@OpenAI why is codex not able to parse pdfs by default. I really want to use this as a competitor to cowork, better model, amazing computer use. But the simplest things are not working.
Simple new eval.
Detecting the correct score from a dartboard photo is unexpectedly hard for LLMs.
The task appears to stress spatial reasoning: Gemini 3 models lead this benchmark by a decent margin.
Counterintuitively, “more reasoning” often reduces accuracy.
Even the top-performing model scores only ~36% of darts correctly.
https://t.co/2sntQn6PxT
@sobedominik It’s just another phone. If the 14 has decent performance and battery life than it’s not worth it inmo. Upgraded from 12pro to 17pro. Battery is much better. This is really the only diff I experience on a daily basis.
Cameras are obv better especially in darker light.
@OpenAIDevs can we get drag & drop support for the vscode/cursor plugins & better search?!
almost impossible to work on a nextjs project where every file is called route.ts or page.tsx
Built VibeIcons over the weekend based on nano-banana. It allows you to generate icons in the style of your favourite icon library - gets you an svg output.
Probably not life-changing, but a fun little weekend project.
https://t.co/O6RpTwQiGO
If you're not building a mini startup with Nano Banana today and launching it tomorrow, you're missing the opportunity of a lifetime
This image model just made 100s of new startups and apps possible
Only limit is your creativity and how fast you can ship a user interface and put a Stripe payment button on it
I think
Mini Weekend Project ✅
VibeIcons: let's you generate AI SVG icons in the style of your favourite OSS library.
Why? At some point in my projects, I am at a point where I want a custom icon that I can't find in my library of choice. VibeIcons solves this.