IGNIS

🧵1/5 AI can recommend a shopping list, but you still have to search every item again. I built DdakDama(딱담아) with GPT-5.6 + Codex: it turns natural language into exact products, sizes, strengths, pack counts, and quantities—then prepares a verified Coupang cart. #GPT #Codex

@thsottiaux Why can't I see my limit right away?

How can we make ChatGPT better. It’s already on your phone, but what do you wish would happen more or less of when you open the app. How can it bring you more joy?

@thsottiaux I am a Korean user, and Korean input behaves abnormally when using remote control on mobile. Typing lags, final consonants are often entered incorrectly, and editing text in the middle causes the cursor to jump to the beginning. This is very inconvenient.

Oops... I did it again. Enjoy reset usage limits for all paid users for Codex and ChatGPT Work. Super grateful for an incredible team who is iterating at lightspeed and keeping the infra up as we scale faster than ever. Enjoy the weekend!

@thsottiaux A 5-hour limit is not necessary.

We've had no 5h limit in Codex plus and pro for a few days. Do you think it is better or are you finding it difficult to manage the usage included in the weekly limit effectively? If we were to make this different, what should it look like in an ideal world?

Introducing GPT-Red An internal automated red teamer on a mission to find our models’ prompt injection vulnerabilities at scale, helping us build stronger defenses before wider deployment. openai.com/index/unlockin…

@mislead 이거 진짜 좋아

GO here in Palworld 1.0 And Get this necklace WAY BETTER than any anti heat or anti cold shirts or armor and its 1 accessory slot

Embarrassment of riches. But looks like we might hit 9M soon. Should we reset the ChatGPT Work and Codex usage again or give it some space?

🧵5/5 Codex helped me build the TypeScript monorepo, MV3 side panel, MCP Apps SDK widget, Cloudflare Worker, parser, cart state machine, and E2E regression suite. 58 automated tests and public MCP checks now pass. GPT-5.6 turns recommendations into accountable action. ✨

🧵4/5 DdakDama does not call a button click “success.” It verifies the product/vendor ID and the before/after cart quantity delta. If 4 of 5 items succeed, the result is a partial failure—not a green “done” screen. Checkout always remains manual.

🧵2/5 Accuracy is the product. “Magnesium 100mg, 240 tablets” means 100mg strength, 240 tablets per bottle, and 1 bottle—not 240 purchases. Two 50mL units can map to 2 singles or one 2-pack; a 3-pack is never silently substituted.

🧵1/5 AI can recommend a shopping list, but you still have to search every item again. I built DdakDama(딱담아) with GPT-5.6 + Codex: it turns natural language into exact products, sizes, strengths, pack counts, and quantities—then prepares a verified Coupang cart. #GPT #Codex

@thsottiaux It appears that plugins and skills can be used in general chat, but too many of them do not work properly. I selected the Sol Ultra model, but for some reason, I can only use a maximum of 3 subagents at a time.

Here you are! Thinking I am about to announce a reset. But no. I’m just scrolling twitter and looking for feedback on ChatGPT Work. What should we improve?

GPT-5.6 sol is half the price and ~twice as token efficient as fable in many cases for accomplishing the same task. happy to deliver at one-quarter of the price.

Tomorrow might be 8M active user celebration day. Just saying

"hard questions are great but only if we deem you worthy enough to not silently downgrade you, or even get access at all"

i gave 5.6 sol access to my camera roll and had it extract pictures of every piece of clothing i own from my photos then, told it to find new outfits for me and render them on me with gpt-image! its kinda cool to see your entire wardrobe in a collection like this

v1 of Command Code is almost here. testing in final stages. faster, crisp, and built around the way you actually work. can't wait to roll it out to y'all. P.S. thanks to all our amazing beta testers who are helping us ship this. 💜

If you missed it, I've published a small framework/workflow on how I've built my various games. From the initial pitch, to generating concept arts and letting Codex set a recursive self improving goal. The Infinite Build the-infinite-build.openai.chatgpt.site

5. Computer Use and Chrome Extensions Codex Computer Use stops when it cannot verify a Chrome URL as safe. It also blocks chrome-extension:// pages, preventing automated extension development and testing. Please allow secure access to verified Chrome and extension URLs.

4. Steering in Codex GPT inside Codex does not support Steering. Users cannot redirect, refine, or add instructions while a response is being generated. Please add Steering so ongoing tasks can be guided without stopping and restarting them.

@thsottiaux 1. File Downloads GPT Classic allowed users to download generated files, but GPT inside Codex does not appear to offer this option. Please clarify whether downloads are unsupported or hidden, and provide a clear way to save generated files.