shiv
90 posts


openai’s imagen 2 is really strong for ui design.
use it to generate ui concepts and visual directions.
use claude for frontend implementation.
use gpt 5.5 for backend logic, architecture, and reasoning.
feels like openai and claude are slowly creating stronger niches in the coding workflow instead of competing in the exact same way.
English


Show us the Codex pets you hatched.
Use /hatch to create your own Codex pet.
We’ll pick 10 favorites to get 30 days of ChatGPT Pro.
OpenAI Developers@OpenAIDevs
Customize your Codex pet with /hatch
English

@GoogleAIStudio Have you upgraded your rate limits?? It's a competition these days
English

@OpenAIDevs What currently lacks in llms is creating ui that's creatively new.
To solve that there should be a mode. Like creative randomizer . What i currently feels imagen 2 is better at ui creative then gpt 5.5.
Imagen truly imagines...gpt 5.5 is limited because it thinks in code.
English

With the Figma plugin, Codex can now turn implementation plans into visual FigJam boards.
Figma@figma
MCP updates in FigJam so you can visualize your systems (and not just read code) → generate_diagram to create architecture diagrams & ERDs → figma-use-figjam skill to add notes, code blocks, and annotations directly to your board → get_figjam tool to read your board & map out next steps
English

@OpenAINewsroom @amazon @awscloud Three things for enterprises.
Integration is easier for AWS infra.
It will feel like a closed box model for enterprises.
Distribution chain improved.
English

Earlier this year, OpenAI and @amazon partnered to bring OpenAI’s frontier capabilities to enterprises, startups, and customers around the world.
We’re taking the next step: making our models, Codex, and Bedrock Managed Agents available to @awscloud customers, in limited preview.
Making OpenAI available on AWS means enterprises can get AI into production faster - across software engineering and other professional workflows.
We’re excited to see what gets built!
openai.com/index/openai-o…
English

@DeepLearningAI @Oracle This can be called moat, either you ship the raw intelligent llm models or you ship the framework on which llm runs..giving them memory!!
English

Make sure to drop by the @Oracle booth at AI Dev 26 to learn more about the unified memory core for AI agents!

English

@OpenAIDevs Well updating a repo to a new model means find and replace right,
Plain and simple find and replace? Don't underuse gpt 5.5 on trivial tasks!!!
English





