After working for 2-3 months with codex, switched to claude for opus 5.5. Its the best model I have ever used but man these safety guards are irritating me. Any routing available for these kinda tasks automatically? Also heard fable 5.5 has this automatic routing, is it true?

Congrats for 1B$ to the product that made me video editor 🫡🫡🫡🫡

Been using opus 5.5 for past 2 days. Pretty impressed tbf. Found 200 bugs in my code, flow and logic. Resolved and improved them. Fyi: I am building this product from past 5 months, so multiple models are being used to build this. Still rooting for codex tho, I am pretty much using both models, my codex and claude code are talking to each other at this point

Personally I will still use Codex for image model itself. Demanding Claude to release one

I am not seeing opus 5.5 proofs or justification of the benchmarks that Claude have released from the community. WHATS HAPPENING?

Good morning to beautiful brains out there!!!

I upgraded my 5x plan to 20x plan in codex yesterday. All my feed is saying openAI is still not allowing new pro accounts. What happened here? Did I upgraded while they are testing something? Anyway thanks for the opportunity @ChatGPT . Hope you won't remove it now :)

Today I realized how to use codex at its best even for frontend. I feel guilty that I didn't realize or know this till now. Till now even with Astra I feel gpt is making UX terrible comparable with opus. I cancelled my claude sub after Astra release and I wondered how can I complete my UI/UX. Then I gave a shot with new gpt image. The UX of my mobile app turned out great than what Astra decided to do. So I created few options of images using gpt image and finalised one and gave it to Astra. It worked. I think this might be the jugaad to make great UI with Codex. I discovered this now though many might already know this, but let me know if the top brains feel there are any other better ways. PS: I tried so many skills.

WSJ: OpenAI quietly bought Glass Imaging in recent months. Ex-Apple camera AI, valuation above $300M. Aims for DSLR-like phone photos. No product plan disclosed.

Siri AI is live as an English beta. The interesting part is still locked: private Inference Providing hooks already hand Apple's planner prompt and tools to GPT-5.6 Terra. Entitlement isn't public. EU iPhones stay blocked.

X and SpaceXAI just got Apple dismissed from their Texas antitrust suit over ChatGPT in Apple Intelligence. Judge Pittman granted it. No reason filed. Claims against OpenAI continue. reuters.com/world/musks-x-co…

Contractors under OpenAI's Project Lily read real ChatGPT chats to rate replies. Model improvement is on by default for Free/Plus/Pro. Turn it off under Data Controls if you assumed only the model was reading.

Temporal raised $550M at $12.55B. Lightspeed led. ARR up 200%+. OpenAI's Temporal usage grew 60x in under a year. Durable execution for agents is where late-stage capital went today.

Schumer just demanded a classified all-senators AI briefing. Three asks: the still-unpublished frontier testing framework, chip/tool flows to China, and AI cyber defense for hospitals and grids. democrats.senate.gov/newsroo…

Trump just named Amodei. Only AI guardrail: a "STRONG AND SMART (High IQ!) PRESIDENT." Calls safety talk a "SICK conspiracy" that helps China. Says Dario is pretending to be a "perfect little angel."

Six-week public draft for MAI models. Never resist shutdown. No neuralese. Rejects AI personhood. Glossary caveat: it does not cover OpenAI/Anthropic models Microsoft hosts. Not training on it today; revised version guides 2027.

UK JCHR report: new AI Bill, statutory regulator with sanction powers, prior approval for high-risk systems, and bans on some biometric/profiling uses. AISI engagement stays voluntary until Parliament actually legislates.

Burry says pacing talk is IPO puffery and cover for slowing growth. OpenAI already delayed past 2026. Anthropic is still aiming Nasdaq in October on a climbing ARR story. The calendars don't match the slogan.