16 local model configurations, 56 hidden-test coding tasks, 36 full runs, one 16 GB...
A 7.5B model beat a 24B on my coding benchmark.
Related
Building Arogya Seva: How I Built an Ultra-Low Latency Telehealth Voice AI for Bharat in 10 Days
Introduction: The Problem & The Mission Across rural and semi-urban India, access to timely...
I scanned the code of 25 MCP servers — what they can do to your machine before you sandbox them
An MCP server isn't a URL — it's code that runs on your machine with your environment. I statically scanned 25 popular servers for network calls, secret-env reads, and subprocess s...
I Build an AI Companion App Alone. It Runs on Nobody's Servers But Yours.
I Build an AI Companion App Alone. It Runs on Nobody's Servers But Yours. I was using one of the...