I learned by having AI write code: 'Working' and 'Correct' are different. ## A 12-week experiment log for humanities majors and non-engineers | Week 2

In Week 1, I consciously used PowerShell and Python for the first time.I learned the basics of writing short code myself, ...

How to get AI to write code for reading CSVs: How to check if it breaks with commas or line breaks

If it’s a CSV, can’t I just read it by splitting by commas?”I ask an AI to write code to read a list of stationery. If it returns a table-like result without any errors, I might feel relieved for a ...
Tech Critter

HTTP2Shell Flaw: Critical Brotli Vulnerability Explained

HTTP2Shell is a critical flaw in Google's Brotli decoder affecting Node.js, Python, Bun and other software. Learn about the ...
Anthropic

LLMs are making formal software verification feasible at a much larger scale

Claude formally proved security properties in the compiled code of 123 open-source libraries. Along the way, it found bugs ...
Anthropic

How AI agents shape experts’ judgment calls

In a randomized trial, professional analysts working with an AI coding agent went with the option the agent was randomly ...

A coding agent’s harness can triple its cost without making it more accurate, study finds

Researchers ran five models through Claude Code, OpenCode and mini-SWE-agent. Swapping harnesses changed scores no more than ...
xenospectrum

Microsoft Releases Mage-VL-Cowork, a 4B-Parameter Computer-Use Model That Scores 66% on OSWorld-Verified

Cowork, a small 4B vision-language model that switches between GUI actions and shell code execution within a single policy, ...