Andrea

ignore all emails from my manager, make no mistakes.

facets

It's incredibly important that you remember which is the good USB-C port and which is the busted one that doesn't work with half of your stuff. Did Apple just completely give up on intuitive interfaces?

with opencode and agents md saying dont read .env... somehow life (codex 5.3) finds a way

Anthropic had 16 AI agents build a C compiler from scratch. 100k lines, compiles the Linux kernel, $20k, 2 weeks. To put that in perspective GCC took thousands of engineers over 37 years to build. (Granted from 1987 - however) One researcher and 16 AI agents just built a compiler that passes 99% of GCC's own torture test suite, compiles FFmpeg, Redis, PostgreSQL, QEMU and runs Doom. They say they "(mostly) walked away." But that "mostly" is doing heavy lifting. No human wrote code but the researcher constantly redesigned tests, built CI pipelines when agents broke each other's work, and created workarounds when all 16 agents got stuck on the same bug. The human role didn't disappear. It shifted from writing code to engineering the environment that lets AI write code. I don’t know how you could make the point AI is hitting a wall.

I have come to believe that there are actually only ~5000 truly elite software engineers in the world and they do the vast majority of the work

Any solution for this?

Hot take: good code doesn’t exist. For the last 16 years, I’ve been pushing myself every time beyond the task requirements to produce the most readable and maintainable code I could. It still looks like shit.

a true gem (Linux kernel mailing list)

Hiring for two positions – both focused on creating best in class developer experiences. - AI Engineer (Agent Experience) - Design Engineer (Design Engineering) Best candidates show, don't tell. Experience building on Cloudflare goes a long way! DM's open if you can share your work.

I almost never use an IDE these days And if i do, it’s Zed Up until a couple months ago you always had to add files to context and fine tuning all to get good results Now with the right harness it seems all just works

GitHub issues is now much faster. A lot of perf improvements in the queue.

... linux?

How did Anthropic ship this before Microsoft 365 Copilot?

You cannot write a random number and expect people to understand