How to Compare the Models Behind a Coding Agent with SWE-bench
How do you measure the model performance of a coding agent?
How do you measure the model performance of a coding agent?
Why the layers piled up, why we returned to the server, and why the stack survives even after agents. And why the survival of a stack and the value of the people who know it are two separate things
The friction that made judgment expensive is the same friction that taught it. Why the skill growing most valuable in the AI era is the one being cultivated least
If the spec is satisfied and the bugs get fixed, do you still need to be able to read the code? Understanding doesn't disappear: this essay traces where it moves.
The claim that job boundaries are collapsing is only half true. What collapses is production; judgment and responsibility remain.
Bun moved roughly 960,000 lines from Zig to Rust in six days using Claude Code. The meaning here is not code quality but OSS governance and resource asymmetry.
I'm definitely coding faster, but why are my codebase and skills staying the same? Examining the gap between perceived and proven benefits.
AI coding tools aren't changing developers—they're changing the nature of what developers do.
Complete overview of core coding agent concepts: Rules, Commands, MCP, Sub-agents, Hooks, Skills, and Plugins
In July 2025, the browser GPU API changed generations for the first time in 14 years
I don't know what the future holds, but shouldn't we just keep developing diligently?