you've probably noticed a new kind of permission prompt popping up: "This site wants to access devices on your local network." It showed up for me on a random dashboard I was building, and my first thought wa...

Why Your Barman Backups Keep Failing (And How to Actually Fix It)

19h ago · 5 min read · So you finally set up Barman to handle your PostgreSQL backups. You followed the docs, configured your server, ran barman check and... a wall of FAILED messages stares back at you. Cool. Very reassuring for your disaster recovery strategy. I've been ...

AI Coding Has Its Own Language Now — Here's How to Decode It

21h ago · 6 min read · If you've tried to follow any AI coding discussion in the last six months, you've probably felt like everyone suddenly started speaking a dialect you never signed up to learn. "Vibe coding." "Agentic workflows." "Context windows." "Prompt engineering...

Running LLMs on Windows: Native vLLM vs WSL vs llama.cpp Compared

22h ago · 5 min read · The Windows local LLM story just got interesting. Someone recently demonstrated Qwen3's 27B model running at 72 tokens per second on an RTX 3090 — natively on Windows. No WSL. No Docker. Just a portable vLLM launcher. If you've been running local mod...