GPT 5.3 Codex has achieved the highest success rate on Next.js evals. View the leaderboard ↓ vercel.fyi/next-evals-2-5

do you remember when codex needed hours or days to implement a feature? these days are gone

Very courageous stand

Today we shipped Handoff in the Codex app: a simpler way to move a thread between Local and Worktree.

Hello, Marcel Pociot 🧪! Love the Opinion feature in Polyscope! Would it be possible to add Reasoning / Fast mode selection for Codex?

as I was reaching my weekly limit set Codex fast mode to off. after 15 minutes I am back on fast mode and don't care about the limit 😅 too slow

figma mcp server is awesome!!

🚨 Major wake-up call for AI users 🚨 A federal court has ruled in United States v. Heppner (Feb 2026): Your chats with public generative AI (ChatGPT, etc.) are NOT protected by attorney-client privilege. Why? You're sharing with a third party → no confidentiality. AI is not

not again ...😒

don't use worktrees. just don't

OpenAI 'lowered' the Business plan by $5 but nerfed Codex so hard the 5h limit runs out after 1-5 prompts. That's not flexibility, that's a price hike with extra steps.

my timeline just now 🤣🤣

I've been using codex cli for months. Tried codex app today - it's amazing 🚀 they got game

While Claude Code is still drafting a plan, Codex has already shipped the change.

is there a difference between high and xhigh? check this out

great 😵‍💫: techcrunch.com/2026/05/01/ubu…