Adnan M.

Software is pure “thought stuff”. One person can write code and billions can run it. If anything, our linear time produces exponential value. Therefore, I’ve never personally believed that developer time is expensive, that we have a “typing” problem. Or that English is somehow a better way to express code than a language as explicit as Zig. Granted, there’s tons of (non valuable) bespoke software that LLMs can now create. But the valuable thought stuff? Great systems coders are becoming more valuable than ever. In the land of the blind, the one eyed man is king.

🚨 We recently discovered that an unauthorized party obtained a token with access to the Grafana Labs GitHub environment, enabling the threat actor to download our codebase. (1/6)

@ThePrimeagen Did you ever write down somewhere, what principles guide your writing of software?

It's weird how much time I spend trying to get agents to write code like me Maybe I should just write code...

The daily standup isn't a status meeting, it's a confession booth. Lining engineers up to recite yesterday, justify today, surrender tomorrow — that isn't management. It's the ritual an insecure manager performs to disguise the absence of a real communication channel.

@yngmisu Life hack: instead of driving to city X and pay for the fuel, you can use your legs and walk.

life hack: instead for paying for a subscription you can write code by hand

@fayazara What happens when I make a mistake and make 500 million worker requests?

The Cloudflare $5/mo plan is honestly absurd Workers - 10M requests D1 - 25B reads, 50M writes KV - 10M reads, 1M writes R2 - 10 GB storage, zero egress Email - $0.35 for 1000 emails Browser - 10 hours per month You also get Durable Objects, Queues, Workflows - included, Unlimited Hyperdrive queries, Vectors.

One of the biggest problems with using LLMs as a google replacement for programming, is that getting zero relevant results on google used to be a signal that you had the wrong idea about the root cause. Whereas LLMs will happily indulge any terrible idea you suggest.

Bait on S3 is faster than Bait on NVMe

How much of SQLite, FFmpeg, PHP compiler can LMs code from scratch? Given just an executable and no starter code or internet access. Introducing ProgramBench: 200 rigorous, whole-repo generation tasks where models design, build, and ship a working program end to end. 🧵

Database transactions don't get enough love but the ability to execute a bunch of code, change a bunch of data, and only commit it when you've validated the results is going to be so critical in the AI era.

the thing with being a leader is you get to take credit for your team's work so when things go wrong it's only fair that you pass the blame onto them

Usually it goes like, look at this tool, it's 50x faster than the one we use. After that you know it's rust. But often it lacks the full feature set, unfortunately.

Oxfmt can now format your JSDoc comments! 👀 All you need is setting the `jsdoc` option in your Oxfmt config file, either as default empty object or with custom options.

If you think you're growing by reading shit on the internet without synthesizing it into wisdom through the struggle and uncertainty of solving difficult problems, you haven't figured out learning yet. Claude is not making you smarter.

Keep an eye on this

$999 might sound like a lot for a fully custom fit wheelchair. But it's 80% cheaper than it used to be. Iykyk.

@michael_chomsky Lgtm

Ok how tf do I review this PR?

you used to spend a day messing with your neovim config, feel self conscious, then get back to work now people are spending weeks on some hyper customized coding agent workflow that definitely is worse than vanilla but they can talk about it like they're ahead of the game

So much bragging about LoC changed and PRs merged. Just pick an ambitious project and brag about it when it’s done. The PRs aren’t the goal.

Both Codex and Claude Code’s desktop apps are using Oxlint & Oxfmt :) Codex also already on Vite 8!