It's me again. I come bearing great news.
First of all, we have hit 20M active users for Codex some time this week. Second of all, this is cause for celebration and during the day we will credit every Codex and ChatGPT Work user with a BANKED reset that you can use at your own leisure. And we will have some other good news later too!
Now, on usage limits draining faster, while we're not seeing anything abnormal, we do take it incredibly seriously and there is an ongoing investigation. I will share if we do find anything and my below post is really a clarification on a specific pattern that we did see that I wanted to call out.
Go do something amazing today.
Claude Opus 5 더 나은 토��� 활용법
1. ‘Medium’ 노력으로 사용
더 나은 성능을 발휘하고 더 저렴, Opus 5는 과도하게 생각할 수 있음.
2. 출력 스타일을 ‘Concise’로 설정하세요
/config
output-style —> ‘Concise’
클로드의 고질적인 문제였던 장황한 응답을 줄여줌.
연속되는 컨텍스트에 더 효율적일 것이라 예상.
Opus 5 is a great model, when used correctly.
Do these two things:
1. Use it on ‘Medium’ effort - performs better and is cheaper. Opus 5 can overthink.
2. Set output style to ‘Concise’ - /output-style —> ‘Concise’. Cuts down on the techno-babble responses.
It also seems very efficient as the tasks get longer, burning down less usage on my Claude Max 5x plan slower than other models for long-horizon tasks.
$200 A MONTH FOR ONE DAY OF USAGE.
OpenAI Pro is cooked. I burned my entire weekly GPT 5.6 Sol limit in a single day of Codex. 0% left. Reset is 5 days away.
This is not the OpenAI I used to know. They went from the best limits in AI to the worst.
Meanwhile SuperGrok Heavy lets me run Grok 4.6 all day and I barely touch 10% of my weekly.
Grok is winning on usage and it is not close.
Fable 5 built this 3D world in Three.js with no external assets.
The project was almost entirely generated autonomously by Fable 5, with only occasional human feedback for visual details the model could not reliably inspect, like moving water and wind effects.
That makes the result a lot more interesting than a normal assisted 3D demo: the terrain, vegetation, atmosphere and world structure were effectively produced by the model itself.
The full prompt is also included in the repository, so the experiment is inspectable.
Repository:
https://t.co/xzY4YObztb
Project by: @SebastianKits