nick

software engineer @google. prev: @aws.
opinions are my own

[nick @thecsguy]

codex on Android - performant and reliable! 🚀

https://t.co/yn6OKSfAsU

[nick @thecsguy]

@linderps Get well soon!! Glad to see you back posting again ! 💪

[nick @thecsguy]

@laurenlself A or H since there is no upper/lower bound 😆

[nick @thecsguy]

Accurate in most cases

[Ray🫧 @ravikiran_dev7]

Unpopular opinion: it's true 👀

Do you agree with this??

[nick @thecsguy]

@jxnlco No but that should change

[ThePrimeagen @ThePrimeagen]

I super don't understand this. I have to believe you're writing no software of any consequence for this to be true.

[nick @thecsguy]

@guinnesschen an official demo and its not using xhigh fast? 😆

[nick @thecsguy]

@Irenezhao_ why not SF?

[joyce @henloitsjoyce]

try it out here: https://t.co/IRqag0OzoA

join the discord and hang out with us: https://t.co/MUVTnWCJvi

[Guinness Chen @guinnesschen]

If you ever get tired of managing your Codex threads, just let Codex manage itself! Codex can now create threads, search them, organize them, pin the important ones, and spin up worktrees for parallel tasks.

[Theo - t3.gg @theo]

I think Codex stopped using Electron 👀

The owl was a big hint, the custom architecture used for the ChatGPT Atlas browser was called "OWL" (OpenAI’s Web Layer)

[nick @thecsguy]

@dear_kxtie when?

[Mitchell Hashimoto @mitchellh]

I've got an agent in a loop optimizing a renderer with the goal to minimize frame times (and tests to measure). It got times down from 88ms to 2ms and allocations down from ~150K to 500. Sounds good, right? Wrong. This is exactly why agent psychosis is a big fucking problem.

As an experiment, I rewrote the Ghostty core render state in Go, with access to identically laid out data structures as Ghostty and the exact same validation tests. I made a purposely naive renderer (simple, correct, but slow). 88ms per frame with 150,000 allocations (horrendous, lol)!

I then kickstarted a Ralph loop to bring the frame times down. I told it it can't modify input data structures or the public API or tests (they're correct), but it can do anything else it wants. It got to work.

It has worked for about 4 hours. I've spent around $350 on this experiment so far. The results?

88ms => 1.5ms
150K allocs => ~500 allocs

Incredible right? Nope.

My hand-written renderer I ported has frame times (same benchmark) of ~20us (0.020ms) and 0 allocations in the update path.

This is the problem with psychosis and lacking systems understanding. If you don't understand the system, you're going to accept that this is an incredible result. If you understand the system, you'll see better solutions immediately and can do roughly 75x better on throughput.

The people who blindly trust agent output are in the former camp. They're sheeple, overdrinking from a fountain of mediocrity.

Standard disclaimer: I use AI all the time. I like AI. The point I'm making is to not blindly accept results. Think. Analyze. Learn.

[nick @thecsguy]

@theo HAHAHAHAHAHAHA that's a record

[nick @thecsguy]

@googlegemma Can we get one that runs Gemma 4 31b?

[nick @thecsguy]

4.6 is trash

[ThePrimeagen @ThePrimeagen]

Now that we have Claude 4.8 how long until 4.6 is called trash?

[nick @thecsguy]

@MarioVerbelen C++ is best

[nick @thecsguy]

@zekramu 80% can be solved by just using a hashmap

[Daniel @growing_daniel]

I've been a software engineer for years but I've never had to interview for a job before. I tried leetcode once out of curiosity and couldn't do it. Seemed really disconnected from what software engineers do.

[Cursor @cursor_ai]

We're hosting an event on June 16th in San Francisco.

Compile is a one-day event that brings together engineers, researchers, designers, and builders of all kinds to discuss the future of software.

https://t.co/8YERlPFooL

[nick @thecsguy]

@skeptrune Don't